Anand my best games of chess (expanded edition)

Vishy Anand: My Best Games of Chess

Revised and expanded edition

Vishy Anand (in collaboration with John Nunn}

CdAI�IBIITI

First published in the UK by Gambit Publications Ltd 200 I Reprinted 2004, 2008 Original edition published by Gambit Publications Ltd 1998 Copyright© Vishy Anand and John Nunn 1998, 2001 The right of Vi shy Anand and John Nunn to be identified as the authors of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. This book is sold subject to the condition that it shall not, by way of trade or otherwise, be lent, re-sold, hired out or otherwise circulated in any form of binding or cover other than that in which it is published and without a similar condition including this condition being imposed on the subsequent purchaser. ISBN-13: 978-1-901983-54-8 ISBN -10: 1-90 1983-54-4 (First edition: ISBN-13: 978-1-90 1 983-5 ISBN-10: 1-901983-00-5)

DISTRIBUTION: Worldwide (except USA): Central Books Ltd, 99 Wallis Rd. London E9 5LN, England. Tel +44 (0)20 8986 4854 Fax +44 (0)20 8533 5821 . E-mail: [email protected]

Gambit Publications Ltd, 99 Wallis Rd, London E9 5LN, England. E-mail: info@ gambitbooks.com Website (regularly updated): www.gambitbooks.com

Edited by Graham Burgess Typeset by John Nunn Printed in Great Britain by The Cromwell Press, Trowbridge, Wilts. Cover photograph by Dagobert Kohlmeyer.

1 0 9 8 7 6 5 4 3

Gambit Publications Ltd Managing Director: GM Murray Chandler Chess Director: GM John Nunn Editorial Director: FM Graham Burgess German Editor: WFM Petra Nunn Webmaster: Dr Helen Milligan WFM

Contents

Introduction 5

I V. Anand- v. Inkiov, Calcutta 1986 7 2 V. Anand - K. Ninov, World Junior Championship,

Baguio City 1987 1 2 3 V. Anand - S. Agdestein, World Junior Championship,

Baguio City 1987 16 4 V. Anand -J. Benjamin, Wijk aan Zee 1989 23 5 M. Thl- V. Anand, Youth vs Veterans, Cannes 1989 29 6 V. Anand - B. Spassky, Youth vs Veterans, Cannes 1989 34 7 M. Kuijf- V. Anand, Wijk aan Zee 1990 40 8 M. Petursson - V. Anand, Manila Interzonal/990 43 9 V. Anand - I. Morovic Fernandez, Novi Sad Olympiad 1990 49

10 A. Beliavsky - V. Anand, Munich 1991 54 1 1 A. Karpov - V. Anand, Candidates match (6), Brussels 1991 59 1 2 V. Anand - G. Kasparov, 1ilburg 1991 67 13 G. Kasparov - V. Anand, Reggio Emilia 199112 74 14 V. Anand -E. Bareev, Dortmund 1992 8 1 15 V. Anand - R. Hiibner, Dortmund 1992 87 16 V. Anand- I. Sokolov, SWIFT rapid, Brussels 1992 90 17 V. lvanchuk - V. Anand, Match ( 1 ), Linares 1992 92 18 V. Anand - G. Kamsky, Alekhine Memorial, Moscow 1992 98 19 V. Anand - V. Ivanchuk, Linares 1993 102 20 B. Gelfand - V. Anand, Linares 1993 106 21 V. Anand- E. Bareev, Linares 1993 1 10 22 V. Anand - F. Izeta, Madrid 1993 1 15 23 V. Anand - L. Ftacnik, Biel Interzonal 1993 1 19 24 L. 011 - V. Anand, Biel Interzonal/993 126 25 M. Adams - V. Anand, European Clubs Cup Final,

Hilversum 1993 130 26 V. Anand - A. Be1iavsky, PCA Qualifier. Groningen 1993 1 36 27 J. Benjamin - V. Anand, PCA Qualifier. Groningen 1993 142 28 V. Anand - G. Kamsky, Linares 1994 148

4 VISHY ANAND: MY BEST GAMES OF CHESS

29 V. Anand- J, Polgar, Linares 1994 153 30 V. Anand-G. Kamsky, PCA Candidates ( 3 ), Las Palmas 1995 157 31 V. Anand- G. Kamsky, PCA Candidates (9), Las Palmas 1995 164 32 V. Anand-G. Kamsky, PCA Candidates ( 11 ), Las Palmas 1995 171 33 V. Anand -J. Timman, Tal Memorial, Riga 1995 176 34 V. Anand - G. Kasparov, PCA World Championship (9),

New York 1995 184 35 V. Anand -B. Gelfand, Wijk aan Zee 1996 189 36 V. Anand - J. Polgar, Amber Rapid, Monte Carlo 1996 198 3 7 V. Anand-V. Topalov, Donmund 1996 202 38 V. Anand-V. lvanchuk, Las Palmas 1996 207 39 V. Anand-A. Karpov, Las Palmas 1996 2 l l 40 V. Anand -J, Lautier, Biel/997 216 41 J. Lautier - V. Anand, Biel/997 220 42 V. Kramnik - V. Anand, Belgrade 1997 225 43 P. Nikolic- V. Anand, FIDE World Ch., Groningen 1997 234 44 V. Anand -A. Shirov, FIDE World Ch., Groningen 1997 238 45 V. Anand- A. Karpov, FIDE World Ch. Final (6),

Lausanne 1998 244 46 V. Anand-V. Topalov, Wijk aan Zee 1998 249 4 7 V. Ivanchuk-V. Anand, Linares 1998 253 48 V. Anand-V. Kr amnik, 1ilburg 1998 2S7 49 V. Anand-L. 011, European Clubs Cup, Belgrade 1999 262 50 D. Reindennan-V. Anand, Wijk aan Zee 1999 269 5 1 V. Anand -J. Piket, Wijk aan Zee 1999 273 52 V. Anand-P. Svidler, Linares 1999 277 53 V. Topalov-V. Anand, Linares 1999 283 54 V. Anand -P. Nikolic, Wijk aan Zee 2000 291 55 V. Anand-A. Khalifman, FIDE World Cup, Shenyang 2000 297 56 V. Anand-M. Adams, FIDE World Ch., New Delhi 2000 304 57 V. Anand - A. Shirov, FIDE World Ch. Final (4), Teheran 2000 3ll

Combinations Solutions

Index of Opponents Index of Openings Symbols

316 326

335 336 336

Introduction

I don't want to spend too much time on biographical details, because this is a book about my games, so I will be content with a brief sketch.

I was born on 11th December 1969 in Chennai (Madras), and learnt chess at the age of six from my mother. A year later I joined the TaJ. Chess club in Chennai. A couple of years later I went to Manila when my father had an assignment there and got caught up in the Philippines chess fever resulting from the 1978 Karpov-Korchnoi World Championship match. Returning to India, it was not until 1983 that I achieved a real breakthrough. In that year I

won both the national Sub-Junior (under- 16) and Junior (under- 19) titles, and qualified for the (adult) national championship. The championship itself was held the following year, and I finished fourth. In 1985 I gained my International Master title; at the time I was the youngest Asian ever to achieve this distinction. In 1986 I won the National Championship and became India's youngest champion.

The following year, 1987, was special. After having played three times before in the World Junior Championship, finishing lOth, 5th and 7th, I finally won this title. In December of the same year I became the world's youngest grandmaster (at that time).

In July 1990, I broke through the 2600 barrier. In the same month, I qualified for the Candidates from the Manila Interzonal. The first round of the Candidates was held in January 1991, and I won my match against Dreev, but in the quarter-finals I lost narrowly to Karpov in a match that hinged on the final game. After this disappointment, I concentrated on tournament play with some success: I won outright at Reggio Emilia 199 112 (ahead of Karpov and Kasparov) and in Moscow (November 1992).

In 1993 I qualified for both the FIDE and PCA Candidates cycles. The following year I was successful in the PCA Candidates cycle, defeating first Romanishin and then Adams to reach the Final of the Candidates. In the FIDE cycle I was less fortunate, losing to Kamsky after having been two games up with three to play.

In early 1995 I defeated Kamsky in the Final of the PCA Candidates and gained the right to face Kasparov for the PCA World Championship in New York. As everybody knows, I started well but faded in the second half.

6 VJSHY ANAND: MY BEST GAMES OF CHESS

However, losing this match was not the end of the world and in 1996 I was back on the tournament trail, finishing joint first with Kramnik at Dortmund and beating Kasparov in the Final of the Geneva Quickplay. However, the most important event of 1996 had nothing to do with chess: on June 27th I married Acuna in Chennai.

1997 was a successful year. In April I finished joint first with Kramnik at Dos Hermanas and in May I became the first person to win the Amber tour nament in Monaco twice. There followed a win in the Frankfurt Rapid, second place in Dortmund and outright first in Biel. Finally, I tied for first place with Ivanchuk at Belgrade in November.

At the end of 1997 I participated in the FIDE World Championship held at Groningen, and since I qualified for the final at Lausanne this event spilled over into 1998. I reached the final against Karpov, but then tiredness took its toll and, despite a 3-3 result in the match itself, I lost the tie-break. Despite this early disappointment, 1998 was my most successful year up to that time. I won the chess Oscar, and took first place in five major tournaments at Wijk aan Zee, Linares, Madrid, Frankfurt and Tilburg.

1999 started well, since I scored +6 at Wijk aan lJ!e to finish half a point behind Kasparov, but the rest of the year was rather disappointing. The problems started at Linares, where I lost a crucial game to Kasparov. Whether this was the sole cause is hard to say, but at any rate I struggled with my form for several months thereafter.

The new millennium began with a modest success at Wijk aan ZJ!e, Where I finished joint second, but once again Linares proved discouraging. However, from this point on my form rapidly improved, and my play regained the freshness and vigour which had been the foundation for my run of succes in 1998. First I won the advanced chess event in Leon and the Frankfurt rapid, and then I was joint first at Dortmund. A further win in the FIDE World Cup in Shenyang left me in excellent shape for the crucial event of the year- the FIDE World Championship. This time there was no disappointment at the last hurdle, and I achieved my greatest triumph to date with a decisive 3112-'h victory in the final against Shirov.

Now I will sign off and leave the readers to enjoy the games.

Vishy Anand

Collado. Spain June 2001

Game 1

V. Anand - V. lnkiov Calcutta 1986

Sicilian, Richter-Rauzer

1 e4 cS 2 lLi3 d6 3 d4 cxd4 4 �d4 .!M6 s 00 � 6 i.gS e6 7 'l'd2 i.e7 8 0-0-0 0-0 9 ll:lb3

In 1986 this move was in vogue thanks to the efforts of Mikhail Tal.

9 aS 10 a4 dS 11 i.bS

A move first played in Tal-Sismega, Taxco Interzonal 1985, even though Tal ascribes the move's invention to Vitolin�. Tal won that game and scored an even greater success when he beat Korchnoi with it at the 1985 Montpellier Candidates Tournament. I figured that I could do worse than to follow in Tal's footsteps.

11 ll:lb4 Sisniega played l l ... lt:lxe4 and

Kon:hnoi l l ...dxe4. l l ...i.b4 is another possibility; after 12 exd5 exd5 13 'l'f4, followed by exchanges on c3 and f6, both sides will end up with fract.•red pawn structures, but

it will be difficult for Black to get at White's c-pawns, whereas White will attack the d5-pawn. The move l l ...ll:lb4 had been played before, in Vitolins-Inkiov, Jurmala 1985.

12 l:the1 (D)

12 .•. dxe4 After 12...'illc7 13 e5 ll:le8 14

ll:ld4! Black is in a bad way. Due to the weakness of the e6-pawn, Black can't play ... f6 and his e8-knight is badly placed.

13 'ii'xd8 Really the first new move of the

game. Vitolins played 13 ll:lxe4lllxe4 14 11rxd8 �xg5+ 15 'ii'xg5/t:\xg5 16 h4, regaining the piece with an unclear position.


13 ... l:l.xd8 Mter l 3 ... ..txd8 14 lllxe4 ..te7

(l4 ... lllxe4 15 ..txd8 lllxf2 16 l:l.d2 is very promising for White; Black has no reasonable defence to the threats of 17 ..txa5 and 1 7 ..te 7) 15

lllxf6+ ..txf6 (l5 ... gxf6 16 ..th6 is similar to the game) 16 ..txf6 gxf6 17 l:l.d6 Black can play neither ... b6 nor ... e5, and so has serious problems developing his queenside.

14 lllxe4 lllbd5 (D)

15 c4 A positionally ugly move, but the

main thing is to exploit White's lead in development. To this end, White must keep the d-file open.

15 - !i:Jc7 15 ... lllb4 occupies the 'hole' cre

ated by White's previous move, but then 16 l:l.xd8+ ..txd8 17 l:l.d 1 ..te7 18 llld6 e5 ( l 8 ... b6 19 ..te3! also nets a pawn) 19lllxc8 l:l.xc8 20 lllxa5 wins a pawn.

16 l:l.xd8+ ..txd8

17 l:l.d1 Everything with tempo.

17 ..te7 (D)

18 ltlxf6+?! White retains the advantage after

this move, but Black could have put up more resistance. It was probably stronger to play 18 llld6!. I would not claim that this is a forced win (Larsen once observed that all long variations are wrong!), but Black is definitely in real trouble:

l ) 18 ... b6 19 ..tc6! (not 19 ..te3 because now Black can reasonably play 19 ... lllxb5! with a satisfactory position after either 20 cxb500! or 20 axb5 transposing to line 2a be· low) 19 ... l:l.b8 20 ..tf4! and White's pieces occupy dominating positions.

2) 18 ... lllxb5 and now there are two possible lines:

2a) 19 axb5 b6 (best; 19 .. . a4 20 llla5! and 19 ... h6 20 ..txf6 ..txf6 2 1 lllc5 are more promising for White) 20 ..te3 a4 21 llla 1 llld5! 22 cxd5

ANAND - INKIOV, CALCUITA 1986 9

.bd6 23 dxe6 i.e? 24 exf7+ �xf7 with fair compensation for the pawn.

2b) 19 cxb5! (this concedes the d5-square, but Black can't make full use of it) l9 ... b6 20 ltlc4 (20 i.e3 �!as in line l) 20 ... llb8 21 ltld4 (intending lLlc6) 2 I ...i.b7 (2l...i.d7 22 ltle5 is also good for White) 22 l0xb6 i.xg2 23 ltlc4! (although Black has the two bishops, the mass of pawns on the queenside is the most important factor in the position; 23 ltld7 is less accurate since after 23 ... llc8+ 24 �bl i.e4+ things are getting quite messy) 23 ... llc8 24 b3 (White only needs to play �b2 and ltlxa5 to decide the game with his queenside passed pawns; however, Black can try to win a pawn) 24 ... i.d5 25 �b2! i.xc4 26 llcl ! (when I checked this position with Fritz, it said that Black was winning! Surprised, I looked to see why) 26 .. . :C5 (this is the reason, but after some thought I found a solution) 27 llxc4 llxg5 28 :lc8+ i.f8 29 b6! (Black is helpless) 29 .. .:c5 (after 29 ... ltld7 30 b7 :ld5 31 ltlc6 Black has no defence to both 32!De7+ and 32 :ld8) 30 b 7 ltld7 31 ltlb5! and Black cannot meet the threat of 32 :ld8.

18 gxf6 19 i.e3 lLlxbS 20 axbS fS!

Black fights back. This gives his bishop some air and prevents White forcing the exchange of bishops by

i.c5. 20 ... e5 is worse, as after 21

i.c5 i.xc5 22 ll:lxc5 i.g4 23 lld5 :lc8 24 b3 Black's queenside pawns are in trouble.

21 &s (DJ The alternative 21 f4 ambitiously

attempts to squash Black, but he can free himself by sacrificing a pawn: 2l...e5! 22 i.c5 (22 fxe5?! i.e6 23 li:ld2 a4 prevents White supporting his c4-pawn by b3, and then the c4-pawn itself is attacked by .. . i.b4 and ... :lc8) 22 ... i.f6 (not 22 ... i.xc5?, when 23 ll:lxc5 gives White exactly what he wants) and now 23 fxe5 fails to 23 ... i.g5+ 24 �c2 i.e6 and Black is doing well as White cannot hang on to all the pawns (for instance 25 �c3 :lc8, followed by some combination of ... a4 and ... b6).

21 ••• eS? The best defence was 2 l... f4! 22

i.d4, and now: I ) 22 . .. e5 23 i.xe5! i.xc5 24

lld8+ i.f8 25 i.d6 and White wins.


2) 22 . . . a4 (by not allowing lt:la4, Black gets some counterplay) and now:

2a) 23lt:ld3 i.d7 (23 ... f6? is very good for White after both 24 i.c5 and 24 tt:lxf4 e5 25 tt:ld5) 24 tt:lxf4 .:.C8 25 i.c3 i.e8 and Black has sufficient counterplay.

2b) 23 tt:le4 f5 (23 ... e5? 24 i.xe5 i.e6 25 c5) 24 i.c5 and White retains some advantage.

3) 22 ... f6 (a solid and sensible defence; Black prepares ... e5 followed by ... �f7) 23 tt:la4! e5 24 i.c5 �n 25 i.xe7 �xe7 26 tt:lb6 .Z:.b8 with a slight advantage for White.

Thus 2 l ...f4, while not equalizing, would have restricted White's advantage. The importance of driving the bishop to d4 becomes clear after the text-move.

22 tt:ld7! From this excellent square the

knight virtually paralyses Black's whole army. The game is already almost over.

22 f4 Too late, as now the bishop need

not block the d-file. 23 i.b6 f6 24 i.c7!

Since 25 tt:lb6 is threatened, Black can no longer delay capturing the knight; the result is that White's rook occupies the seventh rank.

24 i.xd7 25 lbd7 i.cS 26 i.d6 i.xd6

White also wins after 26 ... i.�f2 27 :xb7 :ds 28 :bS! (not 28c5? i.xc5 29 i.xc5 .Z:.c8) 28 .. J[xb829 i.xb8 �f7 30 i.e? and the threat of 3 1 b6 forces Black to approach with his king, whereupon 3 1 i.xa5 gives White an overwhelming mass of passed pawns.

27 :xd6 (D)

Black's position is lost. The active rook, combined with White's queenside pawn majority, guarantees a straightforward win.

27 28 b3 29 �b2 30 �xb3

a4 axb3 �

Just abandoning the b-pawn, but 30 .. .f5 31 :d7 :bs 32 c5 is also hopeless.

31 :d7+ �e6 32 lbb7 e4 33 :a7 e3 34 fxe3 fxe3 35 �c3 :ds

36lW 37 bC'i! 38 b7 39 �b4

ANAND - INKIOV. CALCUTTA /986

40 b81t' 41 1tb7+ 42 'i!t'dS+ 43 c5

lba2 �d3 �e2 1-0

11

The 1987World Junior Championship was a breakthrough for me. It was already my fourth World Junior Championship; 1 had played in every one since 1984, finishing lOth in my first appearance, 5th in 1985 and 7th in 1986, a fairly zigzagging pattern. I started the 1987 event in fairly good shape but I drew 3 out of my first 4 games, and at that point I didn't imagine that I might win the tournament However, the following game was a turning point.

Game 2

V. Anand - K. Ninov World Junior Championship, Baguio City 1987

Sici l ian , Kan

1 e4 cS 2 lbf'3 e6 3 d4 cxd4 4 liJxd4 a6 s �d3 �cS 6 lbb3 fLa7 7 lbc3 lbc6 8 'ife2 d6 9 h3 �xe3

As a result of this game, playel'll with Black started delaying this exchange so as to leave the white queen on e2 and thereby prevent the �e2 line mentioned in the note to Black's 13th move. In time, this led to the development of a whole new branch of opening theory.

10 'ifxeJ lbf6 (D)

w

11 g4! I found this innovation over the

board. Previously White had continued I I 0-0-0, but after 11 ... 0-0White either has to prepare this advance with llhgl, or play g4 as a pawn sacrifice. The idea behind the immediate g4 is quite simple: if you play it when g7 is undefended, then it isn't a sacrifice.

11 •.• bS Black thought for a while, and

then decided simply to proceed with his qutenside counterplay. However, the tempo White saves by missing out llhgl is quite important, and he should have tried l l ...lDxg4. After 12 'ifg3 lbf6 13 'ifxg7 llg8 14 'ifh6 �d7 Black will play ... 'ife7 and both sides will castle queenside. However, when Black castles queenside in the Sicilian, White is better unless there is some mitigating factor, which is not the case here. Indeed, after 'ife3 by White there is an awkward weakness on b6 and the d6-pawn it· self may become vulnerable later.

12 0.0.0 0-0 13 gS lbe8

A poor square for the knight, but after 13 ... lbd7 14 �e2 Black has

ANAND - NINOV, BAGU/0 CITY 1987 13

immediate problems with his d6-pawn, due to the exchange of darksquared bishops and White's extra tempo. Usually White doesn't go after d6 in a Sicilian, but that doesn't mean you should forget about the possibility altogether!

After the text-move, there is no point to :bgl, which would throw away the advantage gained as a result r:i White's innovation, so White has to come up with an alternative attacking plan.

One possibility is h4-h5 followed by :dgl. The arrangement with rooks on gl and hi is very desirable, because they will break through almost any kingside defence, but it is very time-consuming to set up. Unlike many similar positions in the Sicilian, Black's queenside counterplay is rather slow here, which is the only reason White can consider this plan, but in the end I decided on a more conventional approach.

14 f4 b4 15 lDe2

15 � would be a more positional formula. White stops ... a5-a4 and threatens to invade on b6. After l5 ... llb8 White can continue 16 e5, in order to clear the c5-square for the Imight However, I preferred to play for the attack, and for that the knight is needed on the kingside.

15 aS 16 ltlbd4 ltlxd4 17 ltlxd4 (D)

White has the advantage. Normally in the Sicilian, Black's thematic queenside pawn advance gains time because it hits minor pieces on b3 and c3, but here White has evacuated these squares quite quickJy. Moreover, the knight on e8 is very badly placed for supporting the attack. Just about the only useful thing it can do is to shore up the kingside by ... g6 and ... ltlg7.

17 ••• 'ilfb6 At the time I felt that Black should

have gone in for 17 ... a4, but now I don't think so. The line that worried me was 18ltlc6 'iWc7 19 o!l'!xb4 a3 (if Black doesn't play this, then White plays a3 himself, followed by lilbl and c3, and Black will never break through) 20 b3 (threatening to centralize with 'il'd4) 20 ... 'iWc3 (after 20 .. . llb8 21 o!i)a6 .>\.xa6 22 .>\.xa6, followed by .>\.c4, White's queenside position is solid since the poorly placed knight on e8 cannot displace


White's bishop), but in fact 21 Wei! ifh2+ 22 'it>d2 is clearly better for White- Black has surprisingly little compensation for the pawn.

These were still the pre-computer days, when players were much more intuitive. Nowadays everybody goes home and checks everything with Fritz. The use of computers has made people more sceptical and now they are more prone to go pawn-grabbing unless there is definite compensation.

18 eS .i.b7 19 mtfi(D)

19 dxeS Or 19 ... a4 (1 9 ... :d8 20 fS! is simi

lar) 20 fS! dxeS (20 ... exfS 21 e6 is very good for White) 21 fxe6! (not 21 'ibeS :aS!) 2l ... exd4 221Fh3 g6 (22 ... fS 23 l:txfS! gives White a winning attack) 23 exf7+ :xt7 (23...�h8 24 'i'h6 �g7 2S :C6 wins) 24 :Xf7 �xf7 25 'i'xh7+ winning the black queen.

20 fxeS l:td8? (D) Allowing a pretty finish. 20 ... g6!

was best, when White could continue 21 �S 'llrxe3+ 22ll'lxe3 with a pleasant endgame- his knight could head for either f6 or d6. Still, Black has some chances by playing his knight to fS, and he shouldcertainly have gone for this.

21 .i.xb7+! �xb7 22 g6+

Once again Black suffers because of his miserable knight position Here it prevents Black from playing 22 ... fxg6.

22 ••• �g8 Or 22 .. .'>1i>xg6 23 'i1Vd3+ (stopping

... �h7) 23 .. .fS (23...�h6 24 'i'h3+ �gS 2S :gl + �f4 26 :del with mate next move) 24 exf6+ with a decisive attack.

23 'i1Vb3 .!ffii 23 ... fxg61oses to 24 l:txfl!+ 'iPxfl!

2Sll'lxe6+. 24 exf6 (D)

8

ANAND- NINOV, BAGU/0 CITY 1987 15

It isn't every day you see two attacking pawns on f6 and g6! When you get a position like this, you go away feeling very pleased and have a warm glow for the next few games.

24 ••• fxg6 25 fxg7 1-0

Since 25 ... \ii>xg7 26 ltlxe6+ and 25 ... llxfl 26 'l'h8+ �f7 27 :Xn + are decisive.

After this game, which was from round 5, there was a free day and the players went on an excursion. I still hadn't broken free from the pack. The following day, however, saw another good result. I was facing Agdestein, who at 2565 was the highest-rated player in the tournament, although he was not the only grandmaster (Ivan Sokolov was also participating, although he was not very successful).

I prepared as well as I could because Agdestein is very unpredictable in the openings, and sat down hoping to play a good game.

Game 3

V. Anand - 5. Agdestein World Junior Championship, Baguio City 1987

Ruy Lopez

1 e4 tt'lc6 A provocative move. At the time

it was felt that this move had little independent significance because so long as White knew he should play 2 tt'lf3, Black had nothing better than 2 ... e5, thereby transposing into standard king's pawn openings. Subsequently it was discovered that Black could wind White up even more by playing 2 ... d6, but I don't trust this for Black!

2 tt'lf3 eS 3 i.bS a6 4 i.a4 b5 5 i.b3 tt'la5

Agdestein is fond of offbeat systems. I didn't know much about this one, although it is popular amongst Norwegian players. I could only remember a game between Spassky and Taimanov (in fact from the 1955 USSR Championship, held in Moscow). Nevertheless, I was quite happy to see it on the board. White can play natural moves and there is not much risk even if he commits a slight inaccuracy - a pleasant situation when facing the top seed!

6 0-0 7 d4

d6 tt'lxb3

8 axb3 f6 9 tDc3 i.b7

I recalled that Spassky had played lO tt'lh4 in the above- mentioned game, with dxe5, 11Vf3 and l:tdl following in some order, and that later. Spassky sacrificed a piece by means of lL!xb5.

10 tt'lb4 tt'le7 (D)

11 dxe5! dxe5 If Black plays 11 ... fxe5, then White

can strongly reply 12 f4, opening the position up while Black's king is still stuck in the centre. However , taking back with the d-pawn retains control of gS, so that f4 can be met by ... exf4 followed by the fork ... gS·

12 1Wf3

ANAND - AGDFSTE/N, BAGU/0 CITY 1987 17

White avoids the exchange of queens as most of his chances lie in exploiting Black's poor development and centralized king.

12 •.• 'i'd7 13 l:.dl 'i'e6

By now I had worked out that this was indeed what had happened in the Spassky game, and that he now continued with 14 i.e3. For a moment I wondered what Taimanov had played that allowed the sacrifice on b5, then! realized that it was 14 ... g5. After 15 �xb5! axb5 16 W'h5+ 'ilf7 (16 ... �g6 17 �xg6 'l/lf7 18 l:.xa8+ .baS 19 1W g4! also wins) 17 ltxa8+ .ixa8 18l:.d8+ �xd8 19 'il xf7 gxh4 20 'l'xf6 White had a winning posi-lion.

However, I was worried by the idea of 14 ... h5 ! ? , intending .. .'I' g4. I mulled over this for some time, but couldn't see an easy answer. If White plays 15 h3, then 15 ... g5 is now possible because White has no queen check on h5. If 15 li:ld5, then Black just castles queenside. It appeared to me that if White was going to play 00, then it would be better to do it straight away, before Black bad time to set up the threat of ... 'l'g4.

14 li:ldS li:lxdS If 14 ... 0-0-0, then 15 c4 and White

is already starting to make Black's king feel insecure.

IS exdS 'l'f7 16 c4?! (D)

A slight inaccuracy which gives Black the chance to sideline the knight on h4. 16 li:lf5 would have been more accurate, because White can play c4 at any time- Black can't prevent it. Then 16 ... g6 17 li:lh6 .ixh6 18 .ixh6 0-0-0 19 c4 ltd? would have led to a position in which White has an edge, since Black has no really constructive plan.

16 .ie7? Black misses the opportunity he

has been given. 16 ... g6! would have left the h4-knight misplaced It is true that after 17 'l'e2 .ig7 18 f4 0-0 19 f5, follow� by 'l'e4, White has a good centralized position and Black's bishops are hemmed in. However, after 18 ... 0-0-0!? Black would have chances of puhing his two bishops to work. The b7-bishop functions as a kind of 'Dragon' bishop, in that it makes it hard for White's attack down the a-file to strilce home properly. Moreover, playing cxb5 will


leave d5 hanging. White is certainly not worse, but Black would have far more counterplay than he obtains in the game.

11 lbrs (DJ

The exchange of inaccuracies has favoured White, because Black's inaccuracy is actually quite serious -his bishop doesn't belong on e7 at all. Now White is clearly better.

17 ... J:d8 17 ... 0-0 is impossible because of

18 ..th6. 18 ..te3 g6 (D)

More or less forced, because otherwise Black lacks a constructive move. 18 ... 0-0? still fails to 19 ..th6 while after l 8 ... ..tc8 White has a choice of promising lines:

l ) l9l:ac l 0-0 20 lbxe7+(not20 cxb5 ..txf5 21 'i'xf5 axb5 22 J:xc7 J:xdS! and Black escapes) 20 ... 'i'xe7 21 cxb5 J:fe8 22 b6 cxb6 23 ..txb6 J:d7 24 'i'd3! ..tb7 25 d6 gives White a clear advantage.

2) 19 c5! (even more forcing) l 9 ... ..i.xf5 (19 ... 0-0 20 d6 wins a piece) 20 11Vxf5 :Xd5 21 'l'c8+and after 2l.. ...td8 2211fxa6 or 2I.. .l:d8 22 J:xd8+ ..i.xd8 23 l:ld l 'ife7 24 1!fxa6 White wins a pawn while re. taining a positional advantage.

19 lbh6! White could exchange on e7, but

then Black's king could castle or move ton and he would have fair de· fensive chances. White's qu.enside majority is an asset, but it will not win the game by itself because the c-pawn is tied to the defence of d5. Instead, I wanted to keep Black's king on e8, while I broke through on the c- and d-files.

19 'fig7 Intending ... f5, when the knighl

might be in trouble. 20 'l'g3?!

This move, introducing lbf5 ideas, is not so strong as I imagined during the game. Sometimes, when you

ANAND - AGDESTE/N, BAGU/0 Crrr 1987 19

have a good position, the temptation is just to play easy, comfortable moves and wait for the position to win itself. The correct move was 20 cxb5!, when 20 ... axb5 fails to 21 :a7. During the game I rejected it because of 20 ... f5 21 bxa6 .i.a8, with the threat of .. .f4. However, with ten years' hindsight I don't see Black's compensation for the two pawns. In particular, once White has played a7 the bishop on a8 will effectively be dead, since taking on d5 will always allow a combination involving the promotion of the a-pawn. One line is 22 'tlh3! f4 23 .i.d2 .i.g5 24ll:lg4 h5 25 �xe5 and wins.

It is worth noting that 20 h4 was playable. After 20 ... .ic8 we transpose into the game, but White has saved a couple of moves.

20 ... .i.c8 Threatening ... g5, so White's next

move is forced. 21 h4 .id6 (D)

If Black attempts to play actively by 2I...f5 22 .i.g5! f4 then:

I) 23 'i'c3 b4 24 'ilfd2 (24 '6'f3 .txg5 25 hxg5 'fle7 26 ll:lg4 '6'xg5 27 �xe5 0-0! is unclear) 24 ... .txg5 25 hxg5 'i'e7 26 .l:tel �f8! and Black has counterplay.

2) 23 •h2!. A paradoxical move which nails down Black's kingside. Now White simply threatens :el followed by doubling or even g3 (meeting .. . f3 by g4 ). However Black continues_ his e5-pawn will come

under fire and his king is trapped in the centre. The point of .wt-.2 is that after 23 ... .i.xg5 24 hxg5 '*le7 White can defend the g5-pawn and maintain his kingside bind.

w

22 '*if3 Other moves are inferior, for ex

ample 22 c5? .txc5 or 22 lLcl e4 23 .tf4 '*ixh6! 24 .txh6 .txg3 25 fxg3. In this latter line White keeps a slight edge as the opposite-coloured bishops (which normally have a drawish influence) make Black's defence a bit more difficult, when there are still a fair number of major pieces on the board. However, I thought that keeping the queens on offered even more.

The move 'lin is mainly to preempt the threat of ... e4. Now that White has got the knight firmly entrenched on h6, he can proceed with :aci followed by c5 and d6. The manoeuvre '6'f3-g3-f3 may appear odd, but Black has also wasted time

20 V!SHY ANAND: MY BEST GAMES OF CHESS

with the manoeuvres ... iLb7-c8 and ... iLf8-e7-d6.

22 iLe7 Black would like to play .. J:I.f8, but

the immediate 22 ... l:l.f8 allows 23 c5, because after 23 ... iLxc5 24 i.xc5 the rook on f8 is hanging. So Black returns to e7, still preventing c5 and getting ready to play ... l:l.f8 next move. The net effect of both players' oscillations has been to add the moves h4 and ... iLcS, which benefits White.

At this stage I felt that my position was very comfortable, but I had (and still have) a great deal of respect for Agdestein and I wasn't counting on victory yet.

23 l:l.acl Retaining the option of either cxb5

or c5 followed by d6. 23 ••• bxc4

Agdestein finally decides to re-move the cxb5 option.

24 bxc4 :t8 25 c5 f5 26 iLgS iLxgS 27 hxg5 'ii'e7 28 'l'g3! (D)

White had a more complex alternative in 28 d6 'ii'xg5 29 'l'c6+ i.d7 (29 ... l:l.d7 30 'ii'a8! l:l.d8 31 :tel should win for White) 30 'ii'xc7 'l'xh6 31 c6. This would also have been quite promising, as Black would have to return the piece for one pawn, still leaving White with a dangerous passed pawn. However, given

that an effective, solid alternative existed, I preferred to play safe.

In a way it is quite strange that White is not winning already, since Black's king is irrevocably since in the centre - not only call it not castle, but even f7 is denied to it. However, the reason is the knight on h6. It is of course doing a wonderful job, but when it comes to landing a killing blow in the centre, White is effectively a piece down.

28 f4 29 'ii'h4 e4 30 d6 'ii'eS!

The best chance. After 30 ... cxd6 31 cxd6 'ii'e6 (3 l ...'l'e5? 32 llxcS!) 32 l:l.c7 e3 (32 ... iLd7 33 �g4) White can continue 33 fxe3 fxe3 (after. 33 ... '1'xe3+ 34 �h l there is node· fence against 35 :te l) 34 'l'c4!l:l.f2 35 'ii'xe6+ iLxe6 36 l:l.e7+ �f8 37 l:l.xe6 with an extra piece.

31 dxc7 llxdl+ 32 :Xdl (D)

ANAND - AGDESI"EIN, BAGUIO CITY 1987 21

32 e3? Black decides to counterattack, but

now White's win is fairly straightforward. Black's only chance was 32 ... Wxc7, when we can see that it is not so easy for White because his knight is on h6. If it were on any normal square then, for example, lZ\c3 or �4. heading for d5 or d6, would be decisive. Nevertheless, after 33 ll:lg4 (heading for f6) 33 ... i.xg4 34 Wxg4 :tf5 35 b4! (White needs to pause for this move; it strengthens c5 and indirectly supports :td6; after

35 'l1Ve2lhc5 36 'liV:u4+ Wt7 White's attack has got a bit stuck) White has a large advantage. He intends 11'e2 and :td6, taking aim at e4 and a6, when Black's exposed king causes continuing problems.lf 35 ... e3, then 36 fxe3 fxe3 37 11Fe4+ picks up the e3-pawn (37 ... .:.es 3S 'IlVaS+ �7 39 'iiVhS is even worse).

33 l:td8+ 'lie7 34 ltlg8+ lhg8 35 lhg8! ..te6

Or 35 ... e2 36 11'xh7+ 'lie6 37 .l:l.eS+ and wins.

36 c811' 36 'l1Vxh7+ J..f7 37 .l:l.eS+ <lixeS 3S

cS11'+ �7 3911'b7+ <lids 40 11FhS+! is a prettier win, but I preferred the prosaic text.

36 ••• exfl+ 37 •xn!

This game has been published in some magazines with the move 37 <lixf2, but that allows mate in two! !

37 ..bc8 38 lhc8 1-0

This win over the top seed put me well on the way to becoming World Junior Champion. My run of wins continued with further victories over Klinger, Ivanchuk and Blatny, after which I was in clear first place, a position I held on to until the end of the tournament.

Winning the World Junior was my big breakthrough. Nonnally a player from India would have to waste a lot of time playing in mediocre open tournaments, gradually improving his Elo rating and hoping to get some invitations. However, the two factors of my World Junior title and my GM title, which I gained shortly after this event, enabled me to short-circuit the process. I got an invitation to a pleasant open tournament in Lugano (but after


my play there, I didn't get invited again!). Then I was asked to be a commentator at the World Cup event in Brussels (April1988). In Brussels I was able to meet some tournament organizers and this led to my first top-class invitation- to Wijk aan Zee 1989.

I had a tough year in 1988, just after gaining my GM title. I lost about 40 rating points in two tournaments at Biel and Blackpool- I still don't understand why, although many GMs have told me that they had similar experiences after gaining their title. I had a break after Biel and returned to active play near the end of the year at the Thessaloniki Olympiad, making 8lh points out of 12 games. This score was sufficient to regain 10 of the Elo

points I had lost, but later I was surprised to discover that FIDE had not given me any points at Thessaloniki. The reason, I found out, was that the Indiau team had arrived late, after the first round, and had therefore lost one match by default. FIDE had counted this as a normal loss, thereby wiping out my ten point gain from the rest of the event! Later on I had my ten points reinstated.

I then scored 6/9 at the GMA Open in Belgrade and subsequently I played at Reggio Emilia, another good invitation resulting from my World Junior success. This tournament started well with wins against Ivanchuk and Sax, but after that I lost some games and finished on 4/9. However, this event was very useful for me; it was the first time that I had played such a strong field.

My next event was the long-awaited trip to Wijk aan Zee. This was my first really big event and I was thrilled to be there. I had a zigzag course in the tournament. I won my first two games, just as at Reggio Emilia, then lost to Tseshkovsky (who used to be a nemesis of sorts!). lalso lost to Van der Wiel, but then won against Ivan Sokolov. I was still on '+I' when the following game was played in the penultimate round.

Game 4

V. Anand - J. Benjamin Wijk aan lee 1989

Sicil ian, Richter-Rauzer

1 e4 cS 2 lt:f3 d6 3 d4 cxd4 4 tt:lxd4 tt:lr6 s w tt:lc:6 6 �gS e6 7 'l'dl i.e7 8 0-0-0 0-0 9 tt:lb3

Two of the first four games in the book reach this position! I have done pretty well with this line so far.

9 •.• 'il'b6 By the time this game was played

it had become clear that 9 ... a5 wasn't that great a move.

10 f3 IUS (D)

11 <!Jb1

There is an interesting story relating to this move. I used to get a lot of my theoretical infonnation from Ivanchuk. For example, during the 1985 World Junior Championship in Sharjah he had shown me a tremendous idea in the Dragon, which turned an existing evaluation upside down; I was amazed (and thankful!) that he was so open and generous about showing his ideas to me. I would try to give him some titbit in return, but unfortunately my novelties were not that good!

At the Reggio Emilia event the month before Wijk aan Zee, there hadn't been much to do in the evenings. so one day I went to Chucky's room. By this time he had recovered from his first round loss to me and was on a respectable score. He was also feeling bored; we went out for a walk and then returned to his room. I asked him "Why does everybody play the Catalan - it seems such a boring opening." He replied that it was not boring at all and proceeded to show me an interesting idea. The introductory moves were 1 d4 lilf6 2 c4 e6 3 tilf3 d5 4 g3 i.e7 5 .i.g2 0-0 6 �3 dxc4 7 lL!e5 c5 8 dxc5 and


now Vassily said that 8 . .. 'ifc7 was supposed to be a good move. He explained that after 9 �xc4 'l'xc5 ... (D)

... everybody plays 10 'l'b3, but he had found a very strong idea: 10 b3!. The point is that after I O ... :ds White can play 1 1 .llLa3 :Xdl+ 1 2 J:l.xd1 'ilt'c7 13 �b5 trapping the queen. I was very impressed by this line, which I had never seen before. He explained that this was the reason why ... 'ilt'c7 was not good in this particular line of the Catalan, but was good in the similar lines in which Black plays ... .llLb4 and later retreats the bishop to e7 in response to a3. The reason, of course, is that the pawn on a3 prevents the move i.a3. A nice idea, but apparently not much use to me as I didn't play the Catalan. I hope Chucky will forgive me for revealing this piece of analysis!

Returning to the Benjamin game, at this time I didn't really study the openings too deeply. I looked at all

the theory, but didn't really go beyond that. After lO ... :ds I suddenly didn't feel very happy with my posi. tion and couldn't fmd a continuation I felt comfortable with. Then I suddenly brightened up, because I saw a little trick after 11 q;,b I.

11 ••• ciS?! (D) If Black plays ll ...a6, then 12

.i.e3 Wfc7 13 'ilf2 and White gains time owing to the threat of i.b6, so q;,b1 is really to provoke ... a6. Later on it was discovered that this doesn't really matter, because Black's counterplay consists of ... a6, ... lOfl and ... b5 in any case, but we didn't know that in 1989!

After 11 Wbl Benjamin looked surprised, bKause he didn't understand the point of the move. Perhaps he was thinking 'Vishy doesn't know that Black's threat is ... d5'. , when he played ... d5 I checked my idea carefully, although there isn't much choice because after anything else White is clearly worse.

ANAND - BENJAMIN, WIJK AAN ZEE 1989 25

12 .hr6 dxe4? Benjamin falls for it hook, line

and sinker. He could still have bailed out by 12 ... i.xf6 13 exd5 i.xc3 14 •xc3 exd5, although after 15 'il'c5 or 15 i.d3 White has some positional advantage. However, as he admitted after the game, he simply hadn't seen the idea at all.

13 i.xe7 Not 13 i.d4 .fud4 14 �d4 e5.

13 •.• :Xd2 14 lilxd2! (D)

14 exf3 After 14...lilxe7? White wins by

15 lllc4 'ilc7 16 lllb5 and Black's queen is lost almost exactly as in Ivanhuk's Catalan idea. It isn't trapped here as it was in the Catalan, but the threat of mate on d8 means that it amounts to the same thing. At the next tournament where I met Ivanchuk, he came up to me and said "I see you used my idea in the Catalan!".

After the game Benjamin said that he wanted to resign at this point, but decided to play a few more moves; however, I then staned to play so badly that he couldn't bring himself to resign any more.

15 gill? Here's the bad move. Later I real

ized that I5lilc4! would have won on the spot. 15 ... '1ff2 loses to I6lile4, and after 15 ... '1fc7 16 �d6 fxg2 17 �xg2'1Fd818�g3 '1Fe719l:l.hel all White's pieces are active and his knights are going on a queenside rampage (lilb5 or llld5, coupled with ltld6). The fact that White has a lost a pawn is irrelevant.

15 ••• e5 16 �h4?! (D)

Another mistake. 16 �a3 would have been much better, keeping control of key squares such as c5 and d6. In this case White should still win in the long run, although thanks to White's previous error it is likely to be a laborious process.


16 ... .te6 17 lMe4

White already has to take care: not 17 .i.c4? .i.xc4 1 8ltJxc4 'iWb4.

17 lM4 (D)

18 .i.g2?! After this, the position becomes

rather murky, although it is possible that White can still retain an advantage by accurate play. 18 .i.f2 was a much safer way to prove that White is better.

18 ••• .l:c8 Suddenly I couldn't find a line

that worked for White. The threat is . .. f5 followed by ... .l:xc3, and it isn't easy to find a good defence. Finally I found a line that seemed to work for White but I was very nervous because Black has all his pieces aimed at White's queenside.

19 .i.f2 fS! 20 f4!

Not 20 ltJg5? .l:xc3 21 ltJxe6 1Wxe6 22 .i.xd4 (22 bxc3? 1Wb6 + and

Black wins) 22 ... exd4 23 bxc3 dxc3 24 q;>a1 11te2! (24 ... 'itb625 l:lbl 'iff2 26 .i.fl 11txc2 27 .i.c4+ followed by :Xb7 wins for White) 25 .l:hgt11txc2

26.1:bl1ltd2 27 a3 and Black is even slightly better.

B

20 fxe4 21 fxeS (D)

21 ••• :C4 2 l ....l:d8! would have made life

much harder for White: 1) 22 ltJe2? .i.xa2+ 23 �xa2

'li'a6+ wins for Black. 2) 22 .l:he1 ?! .i.g4! 23 .l:d2 '1116!

24 .i.e3 (24 .l:xd4 .l:xd4 25 .i.xd4 'Wd2 favours Black) 24 ... Wh4 25 .i.f2 with a draw by repetition.

3) 22 .l:d2! .i.f5 (not 22 ... ltJb3? 23 axb3 and White wins) 23 .i.e3! 'ilrg6 24 :hd 1 and, surprisingly, there is little Black can do to prevention followed by taking on d4.

While 2I. ...:I.d8 might not have been any better than the text-roove against perfect play, 22l:td2! and 23

ANAND - BENJAMIN, WIJK AAN ZEE 1989 27

.A.e3! aren't easy moves to find over lhe board.

22 l:tbel? (D) Missing 22 &2! l:l.a4 (Black's

sacrificial attempts fail, for example 22 ... 'l'xb2+ 23 �xb2 l:txc2+ 24 �al �xe2 25 i.el or 22 ... l:l.xc2 23 bd4 'il'c6 24 �4 and White wins in both cases) 23 b3! l:txa2 (if 23 ... .1xb3, then simply 24 axb3) 24 .1xd4 and wins.

B

22 ••• .11b4? Benjamin was in time-trouble by

now, and commits another mistake. The best line was 22 ... .ig4! 23 .!:d2 ( not 23 l:txd4? l:txd4 24lt:lxe4 l:txe4 25 .1xb6 l:txel#) 23...'fih6! 24lt:lxe4 (or 24 :Xd4 l:txd4 25 i.xd4 'ilt'd2) 24 ... �6 and Black has avoided losing a piece. Despite Black's slight material advantage, I don't think White is orse, since B ack' k fairly exposed and his pieces disjointed.

23 '.tel!

It would also have been good to play 23 b3!, which looks a bit paradoxical as Black has so many pieces ready to sacrifice on b3. However, after 23 ... .i.xb3 24 axb3 l:txb3+ 25 <t>c l White should win.

23 ... i.g4 If 23 ... lt:lb3+ then 24 axb3 1Wxf2

25 lld8+ �f7 26 :n picks up the queen.

24 .!bdS 'ilt'cS 25 lt:lxb4 .i.xdl 26 �xdl! (D)

Not 26 l:l.xdl? �e2+ and Black wins, nor 26 l:txe4lt'lb3+! 27 <t>xdl 'ilt'xf2 28 axb3 'ilt'xg2 29 l:te2 and only Black can be better.

After the text-move it suddenly dawned on me that I was completely winning. I had been struggling to contain Black's queenside initiative for so many moves that the realization caught me by surprise!

26 e3 26 ... 'ili'xb41oses to 27 l:txe4.


r1 lhe3 lMS 28 �dS+

An important intermezzo. 28 .l:tf3 is a mistake because of 28 .. .'i!fxb4 29 :Xf5 W'g4+.

28 �

Forced; 28 ... Wh8 fails to 29 l:l.c3 29 .:tf3 'il'xb4 30 .:txfS+ WeB 31 e6 1-0

After 31...11Vxb2 32 �c5 it is all over.

Before the last round I was in a tie for fiCSt place with Nikolic, Ribli and Sax. I had the feeling that a draw in the last round would be enough to main tain the status quo, and indeed when my game with Douven ended in a draw my feeling turned out to be justified, as the other leaders also drew.Thus we ended up in a four-way tie for first. It was my first success in one of the world's major international events.

In 1989, Bessel Kok, the chairman of SWIFf and a great chess benefactor, organized a Youth vs Veterans event in Cannes not long after Wijk aan Zee.lt was held in conjunction with the annual Cannes Games Festival and was a very enjoyable event. The veterans were Tal, Spassky, Andersson, Csom and Larsen; the youth team consisted of Lautier, Renet, Adams, Miralles and myself. As the only GM in the youth team, I was acconunodated in an excellent hotel; indeed, the conditions at this event were the best I had experienced un til then.

Game 5

M. Tal - V. Anand Youth vs Veterans, Cannes 1 989

English

I met Misha Tal for the first time when I visited the World Cup tournament in Brussels in 1988. He was the most popular player by far, captivating everyone with his personality and his brilliant chess. This was my first game against my childhood hero and I was obviously quite excited.

1 c4 cS 2 ltl£3 lt:\c6 3 00 lt:\d4

How do you explain a move that violates the rule not to move the same piece twice in the opening? Well, I can't really find a general principle that justifies it, but it does seem to work! One possible explanation isthat when White recaptures on f3 with his queen, he has lost control of the important d4-square. Then Black can bring out his other knight via h6 and f5 to fight for d4.

In any case, 3 ... �4 is a provocative move which leads to a more unbalanced type of position than is usual in the Symmetrical English.

4 e3 lt:\xf3+ S 'i'xf3 g6 (D) 6 b3

Or 6 d4 .i.g7 7 dxc5 (after 7 'Wd 1 �f6 White will have to play d5 and then lose a tempo with e3-e4 in order to get his customary space advantage) and now:

1) 7 ... 'i'a5 8 e4! 'i'xc5 9 lt:\d5! (stronger than 9 .i.d3 .i.xc3+ 10 bxc3 d6) with a slight advantage for White.

2) 7 ... .i.xc3 +!? 8 bxc3 11fa5 9 e4 11fxc5 10 .i.d3 d6. Normally Black should not give up his bishop like this in an 'Indian structure', but due to White's doubled c-pawns this position should be compared with the Nimzo-Indian rather than the King's Indian Defence. Admittedly Black has already played ... g6, but it's not clear how White can exploit this.


6 7 .i.b2 8 g3 9 .i.g2

Two years later, I beat Karpov with 9 . . . llli'6; the tactical justification is that 10 liJd5 .i.g4! equalizes immediately. Actually, it is much more natural than 9 ... ltJh6. Whatever can be said about the objective merits of 3 ... liJd4, I scored 212 against Tal and Karpov with it!

10 'ild1 0-0 1 1 0-0 .i.d7

To support the ... b5 advance. 12 a4

After 12 d4?! Black's plan of exerting pressure·on d4 comes to fruition: 12 ... cxd4 13 exd4 llli'5 ! 14 d5 (14 liJd5 b5! gives Black the edge) 14 ... b5 ! with good counterplay.

B

12 .i.c6 13 d4 .i.xg2 14 �g2 (D)

14 l:l.c8

After 14 ... cxd4 15 exd4 ltJf5 16 d5 a6 17 l:l.bl ! the position is slightly better for White. He has a space advantage and a weak black e- pawn to play against On the other hand, his pieces on the queenside are awkwardly placlld.

1 5 'ild3?! Tal later mentioned the possibility

of 1 5 d5 ! and this move does seem to offer White a stable if minuscule edge. Black will hardly be able to achieve ... b5 and playing ... c7-e6 would create weak pawns on e6 and d6.

15 - ad4 16 exd4 lllrs

Now Black is fine. 17 d5 (D)

After 17 lDe2 d5 18 c5 a5 Black is slightly better. White's queenside pawns are crippled and the pawn on d4 is weak. White's best line was probably 17 �d5! e6 18 ltle3, just playing for the exchange of knights and equality.

B

TAL - ANAND, CANNES 1989

17 1fb6 25 c5

31

The reason why 'ifd3 was bad - There is no other defence against b3 is undefended. 25 ... ll:ld4.

18 t0d1 i.xb2 25 ••• 'ifxcS 19 <fub2 eS! 26 'ffxb7+ .l:.c7

Now Black is justified in playing 27 'lfdS 'ffb4 actively. The open f-file counts for Not 27 ... '1Vxd5+'? 28 l:txd5 l:tc2 more than Black's slightly weak cen- and White can hang on with 29 ll:lc4. tral pawns. 28 lUd1 .l:.cS

20 dxe6 Exploiting the weakness of b3. Forced. If White doesn't do this, 28 ... ll:ld4 is less clear after 29 ll:lc4

Black can aim for ... e4-e3. Moreover, l:tc5 (not 29 ... ll:lxb3'? 30 ll:lxd6!) 30 his knight can settle comfortably on 1Wa8. d4. 29 'IVai

20 ••• fxe6 If 29 ll:ld3'? then 29 ... ll:le3+! 30 21 :Sd1 ltt6 fxe3 11Vxd2+ winning the exchange.

Preparing to double rooks on the 29 ... 'lixb3 f-lile. 30 ll:ld3 (D)

22 :d2 eS! (D) After 30 '1fxa7+ l:tf7 31 11t'a6 (3 1

23 11Fd5+ After 23 ll:ld I ltld4 Black wins a

pawn as White has to meet the threat of 24 .. .'i'c6+.

23 ... 24 'ltbs

'lfa8 l:tc2 is similar, e.g. 32 ll:ld3 'lic3 33 l:txc2 1Vxc2 winning material) 31 . ..l:tc2! 32 ll:ld3 l:txd2 33 l:txd2 'lid5+ 34 �gl e4 White loses the pinned knight.

30 l:tc2?!


30 ... l:l.c7! would have been brutally effective - Black simply maintains his a-pawn and threatens both 3 l ... 'lWxa4 and 3 l .. . �d4.

31 'lWe4?! White could have made life harder

by playing 3 1 'lWxa7+ .l:l.f7 32 'lWa5 (D) (32 'liaS 11fc3 ! wins). I don't know if Tal spent much time considering this apparently greedy continuation. I must admit that it was only after some time checking the position with Fritz that I began to realize that Black's task was not so easy. Still, Black does have a way to win:

1) 32 ... e4'! ! and now: I a) 33 .l:l.xc2 'i'xc2 34 .l:l.d2 'ifc6

35 �b4 'ilfc3! wins for Black. lb) 33 �f4 'iff3+ (33 ... l:l.c5 34

1Wd8 is unclear) 34 �gl e3 looks strong, but after 35 fxe3 'ifxe3+ 36 �hI there is nothing clear-cut, for example 36 ... 'iff3+ 37 �gl �d4 38 .l:l.el .

lc)33.1Wb4! 1Wxb4 34 &b4 saves White.

2) 32 ... 1i'h7+?! 33 �gl lbd2 34 1Wxd2 lDd4 35 �1 and White avoids disaster.

3) 32 ... l:rxd2 33 'lWxd2 (33 l:txd2? 1i'b7+ 34 �gl 'lWbl+ 35 �g2 ta3+ wins) 33 ... 1Wxa4 wins a pawn, but there is still a long way to go.

4) 32...�3+! 33 fxe3 Vxd3 34 .l:l.xc2 ile4+ 35 �gl (35 �h3 :t'5! and there is no way to stop ... l:l.hS#) 35 ... 11'xe3+ 36 �g2 'lWf3+ '57 Wh3 1rxdl and White's king is hopelessly exposed.

31 _ .l:l.c4 32 'lidS 1Wd

Besides the extra pawn, the difference in strength of the knights can be seen.

33 .l:l.b2 :.14 34 .l:l.b7+ wh6 35 'lrbS (D)

35 'i6'gS loses to 35 ... 'i6'c6+.

35

TAL - ANAND, CANNES 1989 33

36 <i>gl

Or 36 fxe3 'l'c2+ 37 �h3 :h4+! 38gxh4:f3+ 39 �g4 'l'g2#.

36 -· 'fic2

37 lin llxd3 0-1

After 38 1i"d7 Black mates by 38 ... 1Fxf2+!.

I was very happy to win in Tal's own style, although I recognized that his poor play in this game was due to illness. In fact, he withdrew from the tournament after the first half for medical treatment. Still, I am proud to have played at least one game against Misha Tal.

Game 6

V. Anand - B. Spassky Youth vs Veterans, Cannes 1 989

Ruy Lopez, Breyer

1 e4 eS During the course of his career,

Spassky has played just about every opening there is, but in recent years he has tended to stick to dual king pawn openings, so this was not a sur-prise.

2 ttlf3 ttlc6 3 �bS a6 4 �a4 ttlf6 S 0.0 �e7 6 l:e1 bS 7 �b3 d6 8 c3 0-0 9 h3 ttlb8

The Breyer Defence, for decades Spassky's main weapon against the Ruy Lopez.

10 d4 ttlbd7 11 c4

Round about this time there had been a modest revival in this old move. I felt that Spassky's knowledge of the main lines would be much greater than mine, so I tried, with some difficulty, to find a relatively unexplored continuation. 1 1 c4 enjoyed popularity in the 1960s and was probably no surprise, but at least it didn't have such a huge body of theory as 1 1 ttlbd2.

11 ••• c6 12 a3

This was an idea I came up with over the board At the time I had no idea if it had ever been played be fore, but I have since found a couple of earlier games. The immediate 12 ttlc3 is met by 1 2 ... b4.

12 ..• bxc4! The best reply. After 1 2 ... b7 ll

ttlc3 I think White is slightly better, since White's structure is a bit more comfortable in this type of 'Old In· dian' position. Black can of course continue with .. ."Wic7, ... l:ac8 and so on, but White can gain more space with an eventual �a2 and b4.

13 hc4 (D)

8

ANAND - SPASSKY, CANNES 1989 35

13 ••• dS Black is not worse after the text

move. but 13 ... li:lxe4 would have equalized straight away:

I) 14 dxe5? d5 15 .i.xd5 cxd5 16 1l'xd5 l:lb8 (16 . . . li:lxf2 may also be good for Black) 17 'ilfxe4 li:lc5 18 'lre2 ltlb3 19 l:la2 .i.e6 and White loses the exchange by force.

2) 14 :.Xe4 d5 with two possibilities:

2a) 15 ltlxe5 (Spassky suggested this, but Black has an escape route) 15 ... dxe4 (not 15 ... li:lxe5 16 .i.xd5 cxd5 17 :XeS .i.d6 18 :eJ and White can be happy because, compared to the Marshall Attack, his piece development is easier as ltlc3 is possible) 16 ltlxc6 'i'e8 17 .i.d5 .i.d6 18 ltle7+ 'i'xe7 19 .ba8 li:lb6 20 .i.c6 j,c7 (the exposed bishop on c6 gives Black's queen a free tempo on its way to h2) 2I ltlc3 'i'd6! (21 ...f5 22 'i'b3+ �h8 23 'i'b4 is annoying for Black) 22 he4 f5 23 .i.f3 'i'h2+ 24 �fl l:le8 25 g3 f4 and White is certainly not playing for the advantage - indeed, Black is probably slightly better.

2b) 15 l:lel dxc4 16 'ilfe2 (after 16 dxe5 ltlc5 the outposts at d3 and b3 give Black the edge) I6 ... :es with equality, as Black's weak pawns are balanced by his active pieces. In fact this had been played as long ago as 1971, in the game HUbner-Lengyel from Wijkaan Zee.

14 exdS

14 dxe5? ltlxe4 transposes to line 1 of the previous note.

14 cxdS 15 .i.a2 e4 16 .!Des .i.b7 (DJ

Here I was quite happy, as it seemed to me that my pieces were well placed to exen pressure on d5. While there is some truth in this, Black has so many pieces available to defend d5 that the inconvenience is not serious.

17 lLlcl li:lb6 By protecting d5, Black sets up

various threats based on moving the f6-knight. The first is to expel the e5-knight by . .. IDeS followed by ... f6, in which case he would probably be better. The second is the simple ... li:lfd7, when Black can exchange on e5 without having a knight attacked after dxe5. Hence White must react quickly.

18 f3! l:lc8 19 .i.b3


White could have maintained some slight pressure by 1 9 .ig5, as Spassky pointed out after the game. 19 . . . .ixa3? 20 bxa3 l:l.xc3 is bad after 21 .id2 ! l:l.c8 (2 1 ....:Xa3 22 .ib4 favours White) 22 .ia5 ! with a very awkward pin.

19 --· .ia8 20 .igS l:l.c7

Black's position is basically safe; for example he could have played 20 ... exf3 2 1 'i'xf3 �c4 with equality, and indeed this would have been the natural way to make use of . . . .ia8, which protected the bishop.

21 .:tel li:Jr d7 22 .if4 .igS (D)

23 .ixgS After 23 �xd7 the line 23 ... .ixf4

24 �xb6 .ixcl 25 �xa8 .ixb2 (25 ... l:l.xc3 26 bxc3 is also very good for White) 26 �xc7 .ixc3 27 �xd5 .ixel 28 'i'xel exf3 29 gxf3 clearly favours White. However, Black can improve by 23 ... l:l.xd7 ! 24 .ixg5

1i'xg5 25 fxe4 dxe4 26 �xe4 'lg6 and he regains the pawn since 27 .ic2 fails to 27 ... f5.

23 --24 fxe4

'ifxgS dxe4

Black had a good alternative in 24 ... �xe5 25 dxe5 dxe4 and now

I) 26 2d4 �d7! (not 26 ... e3 27 l:l.e2 attacking b6 and e3) 27 .6 (not 27 �xe4?? .ixe4! and Black wins) 27 ... �e5 28 exf7+ �h8! 29 'ifd6 l:l.cc8 and Black has dangerous king side threats.

2) 26 e6. During the game, I believed this was good for White, but Black simply continues 26 ... 'lc5+ 27 �hI fxe6 28 .ixe6+ Wh8 and the position is just unclear.

The game continuation is also roughly equal.

25 'ifg4 'ifxg4 Black cannot get away with

25 ... 'ifd2, when White must decide how to take on f7:

I) 26 .ixf7+ (obvious, but this is in fact an error) and now:

I a) 26 ... .:.Xf7 2 7 �xf7 li:i6 28 'i'e6 'ifxd4+ (28 .. Jbf7 29 'ifxb6 e3 30 �e4! wins) 29 �hi J:xf7 30 l:l.cdl and White is winning.

l b) 26...�h8! 27 �xd7 (not 27 .:ted I ? ttlxe5 28 dxe5 'ifxb2 and Black wins) 27 ... '1Vxd4+ 28 �hi �xd7 with an unclear position.

2) 26 �xf7! �f6 (26 .. .'i'xd4+ 27 Wh 1 �c5 28 J:cd I 'li'f6 29 J:fl ! wins, while after 26 ... 'll'xb2 27 'le6! Black is in considerable difficulties)

ANAND - SPASSKY. CANNES 1989 17

27 �g5+ 'it>h8 (27. ..�d5 28 'I'd! ! With a clear extta pawn) 28 'iV g3 and White is clearly better.

26 lllil:g4 g6? Black's first step downhill. After

26 . .. �h8 27 it)e3 f5 Black activates his kingside majority, which should provide enough counterplay to maintain the balance.

r1 �2! Wllite now wins a pawn, although

in view of the reduced material this does not necessarily guarantee winning the game.

r1 ••.

28 d5 29 it)fxe4

:es �g7 lllil:dS (D)

At first I couldn't believe this move; it looks as though White must win material after 30 .txd5 .txd5 31 liXI6, as Black will end up being threatened with �xd5 and lt:le8+. However, Spassky had worked out a defence. After I had calmed down, I did'nt see any way to force a decisive

material gain, so I just went for a pawn.

30 it)d6 30 .i.xd5 .i.xd5 31 �d6 is an

swered by 3 J ...:Xel+ 32 :Xel :cs, meeting both of White's threats. It is easy to miss that the c3-knight is no longer protected twice and therefore b2-b4 is impossible. White can try 33 :e7 �e6 34 �f7 <M6 35 :Xd7 lhc3 ! 36 bxc3 .i.xd7 37 it)d6 �e5, but although he has won a pawn, Black's king becomes too active and he draws easily.

30 31 lbel 32 :e7 33 :xt7+ 34 �c4

J:x:el+ �f6 :c6 �h6 :e6

Black has managed to get some counterplay; White's f7-rook is surrounded and Black can activate his king via g5.

35 'M2 36 �2 37 �dl

�,s .tel» liS (D)


38 �b3?! This would have been the moment

to settle for 38 l:tg7! �e5 (38...�e8 39 �de4+ ! Wh6 40 l:th7+! Wxh7 41 �g5+ leaves White a clear pawn up) 39 J:l.a7 and White's rook gets out of the trap! In fact this is quite awkward for Black as a6 is weak. Black's mass of pieces appears menacing, but White has all the critical squares controlled. The sacrifice 39 ... �eg4+ doesn't work here: 40 hxg4 �xg4+ 4 1 Wgl (but not 4 1 Wg3? h4+ ! 42 Wh3 l:te3+ 43 �f3+ .i.xf3 44 gxf3 l:txf3+ and Black wins) 4 1 . ..l:tel+ 42 �fl defends.

I thought that the text-move was even stronger, but I hadn't really seen Black's 39th move - in fact, I'm quite lucky that it doesn't cause serious damage!

38 39 l:ta7

llJes �eg4+! (D)

This works because White can't play l2lfl any more.

40 hxg4 41 � 42 wgl 43 l:txa6 44 l:txe6?

�g4+ � �c2 ..td7

Though this sets up the nice finish to follow, it is a mistake. Two knights can't win by themselves and there aren't that many pawn left ! White should have played 44 l:ta5+!, followed by J:l.d5-d2. By keeping the rooks on, White has much better chances of pushing the queenside pawns and in fact he still has excellent winning prospects.

44 ·- ..txe6 (D)

45 llJcS ..tc4 46 a4 Wf4?

One reason why I exchanged rooks earlier was the combinnation which now occurs, but in the interim I had realized that 46 .. .'�•f5 prevents it: 47 a5 l2lb4 48 b3 iin 49 li:ld3 l2lxd3 (a forced move, as 49 ... li:la6 50 b4 iic4 5 1 l2lb2 is good for

ANAND - SPASSKY, CANNES 1989 39

White) 50 a6 .ie8 5 1 itld5 �e6 52 lOc 7 + ci.>d6 53 itlxe8+ �c6 and the king catches the a-pawn. Therefore White has to settle for 47 'iii>f2, but Black has good drawing chances.

47 a5 Now we get to see an elegant fin

ish - a lone knight dominating two minor pieces !

� ••• itlb4 48 b3 .if7 49 1Dd3+!!

Spassky had only seen 49 a6?? w6 50 itlxa6 .ixb3 and White has no winning chances since he has only one pawn left.

49 50 a6

itlxd3 £e8

51 itldS+ 1-0 After 5 l ...'iii>e5 52 itle7 (D) the

position deserves a diagram:

Black can't stop the pawn - a nice bit of domination!

In Cannes I made the best score in the junior team, 61/2110, but the 'Senior' Andersson made the best score overall, with 7112 points.

The next game is again from Wijk aan Zee, the year after my success in the 1989 event.

Game 7

M . Kuijf - V. Anand Wijk aan lee 1 990

Ponziani Opening

1 e4 e5 Round about a month before I had

beaten Kuijf in the tournament at Groningen. That game had been a Closed Sicilian, and at one stage it could have been very dangerous for me. This time I didn't want the same 'excitement' so I decided to play more solidly. Hence my choice of first move.

2 ll:lf3 ll:lc6 3 c3

A real surprise. 3 ••• lllf6

I played this move instantly and while he was thinking about his reply I had to spend a few minutes trying to remember the name of the opening!

4 d4 S dS

I don't really understand what Kuijf was aiming for with his choice of opening. It is harmless and only useful if White is aiming for a draw.

6 lOxeS ll:lg6 7 i.d3

However, this indicates that White is not aiming for a draw, towards which he could have made substantial progress by 7 'iWe2 'fle7 8 'flxe4

(8 ll:lxg6?! hxg6 is a little better for Black) 8 ... 'iWxe5 9 ll:ld2 (or alternatively 9 1i'xe5+ ).

7 ... ll:lxeS 7 ... ll:lxf2? isn't even a difficult

!rap: White wins by 8 �xg6 li:lxd1 9 i..xf7+ �e7 10 i..g5+ �d6 1 1 1iJc4+ �c5 1 2 i.xd8.

8 i..xe4 i..cS 9 1lfhs d6

10 i..gS? (D) 10 h3 was a much better move, aJ.

though even in this case Black can play for an advantage.

10 i.g4! Black can play 10 .. .'i'd7 1 1 0-0

'i'g4, with boring equality, but I had seen that the text-move is much

M. KUIJF - ANAND, WIJK AAN ZEE 1990 41

stronger. Technically, it may be a aovelty, but I am reluctant to call it that To my mind, novelties should be at least a little bit difficult to find. If you play the most obvious move and then discover that by an accident of history nobody has played it before, I am not sure that it deserves any special appellation.

11 111'h4? Losing on the spot. The lines l l

11fxh7 111'xg7 and ll 111'xg4 lLlxg4 12 ..ixd8 llixf2 l3 l:l.fl lLlxe4 14 ..ixc7 lc8 15 .ia5 b6 16 ..ib4 ..ixb4 17 cxb4 lc2 also offer White no hope. This leaves l l .ixd8 ..ixh5 1 2 ..ig5 (12 ..ixc7 l:l.c8 l3 ..ia5 b6 14 b4 ..ixf2+ 15 �xf2 bxa5 and 12 ..ih4 0-0 followed by l3 ... l:l.ae8 are also very promising for Black) as the only realistic way for White to play on. Even here Black has a range of tempting options. He could simply play for the two bishops by l2 ... ..ig6 13 llid2 llid3+, but 1Lf6 is probably stronger. Then 14 ..i£4 is impossible, 14 ..ih4 leaves the bishop shut out on the kingside and 14 ..ie3 allows Black to shatter White's pawns. Black could also consider l2 .. .f6, wins ith sim Jar id

11 ... f6 Now White has no reasonable

continuation. 12 ..ict

After 12 ..id2 111'e7 l 3 0-0 g5 White has the unpleasant choice between:

1 ) 14 11'h6 00 15 1fg7 1l'xe4 16 l:l.el ..ie2 1 7 b4 ..ib6 1 8 ..ie3 <:Je7 19 l:l.xe2 (l9lDd2 1l'd3) l 9 ... l:l.ag8 wins.

2) 1 4 1l'g3 f5 15 h3 (the only chance, or else .. .f4 traps the queen) 15 ... f4 16 'llfh2 ..id7 and although material is even. White is playing a whole qu.en down for all practical purposes. After ... 0-0-0 and a subsequent ... g4 the attack should overwhelm White.

If White retreats his bishop to e3 the lines are even simpler: 12 ..ie3 g5 l 3 11fh6 (l3 1l'g3 f5 14f4..ixe3 15 fxe5 ..if4 and Black wins) l3 ... 11'e7 14 0-0 llif7 15 1l'g7 0-0-0 16 ..ixc5

l:l.dg8 17 l:l.el ..id7 and White loses his queen.

12 'fle7 (D)

13 0-0 'This is forced to meet the threat of

. .. ll:ld3+, because l3 f3 is met by l 3 ... ..ixf3.

13 gS 14 'iWgJ rs


The threat is .. .f4, so White re- 17 � J:tg8 sorts to desperation. 18 l:!.ael h6

1S .hi'S 19 :XeS+ dxeS 15 h3 f4 1 6 'irh2 i.d7 shuts the 20 'If xeS+ �d7

queen completely out of play and 21 i.e3 l:1ae8 Black wins by .. . 0-0-0 and a pawn 22 'i!Vr4 11Vxd5 push on the kingside. 23 lbr3 i.d6

1S .hi'S 24 'i!Fa4+ bS 16 i.xgS 'i!Ff7 0-1

This game was finished while some of the others were still in the opening; I spent ten minutes and my opponent a little under half an hour.

It was a nice miniature to play in the first round, but the rest of the tournament didn't go as well as the year before. I lost in the second round to Nunn and continued unevenly throughout the event, finishing on 50%.

Shortly after Wijk aan Zee, I played an open tournament in Rome. This started well, but I lost a miniature to Miles in the penultimate round and f inished with 6/9. Then I went back to India for some rest before travelling to the Zonal tournament in Qatar. I was by far the highest rated player, but even so I was happy to win the tournament convincingly. As a result of this and some other tournaments I gained quite a lot of rating points, and on the 1st July 1 990 list I stood at 2610- I had broken through the 2600 banier. I took part in an open tournament in Manila, went back to India and then returned to Manila for the Interzonal. Based on my new rating I was certainly a potential qualifier, but in an Interzonal you cannot take anything for granted - an Interzonal is a tough tournament even for the top seeds.

My results in the Interzonal followed a fluctuating course: I won in the second round, lost in the third, won in the fourth and lost in the fifth. After a draw with Chandler in the sixth round, I faced the Icelandic grandmaster Margeir Petursson in round 7.

Game S

M . Petursson - V. Anand Manila Interzonal 1990

Queen's Pawn

1 d4 d6 2 c4 eS

This was part of my usual opening repertoire at the time. I liked l ...d6 because it is such a complete system in itself. If you play the Pirc then you have no reason to fear 2 e4, and 2 lLlf3 can be met by 2 ... ..1lg4. Since then, however, White has found ways to keep some pressure. As a result ! lost faith in the system and had to learn a decent defence to I d4.

3 lLlc3 exd4 4 'l'xd4 ll:lf6

An important finesse. If Black plays � • . e:lc6, then 5 'l'e3+ is a little annoying as 5 ... il..e7 6 ll:\d5 snares the two bishops.

5 g3 lLlc6 6 'ild2 g6

White's pieces are not badly placed - for example, the queen is quite useful on d2 if White intends to play b3 and il..b2. On the other hand, Black has gained a tempo because of 'ilxd4 and 'ild2.

7 il..g2 8 lLlh3 9 lLlf4

il..g7 0-0

White is aiming to establish a grip on d5. He has played the opemng

accurately, for example by not playing b3 too early, which sometimes allows a tactical ... d5! by Black.

w

9 a5 (D)

This is designed to prevent White from easily developing his queen's bishop, for if 10 b3, then I O ... a4 I I llbl (11 lLlxa4 lLle4 wins the exchange, while II ..llb2 may be met by l l ...a3 1 2 i.e! i.f5 or l l ...ll:\a5) 1l ...axb3 12 axb3 lle8, followed by . .. i.f5, with active play.

10 0-0 l:le8?! A slight inaccuracy. IO ... a4 was

better, simply preventing b3. 11 l:lel?!

White misses his chance. 11 e4 was correct, and after l l . ..a4 1 2 ll b l


�5 l 3 b3 White has secured his queenside pawn structure. Of course the immediate 1 1 l:l:b 1 is answered by 1 l ....if5.

11 •.• a4 12 l:.b1 llld7!

A good move. By the time White has played b3, Black has arranged to occupy active squares with his knights.

13 b3 axb3 14 axb3 (D)

14 ... �eS! If \4 ... lt'Kie5!? (intending to play

lS ... llld4) then 15 lllbS! is slightly better for White. It is the c6-knight which belongs on eS, so that the other one can occupy cS.

15 'l'c2 lllcS 16 b4?

Premature, because with energetic play Black now gets the better position. After the correct 16 .ib2 c6 we reach a position typical of the Fianchetto King's Indian. White has

the centre, whilst Black has open lines for his rooks and active squares for his knights. Chances are about equal.

16 ... .irS! An important intermezzo.

17 e4 After 17 lbe4 llla4, followed by

l 8 ... ltlc6, Black has excellent control over the long dark-square diagonal.

17 ... lt:le6! (D) Margeir had missed this tactic.

18 lllxe6 Black's play is tactically justified

by the variation 18 exf5 llld4 (·at tacking the queen and threatening 19 ... lbef3+) 19 1fe4 (19 'l'd1 li:lxf3 20 .ixf3 :Xe1+ 21 1Wxe1 li:lxfJ+ and Black wins) 19 ... gxf5! (the queen is trapped) 20 1Wxb7 (20 11ie3 ltlc2) 20 ... :b8 with a clear advantage to Black as White must now give up his queen.

If 18 .i.e3, then 18 ... i.g4 19 12lxe6 :Xe6 and f3 is very weak.

PETURSSON - ANAND, MANILA INTERZONAL 1990 45

18 ••. .be6 Now Black is much better; his

pieces are active and the c4-pawn is weak.

19 �5 19 c5 dxc5 is very good for Black

after :!J bxcH Vd3 ! or 20 J:[d l 'iVf6! . 19 bS!

Winning the c-pawn and forcing White to search for some sort of compensation.

20 .tb2 (D) Not :!J cxb5? .txd5 21 exd5 llJf3+

winning.

20 ... :.2! I prefer this to the line 20 ... llJxc4

21 .ixg7 �xg7 22 'iVc3+ f6 23 llJr4 c6! (23 ... .tf7 24 e5 llJxe5 25 .txa8 1fxa8 26 :e3 :cs intending ... c5 is unclear) 24 llJxe6+ :xe6 25 .tfl llle5. Here Black is a pawn up, but White has some compensation due 1D the weak queenside and the open 7th rank, which would enhance the strength of any rook penelration by

White. Note that 26 f4 can be met by 26 ... 'it'b6+ followed by 27 ... llJg4.

21 llJcJ :xb2! 22 :xb2 .txc4!

Not 22 ... llJxc4 23 :a2 with an unclear position. Mter the text, Black has extremely active pieces and strong dark-squared pressure in return for his small sacrifice. The immediate threat is 23 ... llJd3.

23 :eJ (D) After 23 J:[dl 'iVf6 24 llJd5 llJf3+

25 .txf3 'iVxf3 White is pretty close to being lost.

23 .th6? A really awful move whereby, in

one stroke, Black throws away all his advantage. The problem was that I was so excited about the way all Black's pieces were working well together that I forgot White could still develop counterplay. The correct line was 23 ... llJg4! 24 l:lf3 (24 J:[el .id4 25 'ifcl 'iff6 26 llJdl .i.xb2 wins) 24 ... .td4 25 'ifd2 c5, followed


by 26 .. . lbe5, and Black has a large advantage.

24 f4! More or less equalizing.

24 ••• lllg4 25 IU3 �g7 26 'l'dl

26 :tbl !? was also possible. 26 ••• "l'a8 27 h3 "l'a1+ (D)

The tactics don't work for Black after 21 .. ."tl'a1+ 28 �hl "l'al+ 29 :tbl �xc3 30 li'c2! 'l'a2 3 1 lhc3 'i&'xc2 32 l:l.xc2 �d3 33 hxg4 .bc2 34 .:tel �xe4 35 l:l.xc7 �xg2+ 36 >fi>xg2 .:te4 and the resulting rook ending is probably a draw in view of White's active rook position. I was quite surprised that Black had nothing better in this line; with a King's Indian bishop and a knight rampaging around, you expect something to work but in this case there was nothing.

21 �n �xn

29 lW1 "ifa3 Once again White can hang on af

ter 29 ... "i'a7+ 30 >fi>hl lbe3 by 31 llcl ! lllc4 32 :ta2 'irb7 33 'Wd5 !.

30 lU3 lbf6 31 lle3 (D)

31 dS Black can regain the exchange by

3 l .. .lbd5?! 32 lllxd5 'l'xb2 33 'l'xb2 �xb2, but after 34 lllxc7 Black is suffering, for example 34 ... l:l.b8 35 lld3, 34 ... llc8 35 lllxb5 llb8 36 l:l.b3 ! or 34 ... �d4 35 lllxe8 .ixe3+ 36 �g2 �d2 37 ll:lxd6 .ixb4 38 lbxb5 and White is better in every line.

32 li::JxdS White should avoid 32 e5 d4!.

which gives Black a clear advantage after 33 'irxd4 'l'xb2 34 exf6 :.Xe3 35 'itxe3 �xf6. However, 32 ll:ld!! 'fila? 33 e5 was a perfectly good a!· temative to the text; after 33 ... ll:le4 the position is unclear.

32 .•• "ifa1+ 33 l:l.e1 1Wa7+! (D)

PETURSSON - ANAND, MANIU INIEKUJNAL 1990 47

I could have forced a draw by defensive moves subconsciously re-33...li:'lxe4 34 l:lxal ltlxd2 35 l:txd2 !axes just when his troubles appear J.xal 36 li'Jxc7, but even though to be over, and commits a further erthings had not gone according to plan, ror. In this case time-trouble proba-1 win as st 11 hop ng to w n. Someti es ly a so pla e it is a mistake to play under the influ- 34 ••• lLixds ence of your former advantage, but 35 l:ta2 1rd4! in this case Black's optimism proved 36 11'xd4 i..xd4 justified. Petursson had overlooked that

34 �h2? Petursson returns the favour with

this blunder. The alternatives were: 1) 34 �hi?! ltlxe4 35 l:la2 1Wb7 !

is also bad foe White. 2) 341We3! 'l'xe3+ 35 ttlxe3lDxe4

36 l:lc2 li:lxg3 37 ttlg2 l:d8 38 .l:l.xc7 �f5 with a near-certain draw. White has a nominal material advantage, but Black's pieces are well coordinated and White's pawns are disjointed, so he has no chance of putting it to use.

It often happens that a player who has fought back from a bad position with a long series of accurate

there was no way to catch one of the minor pieces, even though they are temptingly lined up on the d-file. 37 .l:l.dl fails to 37 ... li'Jc3 and 37 .l:l.d2 to 37 ... .J.c3.

37 .:as ltlc3 38 eS .J.b6 39 .l:l.a3 ttlds 40 .ltb3 (D)

40 ttlxb4! Simplest. Now if 41 .l:l.xb4, then

4I ...i.a5 42 .l:!.ebl .i.xb4 43 :Z:xb4 :Z:b8, and 44 ... c5, when theconn��eted passed pawns will romp home.

41 .l:l.dl cS 42 l:ld7 l:lb8

48

43 g4 44 l:a7 45 l:a6

VISHY ANAND: MY BEST GAMES OF CHESS

46 l:a3 47 �g3

0-1

.i.dl b4

This game put me on + 1, but although it put me in the right direction, it was a late winning streak in which I beat Lautier, Miles and M. Gurevich in consecutive rounds that made me a qualifier for the Candidates.

My next major event was the Novi Sad Olympiad towards the end of 1990. I have always enjoyed playing in Olympiads. Dubai 1986 and Manila 1992

were my favourites; in both cases the organizers went all-out to make the players feel comfortable. However, the Novi Sad Olympiad was also pleasant, despite the cold and dismal weather. In general I enjoyed playing events in the former Yugoslavia because of the great public interest, which always ensured a good turnout of spectators.

I started the Olympiad with a good win against Olafsson, but lost a rather silly game to Bouaziz in round 3. Later on things started to get better, and I was satisfied with my final score of7 1h/12. The following eventful game was played in round 7.

Game 9

V. Anand - I . Morovic Fernandez Novi Sad Olympiad 1990

Sici l ian, Maroczy Bind

1 e4 cS 2 lllf3 lllc6 3 d4 cxd4 4 lllxd4 g6 5 c4 .tg7 6 i.e3 lllr6 7 llJc3 0-0 8 .tel d6 9 0-0 llld7 (D)

I had played the white side of the Maroczy Bind before, including a good win against Larsen from the Cannes event mentioned previously. However, in that game Larsen played 9 .. ..i.d7 10 'itd2 li:lxd4 l l .txd4 .tc6. Morovic adopts a rather unusual move, re-deploying his knight to the queenside.

10 llb1 I was just casting around for a

logical move, and since Black intends to play ... lllc5, it seemed reasonable to prepare to meet it with b4. It perhaps looks a little odd to put the rook on b I rather than c I or (after 'l'd2) d l , but in fact White often puts his rooks on b l and c1 in this variation, in order to support a queenside pawn advance. Typically W bite continues l:l.ci-bl to play a2-a3. Black responds with ... Wb6-b4. White then plays l:l.fc I to support the c3-knight. Play then revolves around White's ability to get a3 and b4 in, and Black's ability to stop it. The idea behind the text-move is to save time by going to b1 directly.

10 ••. lDcS When he played this anyway, I

wondered what the idea was, as he seemed to be running into b4. Then I saw that after I I b4 llle6 1 2 lllxe6, he could play 1 2 .. .fxe6! (1 2 ... .ixe6?! 1 3 llld5 is clearly better for White, to be followed by 'trd2, l:l.d1 and possibly c5) 13 lllb5 e5. However, 14 c5 would then give White a slight advantage, so this would have been a valid alternative to the text-move.


11 'ifd2 A flexible alternative; White keeps

open the possibility of b4 while developing his pieces.

1 1 ••• liJxd4 Black decides to prevent b4 by

playing ... a5. If immediately l l ...a5, then 12 liJdb5; the plan is f3, followed by liJd5, and White will play b3, a3 and b4 at his leisure. Black's c5-knight can eventually be kicked away, while the knight on b5 is there permanently. Hence Black's decision to swap knights before advancing the a-pawn.

12 �xd4 aS 13 b3

White is aiming for an eventual b4, but the immediate 1 3 a3 allows 1 3 ... a4.

B

13 �d4 14 'iVxd4 (D)

14 b6?! There are some lines of the Mar

oczy Bind in which ... b6 and .. . �b7

is played, but normally on! where Black is aiming for central and kingside play by ... liJh5 and .. f5 - it's a completely different type of plan to that Black has adopted here. In this position ... b6 is an inaccurate move, which reduces Black's options. His queen can no longer occupy b6 and in some lines where Black plays ... e5, it is useful to have . ..l:ta6 defending the weak d6-pawn.

14 ... �d7 was probably better, although Wbite has the interesting line 15 e5 liJe6 16 '1Fe3 dxe5 17 :tbdl ! and Black still faces difficulties (if 1 7 ... f6, then 18 �f3).

15 J:lfe1 Now White has a very comfort·

able position. He has a space advantage and Black has no chance of playing ... b5 or ... d5, the two breaks which normally give Black counter· play in the Maroczy Bind. Moreover ... a4 is no danger, as Black cannot back it up by ... 'llrb6, so White can always reply b4.

The only question is how White arranges to play h4-h5 to step up the pressure on Black's poorly defended ldngside.

15 ••. �b7 16 .l:[bd1

The rook has done its duty on bl. inducing Black to weaken hi queen· side. Now the need is for play in the centre and in the changed circum stances White reacts by moving his rook to the half-open d-file.

ANAND - MOROVIC FERNANDEZ, NOV! SAD OLYMPIAD 1990 51

16 ••• f6 17 .i.g4

A nice move, activating White's bishop. He doesn't mind exchanging bishops, because in the Maroczy Bind White doesn't need many minor pieces to prosecute his advantage. Black can normally hold a pure major-piece position, but even one pair of minor pieces can be enough for White to exploit his space advantage.

17 ... .ic6 18 b4 'fkc7 19 'i'eJ (D)

19 h5!? is possible, but I didn't want to cornrni t my pawns to light squares too quickly; Black might still set up some sort of dark -squared blockade by continuing .. ,<jJg7, .. . g5 and ... h6.

19 ... 'iWb7?! Black should play 19 ... <jJg7!, when

after 20 h5 g5. followed by . . . h6, Black has set up the sort of blockade

mentioned above. In this case White would have to prepare h5, for example by playing llld5 or f4. It's only a small point, but in such positions you have to put as many obstacles in your opponent's path as possible.

20 bS White can push his h-pawn with

out more ado, because ... g5 can always be met by h6, both creating a permanent danger to Black's king and making f4 much stronger.

20 ·- M7 (D) Now it is too late for 20 ... <jJg7 as

White can play 21 h6+ <jJhS 22 f4, preventing ... g5.

21 lllds Preparing to step up the pressure

on g6 by lllf4. 21 hxg6 would be premature; Black can defend after 2I ...hxg6 22 'ifh6 ltg7.

21 ... gS 22 h6

As intended. Now the g5-pawn is vulnerable to f4, and Black has to


look after e7, as the rook can always be driven away by ..th5.

22 ••• �h8 23 a3

Black's position on the kingside is creaking and now Wbite inconveniences him further by resuming his queenside play.

23 e6 24 b4 axb4 25 axb4 exdS 26 exdS?!

A mistake, not because it doesn't win but because 26 bxc5 ! was much simpler. If Black takes on c4 or e4, White obtains a tremendous passed pawn on d6, while after 26 ... d4 27 :lb.d4 dxc5 28 .l::!d6, followed by e5, Black's king will come under a devastating attack.

26 ..ta4(D)

27 bxcS? Seduced by beauty, I allowed my

self to be distracted from the process of gaining a point! I had seen the

winning move 27 ll:al !,whereupon 27 . ../i',d7 28 .:lb.a4 l:l.xa4 29 •e8+ lt:lf8 (29 .. 1lf8 30 1le7 forces mate) 30 'ilfxa4 is hopeless for Black, but I unwisely decided to 'win' in more flashy style.

27 28 c6

..txdl 1le7! (D)

Somehow this move had escaped my attention. White still has the ad· vantage because of his powerful protected passed pawn, but of course I had spoilt my winning position.

29 'li'c3 White has a slight advantage after

29 .l::!xd l 'li'xe3 30 fxe3 f5, but the pawns appeared so strong that I thought I could play on with a queen against two rooks and a bishop.

29 'li'xel+ 30 "ifxel hg4 31 'li'b4?

Another error. 31 "tfe4! .i.h5 32 'li'e6 wins the d-pawn (32 ... l:ld8? 33 c7 wins) without allowing Black to

ANAND - MOROV/t FERNANDEZ. NOV/ SAD OLYMPIAD 1990 53

take the c4-pawn in return, which should be enough foc a clear advantage.

31 32 "ilxd6

33 c7?!

.ie2 .ixc4 (D)

33 'i'c6 is a better try: 1) 33 .. . :Z.ff8? 34 'il'e7 wins im

mediately. 2) 33...�g8 34 c7 .ixd5 (or

34 ... .ia6 35 d6 .ic8 36 'ii'e8+ l:tf8

37 d7 .ixd7 38 "ilxd7 and wins) 35 c8'ii'+ (35 "ilxd5 .l:te836 "ild8 is also effective) 35 ... .:Xc8 36 1Wxc8+ .l:tf8 37 'ii'd7 .if7 38 "ile7 wins .

2) 33 ... .l:tafll! 34 "ile4 .ia6! (a difficult move to see; after 34 ... .ixd5 35 "ilxd5 Black can't immediately double rooks against the pawn, for example 35 ... .l:tc7 allows 36 "ild6) 35 d6 .ic8 36 'ii'd5 �g8 37 d7 .ixd7 38 cxd7 .l:td8 39 "ile6 Wf8 and Black draws.

33 ••• .l:te8?? A time-trouble blunder. Black

could force a draw by 33 ... .l:taf8! 34 "ilc6 .ixd5 35 "ilxd5 .l:txc7 36 "ild6 .l:tcf7 37 11 xb6.

34 We6! Grabbing my chance.

34 ... .l:tfl'8 35 c8'1f!

A pretty win. 35 .:XeS 36 'ii'e7 1·0

InApril 1991 I played in a tournament held in Munich. While I very much liked the city, I have less happy memories of the chess. In fact, my greatest pleasure was the blitz tournament held at the end, which I won with 14/15, 21h points ahead of the next player.

In round 1 I lost to Nunn, and in round 2 I was fortunate to defeat Zsuzsa Polgar. In round 3 I met Beliavsky, and the result was one of the few good games I played at Munich.

Game 10

A. Beliavsky - V. Anand Munich 1991

Pi rc Defence

In our previous encounter (Linares 1991) I had built up a totally won position only to perpetrate a form of hara-kiri. I was very pleased to get revenge in this game!

1 d4 d6 2 e4 liJf6 3 liJc3 g6 4 (4 ..tg7 5 liJf3 0-0 6 ..te3 b6 (D)

When I was studying this line from White's point of view, I wondered why 6 ... b6 wasn't a more popular reply. Black forces through . . . c5, and if White is to try for an advantage he has to push all his pawns forward in the centre, which is very

committal. If the pawns eventually tum out to be weak, the weakness

wiJI probably be serious. 7 e5 liJg4 8 .i.gl cS 9 h3

White avoids a little trap: 9 dxc5 bxc5 10 'ill'd5 'ill'b6 I I ifxa8 loses to I I .....tb7 1 2 liJd5 ifxb2.

9 liJh6 10 d5 i.b7 (D)

I knew the theory, but at this point I decided to ignore it and just look at the position. It seemed to me that Black could play very natural moves. The point of this one is to play ... e6 and completely destroy White's cen· tre. After the resulting exchanges

BEUAVSKY - ANAND, MUNICH 1991 55

Black may be left with a weak pawn (for example, on e6) but it doesn't matter because Black has generated so much active play for his pieces. While the specific move IO ... ..i.b7 was thought up over the board, I had looked at these lines before and the ideas I had during this earlier analysis germinated into this 'innovation'.

11 "i'd2 After 1 1 "i'e2 ..i.a6 ( I I ...a6 and

I2 ... b5 is also possible) I2 'l'f2 i.xfl 13 "i'xfl lLlf5 Black equalizes comfortably. If White tries to shut the h6-knight out by I I g4, then Black plays I l...dxe5 I 2 fxe5 e6 I 3 i.c4 exd5 ! ( 1 3 ... b 5 I 4 ..i.xb5 exd5 IS i.xc5 favours White) I4 ..i.xd5 with an unclear position. The knight on h6 is bad, but the e5-pawn is weak and gi-bishop is also oddly placed.

11 ... lLlfS 12 ..i.h2

After I2 ..i.f2 dxe5 I3 f xeS e6 White doesn't have time to castle long owing to I4 0-0-0 .i.h6. Hence the text-move.

12 dxeS 13 fxeS e6! 14 0-0-0

Other moves are ineffective, e.g. I4 g4 i0114 is very bad for White, I4 d6 lLld7 leaves e5 collapsing and finally 14..i.c4 .ih6 (14 ... exd5 15 i.xds lLlc6 and I4 ... b5 I5 lLlxb5 exd5 16 i.d3 are also possible, in both

cases with an unclear position) IS

..i.f4 ..i.xf4 I6 1Wxf4 exd5 I7 0-0-0 d4 IS lLle4 lDd7 leaves White without enough for the pawn.

14 exdS 1 S lLixdS lL\c6 16 c3 (D)

Underestimating Black's initiative. If White had time for ..i.c4 and l:hei then he would have a clear advantage, but Black's counterplay is so fast that he has no time for the necessary consolidation, e.g. I 6 ..i.c4 lLlcd4 I7 lLlxd4 cxd4 attacking the d5-knight and, when it moves, creating the possibility of ... lLle3.

16 ••• lLlcd4 17 lL\r6+

I7 cxd4 11fxd5 I S �bi :adS is very good for Black.

17 ..i.xf6 18 cxd4

IS exf6 .'Llxf3 I9 gxf3 11fxf6 gives Black a safe extra pawn.

18 ..i.g7 19 dS (D)


19 c4! Essential. If White could play .i.c4,

then Black would be in some trouble. Now Black locks in the bishop on f1 and frees the c5-square for the manoeuvre ... l:lc8-c5. All Black's moves are very natural; he has the initiative and must make use of it before White completes his development, so he need only consider active moves.

20 .tel? In the critical position White goes

wrong. He should play 20 �bl , with the possible continuation 20 ... l:lc8 21 d6 (21 .i.e2 .i.h6 22 .i.f4 .i.xf4 23 'l'xf 4 .i.xd5 transposes to the game) and now Black can choose between 21 ...1lrd7 and 21...b5. Black has a lot of trumps, not least of which is his lead in development. The position is unclear but I prefer Black.

20 ... l:lc8 21 �b1

21 g4 Ci:Je7 22 d6 00, threatening 23 . .. c3, favours Black.

21 ... .1h6 22 .i.f4

After 22 11rc3 Black can also safely take the d5-pawn.

22 Jud4 23 11rxf4 .i.xdS! (D)

-w •

24 h4?! There is no way White can exploit

the d-file pin, so he tries for a kingside attack. The alternatives were:

1 ) 24 Ci:Jd4 c3 (24 ... Ci:Jg7 25 ttlb51 is bad, but 24 ... 'llrb4 25 .i.g4 1Llxd4 26 llxd4 llcd8 is a reasonable alternative for Black) 25 bxc3 'i'h4 26 .i.g4 Ci:Jxd4 27 l:lxd4 .i.xg2 28 :gl .i.xh3 29 .i.xc8 11rxf4 30 J:lxf4 AxeS is excellent for Black. He has two connected passed pawns for the exchange and White's remaining pawns are weak.

2) 24 lld2 c3 (alternatively, after 24 . . . llc5 25 llhdl 11ra8 Black keeps his extra pawn) 25 bxc3 J:lxcJ 26 llhdl Ci:Je3 and Black is clearly better.

BEUAVSKY - ANAND, MUNICH 1991 57

3) 24 g4 tiJg7 25 'l'h6 tiJe6 26 h4 c3 27 bxc3 (27 �g5 c2+ 28 �a! cxd l'l'+ 29 .l%xdl �xg5 and Black wins) 27 . .. i.e4+ 28 �b2 'ilc7 favours Black. White's pawns are weak and his king exposed.

24 ••• c3 25 bxc3 .l%xc3 26 bS lLle3!

Black's attack is much faster than anything White can muster.

21 lt)gs

After 27 hxg6 fxg6 28 'l'h6 "flc7 Black defends h7 and White will be mated, while after 27 "flh6 ll!xd I 28 hxg6 (28 ll!g5 .i.xa2+ forces mate) 28 ... .i.e4+ 2 9 �a! .i.xg6 Black wins easily.

27 •••

28 ll!m7 0-1

flc7 l:lb3+!

It is mate after 2 9 axb3 flc2+ 30 �a! 'ilc3+ 3 1 � bl "flxb3+ 32 �al lllc2#.

A few days after qualifying from the Manila Interzonal, I received an invitation to play in Linares ( 1 991) from Seiior Rentero. On my way to Linares (which was a couple of months before the Munich event mentioned above) I stopped off for a couple of days in Amsterdam. There I received the news that FIDE had made the pairings for the quarter-finals. Instead of the simple 1 vs 8, 2 vs 7, etc., they had changed the system such that anyone in the top half could be paired against anyone from the bottom half .

Karpov, Timman, Yusupov and Short were in the top half of the draw, with Gelfand, Ivanchuk, Korchnoi and myself in the bottom. I was paired against Karpov, who at that time was a formida ble opponent. Just at that moment I was quite annoyed by this pairing, but later I took the view that you couldn't become World Champion by avoiding people - you just have to take opponents as they come.

I was heartened by the fact that I beat Karpov quite easily in Linares after he misplayed a promising position. Between then and the match in August neither of us had produced any inspiring results. My own performance in Munich was not very satisfactory, while Karpov had drawn a match 2-2 with Agdestein.

In the first game of the match itself I played an insipid system; to be honest we (my second in this match was M. Gurevich) knew that it gave White nothing against best play, but we decided to try it anyway. because Karpov had failed to find the correct solution in a previous game. He got a bad position with an isolated pawn and suffered a lot, but defended very well and, indeed, outplayed me completely. However, he threw away all his good efforts by misplaying the ending.


In the second game I outplayed Karpov completely, but then went wrong and had to acquiesce to a draw. In the third game I could have mated him in a few moves, but somehow just didn't see it. Then I lost the founh game. In both game two and game four I had played the Meran Defence, which I prepared especially for this match; I felt it was a dynamic opening and that Karpov wasn't particularly good against it. In games two and four I had played after I d4 d5 2 c4 c6 3 tbf3 tbf6 4 tbc3 e6 5 e3 tbbd7 6 .ie2 .id6 7 0-0 0-0 8 'i!Vc2 dxc4 9 .ixc4, a plan involving ... a6. In game four Karpov had found a pretty good line against this system, so in game six I decided to switch to 9 ... 'i!Ve7, which also formed part of my preparation. I should add that game five was unfinished when the following game was played, but I was winning the adjourned position.

Game 1 1

A. Karpov - V. Anand Candidates match (6), Brussels 1 991

Semi-Slav

1 d4 dS 2 c4 c:6 3 lllf3 �f6 4 0.c3 e6 5 eJ �bd7 6 'i'c2 i.d6 7 i.e2 0-0 8 0-0 dxc4 9 hc4 'ii'e7

10 h3

To avoid the exchange of knights after ... �e5, but it is not a very ambitious move.

13 ••• iDeS 14 .i.b3 i.d7 15 i.eJ (D)

In game eight he finally found the right recipe, which is to play 10 a3.

B

10 ••. cS In game four I had played 10 ... a6.

11 dxcS i.xcS 12 e4 (D)

12 • • •

13 �d4 i.d6

15 ••• �g6 15 ... l:tfd8? is a loss of time; White

continues 1 6 f4 �c6 17 e5 �xd4 1 8 i.xd4 i.c5 19 l:tad 1 (after 1 9 i.xc5 'ihcS+ 20 'ii"f2 'ihf2+ 21 :x£2 iDeS 22 l:td 1 i.c6 23 l:tfd2 l:txd2 24 l:txd2 �f8 Black should hold on) and now:

1) 19 ... i.c6?20 exf6 l:!.xd4 (White also wins after 20 ... i.xd4+ 2 1 l:txd4 1Wc5 22 lbe2!) 21 fxe7 l:txd l+ 22 �h2 l:txfl 23 'i'd3 i.gl+ 24 �h 1 and White is winning.


2) 1 9 ... .ixd4+ 20 :Xd4 ll:le8 21 l:l.fd1 and White is clearly better.

After the text-move Black has more or less equalized.

16 J:l.ad1 J:l.fd8 17 ll:lr3 .ic6 18 J:l.fe1 l:tdc8

This looks a little odd, but I felt it was important to inconvenience White's queen by playing a rook to c8. The aS-rook can't go there because the a7-pawn is hanging, so it has to be this one. White must move his queen because 19 ... .ixe4 is a threat.

19 'l'bl (D)

. 0,\ -B -

19 NO .ib4 Everything is based on threats

against the e4-pawn. 20 .id2

After 20 .id4 e5 the e-pawn would be in serious trouble. In this position only White can be worse, because of his exposed e-pawn.

20 l:td8

The rook returns now that White has been forced to block the d-file.

21 a3 .icS 22 ll:la4 .id6 (D)

Black is more or less committed to this piece sacrifice, as 22 ... .ixa4 23 .ixa4 would give White a slight advantage based on his two bishops. However, I had no objections as I felt that it was promising for Black.

23 ll:lc3? After this cop-out, Black's pieces

are more harmoniously placed. The critical line is 23 e5! .ixf3 and now:

1) 24 exf6 (bad) 24 ... '1'xf6 25 gxf3 .if4! and now:

Ia) 26 .ic3 "irg5+ 27 �h1 (27 �fl "irb5+) 27 ... 1Wh5 28 �g2 lllh4+ and wins.

1b) 26 .ixf4 ll:lxf4 also wins. 1c) 26 .ib4 'ii' g5+ 27 �fl 'li'b5+·

followed by 28 ... a5, with a distinct plus for Black.

2) 24 exd6 (best) 24 ... '1'xd6 25 gxf3 ll:lh4 and now:

KARPOV - ANAND, BRUSSELS CANDIDKTES 1991 61

2a) 26 .tc3?! lllxf3+ 27 �g2 ®14+ 28 1i>fl (28 'iii>g1 'ifc6 is dead lost) 28_.'i'a6+ 29 l'Z.e2 'il'c6 and Black should win.

2b) 26 .i.e3? lilxf3+ wins. 2c) 26 l'Z.e3 (D) and now:

2c l) 26 ... 1i'f4?! 27 l'Z.d3! lilxf3+ 28 �g2 lilh4+ (28 ... lilxd2 29 l'Z.1xd2 'i'gS+ 30 �fl 1i'xd2 3 1 l'Z.xd2 l'Z.xd2 gives White an edge) 29 �fl 'il'h2 30 .tg5 ! (30 .i.c3 lild5 31 .i.xd5 l'Z.xd5 32 lilc5 is unclear) 30 ... l'Z.xd3 31 'i'xd3 'il'g2+ 32 �e2 lilg6 33 l'g3 favours White.

2c2) 26 ... 1i'c6 27 .i.c3 .!l'lxf3+ 28 �fl b5 with a final branch:

2c21) 29 l'Z.xd8+ l'Z.xd8 30 .i.d1 IOd4 (30 ... lild2+ 31 .i.xd2 l'Z.xd2 32 .tn defends, while 30 ... lilh4 31 .txf6 gxf6 32 l'Z.g3+ �h8 33 .!l'lc3 'i'h I+ 34 �e2 lilf5 35 l'Z.d3 lild4+ 36 �d2 is at least equal for White) 31 i.xd4 l'Z.xd4 32 lilc3 1i'h1 + 33 �e2 lllh5 34 J:f3 1i'h2 35 .i.c2 lilf4+ 36 �3 lilg2+ 37 Wxd4 1i'd6+ 38 �e4

'il'c6+ 39 �e5 and this exciting line ends in a draw by perpetual check.

2c22) 29 lilc5 lilh2+ 30 �e2 'il'xc5 31 l'Z.xd8+ l'Z.xd8 32 .i.xf6 gxf6 33 1i'g1 + �h8 34 'il'xh2 'il'cl is unclear.

The conclusion is that White may be able to hold the balance by accepting the sacrifice, but it would have been very difficult for Karpov to find all this at the board!

23 ••• 'il'c1 (D) If we look at the total effect of the

last five moves, White has played .i.d2, 'ifbl and a3, while Black has achieved ... 1i'c7. Thus White's queen and bishop have been pushed back, while Black has improved his queen position and now controls e5 -Black's manoeuvre can be counted a success.

Now Black is slightly better because he can expand on the queenside, while in the meantime White's pieces are only crawling back to their earlier positions.


24 �a2 It's hard to say what this move is

for. Perhaps he was worried that after ... j_cs, there might be a threat of ... j_xf2+ and ... 'l'b6+, but it's only a guess.

24 •.• a6 25 j_eJ b5

White's 24th move is revealed as a mistake. It may have solved one problem, but now the bishop has to return in order to meet ... 'l'b7 by j_c2.

26 j_b3 j_b7 27 .:tel 'l'e7 (D)

I was very proud of this game, because I felt that I had outplayed Karpov in the manoeuvring phase of the game. The idea now is . .. �7 followed by either .. .lllcS or ... lLldeS heading for c4.

28 .1b6 l:.dc8 29 .1d4

Now 29 e5 .ixf3 is good for Black, as 30 exf6 1i'xf6 leads to variations

similar to those in the note to White's 23rd move, while 30 exd6 'l'xd6 attacks the bishop on b6 so there is no win of a piece.

29 ·- �7 30 l:.cdl

Over the last few moves White has only been moving his pieces backwards and forwards, simply responding to Black's various threats, while Black has gained space on the queenside and created an outpost at c4. Nevertheless, the symmetrical nature of the position exerts a drawish tendency. Black's position is more comfortable and easier to play, but one cannot say more than that

30 liJgeS 31 liJxeS lillce5 (D)

32 liJe2? An error, overlooking the reply.

32 ••• 1i'h4 33 f4

After 33 lt:lc3 Black continues 33 ... lllf3+ 34 gxf3 1i'xh3 35 l:.d3

KARPOV - ANAND, BRUSSEU CANDIDKI'ES 1991 63

.A.h2+ 36 �hl .i.£4+ 37 'iilgl :tc6, and the possibility of playing ... e5 and switching the rook to the kingside gives Black an easy win. The alternative 33 lLlg3 is similar; then 33 .. .<�f3+ 34 gxf3 .bg3 35 'iilg2 .A.f4 doesn't lead to a winning attack, but White's pawn structure is significantly damaged.

The text-move is therefore forced, but it is a move White certainly doesn't want to play with Black's bishops pointing at the kingside.

33 ••• lLle4 34 .if2 .icS 35 .!xeS :XeS (D)

36 .l:l.el? 36 .l:l.d4 was a better defence, when

36...ll:lxb2? loses to 37 :tfl lLlc4 38 .ixc4. However, Black can continue 36 ... .1:1.ac8 with a clear advantage.

36 •.• li)d2 37 Wd3 li:Jxe4?!

One of my weaknesses during this match was my inefficiency in

converting technically winning positions. Too often I played moves which just maintained my advantage, instead of pressing it home forcefully. On the other hand, Karpov, although he only gained a large advantage twice, pushed it home both times, and this effectively decided the match.

Here 37 ... .l:td8! 38 'ire3 lLlxb3 39 'irxb3 .ixe4 would be a much improved version of what happens in the game.

38 'ife3 39 .l:l.edl

.l:l.ac8 h6

39 ... g6 would have been slightly more precise.

40 :n 41 lLlc3

lLlf6 'ii'hS (D)

Black has a clear extra pawn, and despite White's stern resistance, the win should only be a matter of time.

42 :1d6 .I:I.Se6 43 .l:l.xc6 .ixe6 44 :1f2 li)d7


45 �d2 I decided not to try to find a mid

dlegame win, but just to liquidate down to an ending.

46 'i'xe5 47 .tel 48 <M2 49 �e3 50 g3 51 ..tdl

ltJxe5 Wf8 �e7 liJd7 �6 ltJe8 (D)

At some point I should play ... g5, so that if White plays h4, then Black can reply ... gxh4 followed by .. .J:tg8, when White has the possibility of a g-file penetration to worry about.

52 .tel 53 .i.d3

ltJd6 .i.b7

53 ... g5 would still be good, but this is the last chance as White now prevents it (of course, he could have played h4 earlier himself).

54 h4 Now any attempt to prepare ... g5

will just lead to a lot of simplification.

54 "" ltJc4+?! A real mistake. Again it doesn't

give away Black's advantage but it makes the win fairly complicated 1n endings with bishop and knight each, exchanging dissimilar pieces (i.e. a bishop for a knight) has the effect of increasing the defender's drawing chances. The superior side should therefore only exchange similar pieces.

55 he4 lbc4 Now White has much better pros

pects of setting up a dark-squared blockade.

56 :td4 :tcS (D)

57 a4 At this point I decided to sit down

and really calculate, because I could see that the win was no longer going to be trivial.

57 ••• eS Over the next few moves Kar

pov's resistance starts to weaken. 58 l:lb4

KARPOV - ANAND, BRUSSEI.S CANDIDATES 1991 65

Already 58 fxe5 l::txe5+ 59 'it>f2 was more accurate, when Black has a long way to go to create a passed pawn on the kingside.

58 .i.c6 59 axbS axbS 60 �2? (D)

This was White's last chance to play 60 fxe5 l::txe5+ 61 'it>f2.

B

60 f6!

Now Black retains control of the key dark squares d4 and f4, and the

Another bad move. White should have tried 6 l ll'lc3.

61 ••• 'it>e6 62 l:la3 l:lc2

Now Black is winning. 63 fxe5 fxe5 64 l:la6

The only line I needed to calculate was 64 l:lc3 l:lxc3+ 65 ll'lxc3 'it>d6 66 b4 'it>e6 67 �d3 'it>£5 68 �e3 �g4 69 �f2 e4 70 ll'ld I 'it>£5 71 �e3 �e5 72 ll'lc3 g6 (but not 72 ... g5? 73 hxg5 hxg5 74 g4 and White escapes with a draw) 73 g4 g5 74 h5 .i.d7 and Black wins.

64 ••• �d6 65 b4

65 l:la7 l:lxb2 66 l::txg7 b4 is also an easy win.

65 i:ia7 :C4

66 .i.d7 67 l:la6+ �7 68 l:lg6 �7 69 l:ld6 .i.g4 70 l:ldS l:lc2

white knight will feel a lack of good 0-1 squares. As the knight is trapped after 7 1

61 :b3? ll'lgl l:lg2.

This game was finished after the conclusion of the fifth game; the two adjournament were played on the same day. Unfortunately , in the adjourment of the fifth game I first made a simple win rather complicated and then, when I need to calculate a long forced line in order to win , I made a mistake that allowed him to gain a tempo. The game ended in a draw. Here Karpov's experience also played a part. He had two lost positions (games five and six), but he concentrated all his efforts on finding resources in game five (where had a better chance to save the game) and eventually managed to salvage a half-point


I missed a forced win in the seventh game and then he won the eighth game to take the match. Cenainly he defended very well in this match, but he didn't demonstrate the form which would have taken him to the world championship, and he subsequently lost to Short in the semi-finals.

The two matches I played in this Candidates cycle were virtually the first matches I had ever played. Before, I had only played a not especially serious four-game match with Levitt. When the next Candidates cycle came around, the experience of match play that I gained against Dreev and Karpov turned out to be very useful.

The 1991 Til burg tournament was a double-round event. Prior to this tournament I had only played Kasparov once, the game ending in a draw. In the first cycle at Tilburg I lost, but gained my revenge in the game immediately following. I won again in Reggio Emilia (see Game 1 3) but, to date, that was the last I saw of a plus score against him!

Game 12

V. Anand - G. Kasparov Tilburg 1991

Sicilian, Scheveningen

This was quite a pleasing game.ln fact, I hadn't really bothered to prepare for this game - I decided that

whatever I did, it would be inadequate. He'd played this line so many times I couldn't hope to out-prepare him, so I preferred to concentrate on keeping a clear head for the game. Although the strategy worked well on this occasion, it would be easy to exaggerate its advantages!

1 e4 cS 2 �3 d6 3 d4 cxd4 4 �d4 �6 s ltlc3 a6 6 f4 e6 7 i.d3 lllbd7

1 later discovered that Kasparov and Nikitin's book on the Scheveninge gives 'The position after 8 0-0

'l'b6 9 �e3 'i'xb2 10 lllcb5 axb5 1 1 �xb5 J:ta5 is interesting for analysis' , but during the game I wasn't aware of this. At the board I was considering 8 0-0, and my analysis ran '8 0-0 'i'b6 9 �e3 'ii'xb2 and there must be something strong.' After 1 played 8 0-0, he instantly flashed out 8 • . 'i'b6 and I thought for a

while, during which I saw that it was not so easy.

8 0-0 Vb6 9 �e3 11'xb2

Here I realized that I had to sacrifice on b5, or else I would just be a pawn down for nothing.

10 llldbS (D) 1 0 'ii'd2 ti)g4 is clearly better for

Black.

10 ••• axbS 11 lllxbS :as

The line which had convinced me to play 10 llldb5 was 1 1 ... 'ii'b4 1 2 lllc7+ �d8 1 3 lllxa8 'ii'a5 1 4 lllb6 lllxb6 15 'ife1 ! 'ifa7 16 a4 lllg4 1 7 a 5 and White wins. However, Black can improve on this by 12 ... �e7! 1 3


�xa8 'iVa5 which leads, just like the game, to a complex and unclear position.

Other moves are bad, for example 1 1 .. .\t>dS? 12 a3! (threatening 13 �d4) 12 . . . lh4 (12 ... d5 1 3 l:.b1 -.a2 14l:tb3! followed by �3) 13 c4! and wins, or 1 1.. .l:.b8? 12 l:l.b1 'iVxa2 13 l:tal 1rh2 14 �d4 1rb4 15 c3 trapping the queen.

12 l:.b1 Better than 12 a4 l:txb5 (12 ... d5

13 l:.b1 'i'a2 14 l:tal is a draw) 13 axb5 'iVc3, which is better for Black, e.g. 14 l:taS �e7 15 'iVel 'iVc7 and 16 b6 �xb6 17 ..llxb6 fails because Black takes on b6 with check.

12 ••• lhbS Not 12 ... 'i1Vxa2? 13 �c3 'iVa3 14

l:l.b3 and Black does not get enough for the queen.

13 l:txb2 lhb2 (D)

was still analysis as he was playing very fast.

14 ... l:tb6 Not 14 . . . l:.b4? 1 5 'ifc3 and wins.

15 �xb6 �xb6 16 'iVc3 (D)

This move was the product of long thought. I realized that any other move would allow Black to regroup his b6-knight to c5. Then it is hard to judge whether the queen or the three minor pieces would be better, but Black's solid pawn structure gives him a head start. I therefore decided that it was necessary to keep the knight fixed on b6 as a target, and force the other knight to occupy d7.

16 i&.e7 Alternatively: 1 ) 16...�d8 17 'ii'a5 �d7 18 llbl

and now: 14 'i1Va1 Ia) 1 8...�c7 19 ..1lb5 �b8 (White

I was quite surprised that Kaspa- wins after l9 ... �c5? 20 i.e8) 20 rov went in for this, but I was sure it ..llxd7 �xd7 21 'ii'dS ffJ (2l...d5 22

ANAND - KAsPAROV, TtLBURG 1991 69

� dxe4 23 l:tb3 ta:5 24 l:lc3 is again winning) 22 <Jthl ! i.g7 23 'fle7 '/;c7 24 l:ldl and White is clearly better.

lb) 18 ... d5 19 exd5 exd5 20 i.f5 i.c5+ 21 �fl �7 22 i.xd7 lill.d7 23 'tc7 ffi 24 lldl d4 25 c3 also favours White.

2) 16 ... lilld7 17 l:lbl d5? (Black can transpose into the game by 17 ... i.e7) 18 l:lxb6 and White wins after 18 ... .ic5+ 19 <Jtfl d4 20 'i'c4 or 18...�xb6 19 'i' c7.

17 l:tbl (D)

17 li::Jfd7 Black can get castled at the cost of

his d-pawnby 17 ... i.d8 1 8 'i'd4 �7 (18 ... �fd7 19 'ii'xd6 also favours White) 19 'ii'xd6 i.e? 20 'ii'c7 0.0 21 �fl . but White retains a slight advantage.

18 'ii'xg7 Now 18 'ii'c7 fails to 1 8 .. . 0-0! 19

llxb6? .idS 20 'l'xc8 i.xb6+. 18 i.f6

Now that Black cannot castle, he finds it hard to coordinate his pieces.

19 'i'b6 (D)

19 .•• �e7 The alternative is 19 ... l:lg8 20 e5

dxe5 and now: 1 ) 21 i.xh7 with two lines: Ia) 2 I ...l:lh8 22 l:lxb6 lbxb6 (if

22 ... e4, then 23 llb4 �e7 24 llc4 wins) 23 'ifxf6 l:lxh7 24 'ii'xe5 lbd5 25 c4 lbe7 26 a4 with the plan of 'ifc7, c5-c6 and pushing the a-pawn forward to queen. Of course, matters are not so simple as this, but White has the advantage.

lb) 2 I . . .llg4! 22 llxb6 llh4 23 llxe6+ fxe6 24 'ifg6+ �e7 25 i.g8 llxf4 26 'ii'f7+ �d8 27 'ii'xe6 is unclear.

2) 2 1 llxb6! e4 22 llb4 (22 i.b5 i.d4+ 23 �fl i.xb6 24 'ifxh7 <Jtf8 25 'ii'xe4 is unclear, but 22 llxe6+ fxe6 23 i.xe4 should be slightly better for White) 22 ... exd3 23 'ii'xh7 with advantage to White.


20 i.bS? Missing a stronger possibility: 20

g4 ! (while not winning, this gives White an advantage) 20 . . . l:.g8 (not 20 ... i.d4+? 21 ..t>n tas 22 l:.b4, nor 20 ... �c5? 21 e5 dxe5 22 g5 and White wins in both lines) 2 1 g5 il..g7 (2l ... .id4+ 22 'it>fl lt)f8 is possible, but still better for White) 22 1i'h4 (D) (not 22 'i'xh7?? .id4+ 23 'it>fl !::th8 and Black wins) and now:

I) 22 ... h6 23 �fl and now both 23 ... hxg5 24 'ilfxg5+ and 23.. .�a4 24 e5! dxe5 25 fxe5 favour White.

2) 22...�a4 23 �hl (23 l:.b4 �ac5 24 'ilfxh7 �f8 25 'ilfh4 is also good for White since he has the e4-pawn well supported) 23 ... �ac5 24 g6+ (24 'i'xh7 �f8 25 'i'h4 l:.h8 26 1Wg3 b6 is only slightly better for White) 24 ... il..f6 25 gxh7 l:.h8 26 'i'h6! (now the manoeuvre l:gl-g8 will also introduce the possibility of 'ilff8+) 26 ... e5 (26...�xd3 27 cxd3 �f8 28 e5 dxe5 29 fxe5 .ixe5 30

'ilt'g5+ i.f6 31 11fc5+ �d8 32 'l'd6+ �d7 33 .l:l.gl is very unpleasant for Black) 27 l:.gl (27 l:.fl �f8 28 fxe5 i.xe5 is less clear) 27 ... exf4 28 ltg8 �e6 29 11fxf6+! �xf6 30 .!Z.xh8 and wins.

3) 22 ... e5 23 'it>hl exf4 and now: 3a) 24 g6+ .if6 25 gxh7 1th8 26

'ifb5 .ie5! 27 11fg5+ �f6 28 :Xb6 1txh7 29 ltb5! (29 .ie2 f3 30 i.xf3 l:.xh2+ 31 '1t>gl lbc2 is not easy for White - Black's pieces are Vf:I'J active) 29 ... l:.h5 30 l:.xe5+ dxe5 31 'ifgl i.d7 3 2 .ie2 (32 'ifc5+ 'ite8 33 .ib5 hb5 34 1Wxb5+ �f8 35 'lxb7 '1t>g7 is less clear as Black has counterplay with ... �g4 or with his fpawn) followed by either t!fc5+ or 'i'g5, with advantage to White.

3b) 24 e5! (also strong) and now: 3b1 ) 24 ... dxe5 25 g6+ .if6 26

gxh7 lth8 (26 ... l:tf8 27 'l"h6 lLld5 28 l:.gl i.h8 29 l:.g8 �ffi 30 l:txf8 �xf8 31 'ifg5 �g6 32 .ixg6 fxg6 33 '6'xe5+ wins) 27 '6'h5 (27 '6'h6�d5 28 l:.gl is also promising for White. 27...�4 (27...�5 28 i.c4 liJe3 29 '6'xn + Wd6 30 l:.b6+ is winning) 28 llgl �ac5 29 .ic4 � 30 i.xe61 '1fi>xe6 31 ltg8 and White should win

3b2) 24 ... .ixe5 25 g6+ .if6 (the line 25 ... We8 26 gxh7 l:.h8 27 ltb5!' �a4 28 llxe5+ dxe5 29 'ifg5 is deci· sive) 26 gxh7 llh8 27 '6'h6, again in· tending llgl -g8, and Black faces serious problems.

Kasparov pointed out many of these variations after the game. My

ANAND - KASPAROV, TILBURG 1991 71

inaccuracy arose because I failed to

appreiate that this was really a critical position in which White had to continue very precisely, and not just play natural-looking moves.

ZO llg8?! (D) Missing a simple chance: after

20 ... e5! (not 20 ... li:lc5? 21 e5 dxe5 22 fxe5 i.xe5 23 'if g5+ �d6 24 J:ldl+ lDd5 25 o4 and wins) 2 1 llfl (after 21 f5 li:lc5 Black has wrested control of some dark squares; 21 a4 � 22 a5 li:lbd7 is also fine for Black) 21 ...llg8 22 fxe5 .i..xe5 23 •xh7 J:lg7 24 1Wh4+ li:lf6 Black has regrouped his pieces with a solid position.

21 lldl! eS?! After 2l ...li:lc5 White can gain the

advantage: 1) 22 J:lxd6?! is inferior after

22 . .. J:Ig6 23 'l'h.S ! ll:lxe4! (23...�xd6 24 e5+ �c7 25 exf6 li:lbd7 26 'i'xh7 lhf6 27 g3 is slightly better for White ) 24 lld3 li:ld5 and Black's

active pieces provide him with sufficient counterplay.

2) 22 e5 ! dxe5 23 fxe5 ..txe5 24 1We3 and now 24 ... ..td6 25 'i'd4 .l:td8 (25 ... li:ld5 26 c4 wins for White) 26 'i'h4+ f6 27 'i'xh7+ �f8 28 'i'h8+ �7 29 1lg7# is mate, so Black must play 24 ... li:lbd7 25 �xd7 li:lxd7 26 .:.Xd7+ �xd7 27 11fxe5 �c6 28 g3, when White has fair winning prospects.

The best move is 2 1 ...llg4! (D), which was Kasparov's original intention, but when he was about to play it, he saw a hole in his analysis. However, it appears that the move is playable after all:

I) 22 g3? e5! 23 a4 exf4 24 a5 fxg3 25 axb6 gxh2+ 26 �hi .ie5 favours Blade - the h2-pawn is very strong.

2) 22 e5 dxe5 23 'ifh3 .:.Xf4 24 'i'a3+ �e8 (not 24...�d8? 25 'i'f8+ �c7 26 'i'd6+) 25 'ifc5 .i..d8 26 a4 and now:


2a) 26 .. J:I.f5? 27 l:l.f! l:l.g5 (White wins after 27 .. Jhfl+ 28 �xfl as Black cannot defend against a5) 28 h4! is very good for White.

2b) 26 ... e4 27 l:lfl ! (now not 27 a5? llf5 28 � 00 29 'ifc4 fi'Je7 and Black is better) 27 ... l:l.g4! 28 h4! (28 a5 .l:l.g5 29 'iff2 f6, intending 30 c4 fi'Ja8, is slightly better for Black) and now Kasparov goes into some spectacular lines, but he missed 28 ... e3 !, which looks fine for Black, e.g. 29 a5 (29 'ifxe3 lDd5 is very good for Black) 29 ... e2! 30 l:l.el .if6 3 1 c3 �d8! with advantage to Black.

3) 22 .ie2!? (probably objectively best) 22 ... l:l.g8 and White may have nothing better than to repeat moves.

22 rs (DJ

8

22 It is now too late for this move. I

was rather relieved when he didn't play 22 ... l:l.d8!, after which White is still better, but the position remains

complicated. The analysis runs 23 g4 fi'Jc5 (23 .. .1lg8 24 h3 is a simple win) 24 g5 fi'Jxe4 (not 24 ... .th8? 25 f6+) 25 gxf6+ fi'Jxf6 (Kasparov ana lysed this in excruciating detail and decided it favoure White; however, when you go for your best practical chance, you shouldn't be too fussy!) 26 'ife3! (26 c4?! .l:l.g8+ 27 �I .ixf5 28 c5 dxc5 29 11t'e3 fi'Jbd7 and 26 l:l.fl .l:l.g8+ 27 �hi l:l.g4! are less clear; Black has real counterplay in both cases) 26 ... fi'Jbd5 (26 ... �bd7 27 'ifa7! fi'Jc5 28 llxd6! :.gS+ 29 �I �d6 30 'ifb6+ wins) 27 'lb3 fi'Jc7 ! (27 ... fi'Jf4 28 'ifc4 heading for c7) 28 'ifc4 fi'Jxb5 29 'ft'xb5 ltg8+ and Black can play on, but White has excellent winning chances.

Other 22nd moves are bad, for example 22 .. . l:l.g4? 23 "ifd2 dS 24 .ixd7 fi'Jxd7 (24 ... .ixd7 25 'lb4+ and 24 ... �xd7 25 'ife2 are also win· ning for White) 25 'ifxd5 wins,or 22 ... fi'Ja8 23 .ixd7! .ixd7 24 ltxd6 .ig5 (24 ... �xd6 25 'ifxf6+ �c7 26 'l'xf7) 25 'ifxh7 and White wins.

23 l:l.xd6! Black's position crumbles.

23 ••• .tgS 23...�xd6 loses to24 'ifxf6+�c7

25 'ifxe5+. 24 'ifxb7

Now all the tactics work out for White.

24 ••• fi'Jxe4 Or 24 ... �xd6 25 'il'xg8 .ie3+

(25 ... fi'Jxe4 26 'ifxO and wins) 26

ANAND - KA.SPAROV, TILBURG 1991 73

�I �xe4 (26 ... </we7 27 f6+ 'Jwxf6 28 'l'd8+ and 26 ... .i.d7 27 .i.xd7 �bxd7 28 1fxn are decisive) 27 'lxn with too many passed pawns.

25 :xb6 .:td8 25 . .. .i.e3+ 26 �fl :g4 (26 .. Jld8

27 'i'h4+) 27 f6+ �8 (27... �dB 28 'i'xn) 28 �e2 would be fatal for Black.

26 .id3 .te3+ 21 �n (DJ 27 ... .txb6

27 ... 00+ 28 �e2 hb629 'ifh4+ �e8 (29 ... �d7 30 �xd2) 30 .i.b5+ i.d7 31 'l'h8+ �e7 32 'l'xe5+ � 33 'i'd6+ decides the game.

28 .i.xe4 :d4 29 c3 1-0

1n view of 29 ... :Xe4 30 ffi+ �xf6 3 1 'l'xe4.

The result of this game was a pleasant surprise for me. At the time Kasparov already had a huge lead over the other players, but now I had visions of overhauling him. I already had a winning adjournment against Kamsky in the bag, then I won this game, and the following day I had a winning position against Karpov. However, I lost the game against Karpov and then I blundered into a mate in two in the Kamsky adjournment, so to win the tournament Kasparov didn't have to do anything other than to watch my mistakes.

Game 13

G. Kasparov - V. Anand Reggio Emilia 1 991/2

French Defence

This game was played in the New Year tournament at Reggio Emilia, which at the time was the strongest tournament ever held and was the first to reach category 1 8. Nowadays this has become par for the course, but at that time it was something special.

What was also special about this event was that I was the only player in the tournament who didn't speak Russian, the other nine participants all being from the Soviet Union (after the fifth round, they were from the former Soviet Union!) - the event was effectively the last Soviet Championship.

1 e4 e6 I chose this because I dido 't want

to challenge Kasparov again in the Sicilian. He had already shown in Tilburg how well prepared he is for the Sicilian and I didn't see the point of provoking him again.

2 d4 d5 3 llld2

In Tilburg, he had started playing 3 exd5 and 4 lllf3 against the French and it was partly in the hope that he would repeat this insipid system that

I chose the opening. 3 llld2 was a bit of a surprise.

3 c5 4 exd5 1i'xd5

I had studied this line fairly extensively for the Dreev match.

5 dxc5 A very surprising move, which I

had never seen before. It seemed quite unlike Kasparov to step out of theory into unknown territory.

5 .bcS 6 lllgf3 li:16 (D)

7 .td3 Here I realized that he wanted to

get the type of set-up that often arises in the Rubinstein FreiiCh (i.e. 3 ... dxe4) - White castles long and

KAsPAROV - ANAND, REGGIO EMIUA 1 991/2 75

has a!lacking chances on the kingside, but Black has an extra centre pawn.

7 ••• 0-0 8 'l'e2 �bd7

Better than 8...�6 9 �e4 .i.e? 10 0-0, with an edge for White. Here it is better to have the knights connected, and in some lines with � and �xeS, Black can reply ... �xeS and hit the bishop on d3.

9 &4 b6 10 �xeS 16xc5 (D)

IO ... �xcS is also fine for Black, e.g. 11 i.c4 'iWfS 12 .i.e3 .i.b7 with equality. All Black's pieces are in play and the advantage of the two bishops is purely academic.

ll hJ White's idea is to put the bishop

on the long diagonal and aim for �S followed by f4. 1f he could achieve this then he would have an advantage, but there just isn't time for it. The quiet 1 1 0-0 may be better.

11 'fkc7 12 .i.d4 .i.b7 13 0-0-0

Thanks to the loss of time with .i.e3-d4, Black can safely meet 13 0-0 by 1 3 ... �c5. The exchange on f6 is not dangerous as White's pieces are not active enough to achieve anything before Black plays ... �h8 and ... l:l.g8.

13 �! (D)

14 .i.eS After 14 .i.xf6 Black has a range

of satisfactory options. 14 ... lbxd3+ 15 l:l.xd3 'lWf4+ 16 �bl 'lWxf6 is the simplest possibility, but Black can even play for the advantage with 14 ... 'lWf4+ (more accurate than the immediate 14 ... gxf6 15 1We3 �g7, although that is also playable) 1 5 �bl gxf6!? and the active queen on f4 immobilizes White's queen, while the f6-pawn prevents �e5.

14 �xd3+ 15 lbd3?!


15 'lil'xd3 was better, leading to an unclear position after 15 ... 'lil'c6 16 i.xf6 gxf6 17 Lbd4.

15 ..• 'lil'c4 16 lbd4

Other moves are not very impressive:

1 ) 16 \Pbl?! i.e4 17 J:l.e3 'ilfxe2 IS J:l.xe2 i.xf3 19 gxf3 Lbd5 with a clear endgame advantage for Black.

2) 16 i.xf6 'ilff4+ 17 IPbl 'ilfxf6 is at least equal for Black.

3) 16 Lbd2 'ilfg4! (16 ... 'ilfxa2 17 i.xf6 gxf6 IS "i'g4+ \PhS 19 'ilfh4 l:lgS 20 'ilfxf6+ J:l.g7 21 J:l.d7 is too dangerous for Black) 17 f3 'ilfg6, threatening 1S ... i.a6 or l S ... J:I.acS, again with a comfortable position for Black.

16 .•. i.e4! {D) I thought for some time about

16 ... 'ilfxa2 !?. I couldn't see anything definitely wrong with it, but it appeared more prudent to have the bishop on e4, from where it could defend the kingside. After 16 ... 'i'xa2 17 hf6 Black can play:

1 ) 17 ... 'ilfal+ (this move is definitely too risky) lS IPd2 'lil'a5+ 19 b4 'lil'xb4+ and now:

Ia) 20 �c l gxf6 21 'lil'g4+ \PhS 22 'lil'h4 (22 J:l.h3 'lil'a5t is favourable for Black) 22 ... J:I.gS 23 'lil'xf6+ J:l.g7 24 J:l.g3 with a likely draw.

lb) 20 c3 t 'lil'b2+ 2J li)c2 gxf6 22 'i'g4+ \PhS 23 'lil'h4 J:l.gS 24 'lil'xf6+ J:l.g7 25 J:l.el t with a dangerous attack.

2) 17 ... gxf6 J S li)b3 (I don't see a direct try that works, so this sensible move seems best - it stops ... eat+ and temporarily shuts out the queen ; IS 1Vg4+ \PhS 19 'ilfh4 'llfal+ 20 \Pd2 'lil'a5+ and I S J:Ia3 'llld5 19 lld l 'lil'e5 t favour Black) IS ... 'llfa4 (after I S ... IPhS 19 'lil'g4 the enemy queen is totally isolated) 19 J:l.d4! 'ifc6 20 l:l.hdl is a hard position to evaluate, but White certainly has compensation.

17 J:l.eJ "i'xa2! Not 17 ... 'lil'xe2?! IS :Xe2 i.xg2?

19 J:l.g I and White wins. 18 i.xf6

I S J:l.xe4 doesn't work because of IS ... 'ilfal+ 19 Wd2 lbxe4+ 20 'lxe4 'lil'xhl 21 'lil'g4 f6 22 lbxe6 :n defending, for example 23 i.xf6 'lxh2 24 i.xg7 J:l.xf2+ 25 Wc3 'lil'xg2 and the attack collapses.

18 ... i.g6! (D) I felt much more secure with my

bishop placed on g6, safeguarding

KASPAROV - ANAND, REGGIO EMIUA 199112 77

the kingside. 18 ... 'l'a 1 + ?! is risky: after l�d2 'l'xh 1 20 J:!.xe4 gxf6 21 l'g4t �h8 22 '1Vh4 'l'xg2 (22 . . J:tg8 23 l'xf6+ l:!.g7 24 J:!.g4 l:tag8 25 lilf3 wins) 23 l:tg4 Black has to give up his queen.

19 lW 'l'dS The point of Black's play; White

can't defend his knight because 20 .tes is met by 20 .. .f6.

zo b4?! A risky try which I hadn't really

looked at -White could have settled for sterile equality with 20 .ie5 (20 l'eS is met by 20 ... '1'xg2! and not 20 ... gxf6 21 'l'xd5 with excellent play for the pawn) 20 .. .f6 and now:

I) 21 .id6?! l:tfc8! (2 l ...'l'xd6 22 l'xe6+ 'l'xe6 23 lilxe6 is equal, while 21...l:tfe8? 22 lilb5 a6 23 lilc7 is good for White) 22 'l'xe6+ (22 c3 •xd4 23 •xe6+ .if7 24 11e7 'l'xf2 and Black wins) 22 ... '1'xe6 23 lilxe6 l:txc2+ 24 �d 1 l:tc6 and Black will be a pawn up.

2) 21 .ig3 1lxd4 22 11xe6+ with a draw.

zo ... Zl bS 22 bxg6 23 .l:lab3 24 l:lb4

gxf'6 1lxd4 bxg6 fS f4! (D)

Kasparov had placed many of his hopes on 24 .l::th4, based on the line 24 ... 1lf6 25 11e3 l:lfd8 26 11h3 �f8 27 .l::th8+ '1;e7 28 1la3+ '1;d7 29 l:!.d1 + '1;c6 30 1la4+. After 30 ... �c7 (not 30 ... '1;b7? 3 1 .l::td7+ l:lxd7 32 1lxd7+ '1;a6 33 l:txa8 and White wins) 3 1 lldxd8! (3 1 l:!.d7+ .l::txd7 32 l:!.xa8 'lfd4 is unclear) 3 1 ...l:lxd8 32 1i'xa7+ '1;c6 33 l:lh3 l:ld5 White has a dangerous attack for the pawn and can force a draw whenever he wants.

I had pinned my own hopes on 24 ... f4!, keeping the queen in the centre. My queen functions like a Dragon bishop in preventing mate at h8. Kasparov is very good at long forcing lines, but it is the nature of


such lines that one cannot he absolutely certain about them. In this case he was just wrong.

25 'llffJ? After the game Kasparov pointed

out that 25 g3 ! would have given him excellent drawing chances, for example 25 ... :1.ac8 (25 ... e5 26 .cth8+ �g7 27 :1.8h7+ �6 28 .ct.dl 'iVb4 29 'i'g4 gives White at least a draw) 26 gxf4 (D) and now:

l ) 26 ... '11ff6 (my intention during the game) 27 'lie5 (27 'ife3 :l.c5 28 'llfh3 'li'xh4 ! 29 'ii'xh4 .cth5 is also drawn) 27 ... 'llfxe5 28 fxe5 g5 29 l:h5 :.Cd8 with a near-certain draw.

2) 26 ... .ctc5 27 f5 'iff6 28 fxg6 fxg6 29 l:h8+ 'ilfxh8 30 1Vxe6+ �g7 3 1 'fre7+ (or 31 'llfd7+ �g8 with a repetition) 3 1 ... :tf7 32 1Vxf7+ �xf7 33 .ctxh8 draws.

The text-move is a mistake hecause it gives Black time to bring his rook into play and thereby gain a tempo by threatening mate on c2.

Kasparov should have abandoned his winning attempts and gone for the draw, but he decided to 'fish' for a move too long! Mter 25 'l'f3?, the game followed his analysis but the position arising favours Black rather than White.

25 ••• :.acs 26 .lhf4

Not 26 'ilfh3 'i'xf2! 27 :l.h8+ <jJg7 and Black wins after 28 'i'b6+ lt>ffi or 28 l:h7+ �f6 29 c3 'l'e3+.

26 'llfc5 27 c3 rJig7 (D)

By now Black is slightly better. 28 l:bh4

fullowing the game I pointed out that 28 .ctfh4! 1lrg5+ 29 <jJc2 'lf.st 30 'l'xf5 might have been a better chance. However, 3Q ... gxf5 (after 30 ... exf5 31 l:d4 it would be vert hard for Black to win) 3l l:a4 :c7 32 l:hal a5 33 b4 :.Cc8 34l:la3 lth8 35 bxa5 bxa5 36 J:xa5 .lllh2 is still quite promising for Black.

KASPAROV - ANAND, REGGIO EMJUA 199112 79

During the game I felt happier with queens on than in a pure rook ending and so I was pleased to see the text-move.

28 29 g3 JO 'it>c2 31 l:td4

'ife5 'ifel+ :cd8 'it'e5

Now Black has secured his extra pawn.

32 l:tbf4 'it'c7 33 'ife3 e5

Forcing an exchange of rooks without allowing the other rook to come to d4.

34 :.Xd8 l:txd8 35 l:te4 (D)

35 ••• l:tdS 36 g4?!

36 f4 ! would have been better, reducing the number of pawns and possibly exposing the black king to more checks.

36 b5 37 g,5

White's plan is to cripple Black's pawn majority, but it gives the f5-square to Black's queen.

37 1Vd6 38 f3 aS 39 'ife2 'it'e6 40 1Vb2 1Vf5

Perhaps White could have put up more resistance, but the position should be won for Black.

41 'it'gJ Or 41 'ii'h6+ �g8 42 'it>b3 l:td2

and wins. 41 'it'd7

8

42 Vel (D)

42 b4! 43 cxb4

Black also wins after 43 J:txe5 'ifa4+ 44 'it>cl bxc3! 45 bxc3 (45 lbd5 'iVai+) 45 ... 'iff4+ 46 l:te3 l:td3, so the best chance was 43 b3, which avoids an instant catastrophe. After the text-move Black decides the game by a direct attack.

43 'ifa4+

BO VISHY ANAND: MY BEST GAMES OF CHESS

44 b3 The continuations 44 �c 1 axb4

45 l:lxe5 l:ld8 and 44 �c3 1!fc6+ 45 l:lc4 axb4+ 46 �xb4 l:lb5+ 47 �c3 'i'xf3+ 48 �c2 'i'b3+ are equally hopeless for White.

44 'i'a2+ 45 �c3

45 �cl loses to45 ... axb4 46 'l'xb4 'i'al+ 47 '.tc2 'i'dl+.

45 46 bxa4 47 �c2 48 'ili>c3 49 �c2

0-1

I started Reggio Emilia with my traditional two wins. In the next round I drew with Khalif man but lost in round 4, which allowed everybody to catch up with me. The rest of the tournament was a race between Kasparov, Gel· fand and myself and in the end I finished half a point ahead of them - my greatest success up to that point, and even today I would consider it oneofmy best results.

Game 14

V. Anand - E. Bareev Dortmund 1 992

French Defence

This was my best game from Dortmund 1992. I also won a nice game against Hiibner (see the following game), but I particularly like this one as it is a kind of model game for dark-squared play against the French.

1 e4 e6 Bareev's favourite defence is the

French. Although he has also experimented with other lines, all our encounters in which I was White have been French Defences.

2 d4 dS 3 lbc3 lLir6 4 eS lill'd7 s f4 cS 6 m lbc6 7 .i.e3 a6 8 .. dl bS 9 dxcS .i.xcS

10 .i.xcS lbxcS 11 .. fl ..,6 12 .td3 .l:b8 13 0.0 lbb4

This is all fairly standard stuff in this opening. Black must develop play on the queenside, whilst White tries to play on the kingside.

14 .l:fdl 0.0 After 14 ... lba4 15 lbxa4 bxa4 16

b3 Black isn't doing too badly, but he

has a long evening ahead of him as White tries to exploit the d4-square and Black's bad bishop. However, heading for an ending may be Black's best chance once he has decided to play ... .l:b8 and ... ltlb4. My personal view is that if Black wants to keep the queens on, he should adopt the plan with ... b4, ... aS and ... .i.a6.

15 lbe2! .i.d7 The option of ... ltla4 has already

gone, as now it would just be punching thin air.

16 ltled4 (D)

16 ••• lbbxd3 l 6 ... ltle4 is bad in view of 17

.i.xe4 dxe4 18 ltlg5, but 16 .. . a5 is a


possibility. From Black's point of view there is no advantage to this move, because White can just reply 17 b3 and Black has nothing better than 17 . . . tllbxd3 1 8 cxd3 b4, transposing into the note to Black's 17th move. However, White can try to refute 16 . . . a5 out of hand by 17 .i.xh7+!?. During the game I looked at some of the following lines; the analysis is quite interesting, but the tree of variations is so large that I will only give the basic variations: 17...�xh7 1 8 1i"h4+ �g8 19 tllg5 J:.fc8 and now:

I ) 20 � ..te8 21 �h l (21 11rh7+ Wf8 22 f5 exf5 23 tllxf5? tllcd3+! mates) 21...tlle4 22 "ffh7+ �f8 23 tllxe4 dxe4 24 f5 llld5 (the knight comes back to aid the defence) 25 "ifh8+ �e7 26 "ifxg7 �d8 27 fxe6 fxe6 28 11t"g8 �e7 29 "ifg5+ �d7 30 "ifg7+ �d8 is unclear.

2) 20 �hi !? (D) and now there are three possible lines:

2a) 20 ... lbc:d3? 21 c3 lUc2 22 "ifh7+ �f8 23 Uxd3 tllxal 24 'i'h8+ �7 25 'it'xg7 J:.f8 26 f5 exf5 27 e6 .i.xe6 28 Ue3 winning.

2b) 20 ... �f8 21 11fh8+ �e7 22 "ifxg7 J:.f8! (22 ... .i.e8 loses to 23 15, while 22 ... �d8 23 11t"xf7 clearly favours White) 23 tlldO ! ! (threatening 24 tllh4; 23 a4 is less effective - the idea is to play :a3-h3, but 23 ... � is quite an annoying reply) 23 ... � (23...lLlxc2 24 l:lacl lLle3 25 � with a crushing attack) 24 li)xe4dxl4 25 li)h4! Ufc8 26 "iff6+ WeB 27 "ifh8+ �e7 28 lLlg6+ fxg6 29 'i'g7+ �e8 30 Uxd7 and White wins.

2c) 20 ... lt:le4! (best) 21 11ih7+ � 22 "ffh8+ (22 lLlxe4 dxe4 23 .!ilf5 exf5 24 Uxd7 'ilh6 wins fer Black) 22 ... �7 23 "ifxg7 (D) and now:

2c I ) 23 ... lLlf2+ 24 �gl tzlxdl 25 "iff6+ �e8 26 Uxdl Uxc2 (26 . .. :C4 27 c3 is hopeless) 27 li)xf7 llc6 28 f5 ..tc8 29 "ifh8+ �7 (29 ... �xmo f6 wins) 30 li)g5 �c7 3 1 'i'h7+ id7

ANAND - BAREEV, DORTMUND 1992 83

32 fxe6 l:d8 33 'l'e7! and White ends up well ahead on material.

2c2) 23...ltlxg5 24 fxg5 (if 24 'l'xg5+, then 24 .. .'.ate8) 24 .. J:tg8 25 'l'f6+ and now 25... <.ate& 26 l:l.fl 'l'xd4 27 'l'xf7+ <.atd8 28 'il'xg8+ is good for White, but 25 ... <.atf8! is unclear.

2c3) 23 ... l:l.f8! (this defence appem to hold out for Black) 24 c3 (24 lbxe4 dxe4 25 c3 ltld5) 24 ... ltlf2+ 25 ¢'gl ltlxdl 26 ltlh7 and now 26 ... :g8 27 'il'f6+ <.ate8 28 'il'h4 is probably a draw, while 26 ... ltlxc3 !? 27 bxc3 'il'c5 28 'iWg5+ <.ate8 29 lQf6+ �d8 is unclear.

One can see why Bareev decided not to invest a lot of time in working through these variations and opted foc the simpler text-move.

17 c:xd3 ltla4? A serious error - Black shouldn't

allow White to fix the enemy pawns m light squares. Black should play either 17 ... b4 or 17 ... a5, which gives him chances of eventually activating his bishop at b5. After 1 7 ... a5 18 b3, fer example, White is just slightly better.

18 b4! (D) At first it seems that Black's

knight can reach a good square by ... �a4-c3, but in fact it is not very effective at c3 because it lacks adequate suppon. By contrast, White's knights, after ltlb3 and ltlfd4, have well-supported and useful squares to land on at d4 and c5. Alternatively,

White can use the fact that Black has no queenside counterplay to start kingside operations.

8

18 aS 19 a3 l:l.rc8 20 l:l.dcl axb4 21 axb4 l:l.xcl+

After 2 1 ...l:l.c7 22 ltlb3 l:l.xcl + 23 l:l.xc I White retains a clear advantage. The position is very pleasant for White because it doesn't matter whether he swaps rooks, queens, or both - his advantage persists in any case. The plan of h4-h5 followed by g4 and f5 is effective both in the middlegame and in the endgame.

22 l:l.xcl l:l.c8 23 l:txc8+ .i.xc8

White could play for the exchange of queens, but keeping queens on the board allows him the option of starting an attack using his space advantage on the kingside.

24 1Wc2 25 �f2

.i.d7 �


26 g4 �e8 27 �e3 (D)

27 �g3 !? might have been more precise - White's king is safer on g3, and the king would not obstruct White's kingside attack. I intended to bring the king to d2 to complete the domination of the a4-knight, but this is not necessary and in fact allows Black some slight counterplay.

B

27 f6 28 b4

White could have played more directly by 28 exf6 gxf6 29 gS fxgS 30 �xgS 11t'd6 (the attack on b4 and the threat to play ... eS look awk-ward, but...) 31 11fe2! ( ... White sirnply abandons the b-pawn) 3L.11fxb4 32 11fh5+ �d8 33 �gxe6+ �c8 34 11fe5 'ifel + 35 �2 .1xe6 36 11fxe6+ �c7 37 11fe5+ �b6 38 fS and Black's pieces are practically powerless to stop the f -pawn.

When you have a grip such as White has in this game, it is not easy

to decide on the right momentto convert the advantage. Once you have opened the position you can't close it again, so a heavy commitment is involved in playing a line such as the above. It mustn't be too early but you must also avoid enjoying your grip for too long and letting Black gradually free himself.

During the game, I thought that exf6 and gS was still premature, but it turns out to be possible fa- tactical reasons. Still, I prefer the game continuation; as yet there is no need to hurry. After the text-move White clearly threatens exf6 and gS, because ... fxgS can be met by hxgS.

28 _ 'l'b8 After 28 .. /�f7? White just con

tinues with his plan by 29 exf6 gxf6 30 g5.

29 'ifcl (D) Now 29 exf6? gxf6 30 gS allows

30 ... e5!.

8

29

ANAND - BAREEV, DOKI'MUND 1992 85

30 li>e2 Imprecise. 30 li>f2!, heading for

g2 or g3, would have been more efficient

JO ... �f7 31 1i"e3 'iVf8

After 3I...l!X:3+ White has to take some care: 32 �d2 would allow a measure of counterplay by 32 ... 'flc7! (but not 32...lila2? 33 lilc2! "ilc7 34 l'd4 and the a2-knight is trapped), but 32 li>CI ! is very good for White after 32 .. .'1"a7 33 'iVel or 32 ... 'lic7 33 !5.

32 f5 'fle8 33 &5 (D)

Not 33 exf6 gxf6 34 fxe6+?! .he6 35 lilxe6 'i'xe6 36 'i'xe6+ �xe6 37 o&l4+ �e5 38 �b5 \i>r4 with a likely draw. When you have such an advantage, simplifying everything just to win a pawn is insane. However, White could also have maintained the pressure by 33 �fl, &epping out of the way of some checks.

B

33 ••• exf5 34 gxf6 gxf6 35 1!rh6 fxeS

After 35 . . . 'i'h8 White wins by 36 e6+ i.xe6 37 lbe5+!.

36 1Wxh7+ �6 37 '16116+ <M7 38 il:lgS+ �e7 39 1Wg7+ �d6 40 il:lf7+ 'l;c7 41 il:lxeS (D)

White has established complete dominance and should win without difficulty. On top of all his other advantages, he has an outside passed pawn.

41 •••

42 il:lxbS+ 43 il:ld4 44 il:ldf3 45 �2 46 ile7

lLlb6 �b8 'l"bS+ i.e8 f4

Threatening 4 7 b5 .i.x b5 (or else lLlc6+) 48 ilc5.

46 lLlc8


47 'i'f6 48 'iVxf4 49 lDg5 so �g3 51 'i'f3 52 lDgf7

�b7 'l!i'h8 'i'h6 lDd6 �c6 lD:d7

If 52 ... 1Wg7+, then 53 'M2. 53 'lfxf7+ �b6 54 11f4 ire6 55 hS �b7 56 ST6! t-O

A neat finish.

Bareev subsequently recovered and had a very good tournament. The above game was played in round three. Over the next few rounds I had a number of sharp draws, including a very exciting one against Shirov,and then in round seven I faced Hubner.

Game 15

V. Anand - R. Hubner Dortmund 1 992

Petroff Defence

1 e4 2 ltlf3

eS ll:\f6

Round !bout this time I was very well prepared against the Petroff, beca� I had played the opening myself fa-many years, so I was quite happy with HUbner's choice. I had many interesting ideas stored up for White and was hoping that I would be able to use one of them.

3 d4 4 .td3 5 ll:\xeS 6 li:Jxd7 7 0-0 8 c4 9 c5

ll:\xe4 dS ll:\d7 �xd7 11'h4 0-0-0 g6

An unusual move instead of the normal 9 ... g5. HUbner had played it once before, in 1983, but I had never

1 2 b4 ll:lh5 1 3 b5 and White gained the advantage, but presumably there was an improvement somewhere.

11 •.• 11'f6 Perhaps this is the point behind

9 ... g6; with the pawn on g5 this retreat would not be possible.

12 �e3 (D)

seriously looked at it. One reason was that Hiibner had annotated the earlier game in lnformator, and had spent about two pages explaining why the move was really lousy!

For the moment I decided to continue as if Black had played 9 . .. g5.

10 lOc3 .tg7 ll g3

The earlier game, Timman-HUbner, Til burg 1983,had gone 1 1 ll:\e2 li:Jf6

12 �fS? Black should have attempted to

make use of the fact that g5 is free by playing 12 ... ll:\g5!?. Then White can try:

1 ) 13 ll:lxd5 ll:lh3+ (if 13.. .�c6, then 14 1l'g4+ wins) 14�g2 ( 14 �J �c6 15 11'g4+ 11'e6 and Black wins) 14...�c6 15 11'g4+ (15 �xh3 �xd5 leaves the king miserably placed)

BB V/SHY ANAND: MY BEST GAMES OF CHESS

15 ... �b8 16 11'xh3 :xd5 and Black is much better.

2) 1 3 �e2 �h3 14 :e1 l0e6 is unclear.

3) 1 3 f4 and now: 3a) 13 . . . ll:le6? 1 4 lt:lxd5 wins. 3b) 13 ... lt:lh3+ (a difficult move to

make; of course the knight is rather annoying for White, but it has no way out) 14 �g2 :he8 (14 ... h5? 15 f5! is very good for White) 15 ... d2 with an unclear and double-edged position.

3c) 13 ...... e6! 14 lle1 ll:le4 15 �xe4 dxe4 16 d5 is another unclear line. Black can try to develop lightsquared counterplay, while White can use his pawn-mass for attacking purposes.

The text-move is a clear mistake. I didn't realize this immediately, because I was trying to imagine what this position would be like with the pawn on g5. Suddenly I noticed that lt:lb5 was very strong, precisely because with the pawn on g6 my bishop has access to f4.

13 lt:lbS! �h3 After 13 ... a6 White continues 14

lt:lxc7! �xc7 (14 . . .... c6 15 lt:la8! and the knight escapes via b6) 15 �4+ �c8 16 �e5 ... c6 (16 ... 11'e6 is the same) 17 �xg7 llhg8 1 8 �e5 f6 19 �f4 g5 20 �e3 with a safe exira pawn and an attack by b4-b5.

14 lt:lxa7+! �b8 15 lt:lbS �xfl 16 �xfl (D)

I didn't think long about this exchange sacrifice or even bother to calculate variations; it's clear that without the black a-pawn, White's own a-pawn can just run all the way through.

8

16 More or less forced, to meet if4

by ... :e7. 17 ... a4 ... a6 18 W'b4 ... aS

White wins after 1 8 ... 11fc6 19'ila5 f5 ( ...... a6 is no longer possible be cause c7 is also attacked) 20 if 4 :e7 21 lt:lxc7 :Xc7 22 �b5.

19 a4! Intending to tuck in her majesty

with a5-a6. 19 g5

Black even has to spend a tempo preventing �f4+ before he can play . . .c6.

20 21 22

aS lt:lc3 bxc3 (D)

c6 lt:lxc3

ANAND - HOBNER, DOKTMUND 1992 89

Black is dead lost 22 h6 23 a6 rs 24 ..th3 :rs 25 a7+ �c8 26 '6bl

Forcing the g-pawn forwards and so gaining f4 for the dark-squared bishop.

26 g4 27 w �d7 28 ..td3 �e6 29 ..tf4 l:tf7

I was tempted to toss in ..tb8, but it isn't rea! I y necessary!

30 'IVcl ..trs 31 '1Ve2+ 1-0

White finishes off by 31 ... �6 32 'ife5+ �g6 33 'ife6+ llf6 (33...'it>g7 34 ..te5+ 'it>g8 35 hf5 l:te8 36 ..th7+! mates) 34 ..txf5+ 'it>g7 35 ..te5 and mate in three more moves.

A nice crisp win, though Black's 12th move made it easy for me.

The above two games were my only two wins from Dortmund and I finished fourth with a score of 5/9 - not one of my most memorable results.

In July 1992 we had the GMA 'farewell party'. The organization had fallen into difficult times and with the failure of the second World Cup cycle it was recognized that its tournament-organizing days were over. SWIFT sponsored a final rapid-play knock-out event held in Brussels. It was superly organized.

Game 16

V. Anand - I . Sokolov Brussels SWIFT rapid 1 992

Sicil ian, Scheveningen

This was the second game of a twogame mini-match. I had won the first game and so only needed a draw to go through to the next round.

1 e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4 a6 5 �c3 d6 6 a4 �6 7 .te2 �bd7

I found this move a bit strange, but 7 ... �6 would have transposed into one of the main lines of the Scheveningen and I suppose Sokolov wanted something a bit sharper.

8 0-0 �c5 9 .tr3 .te7

10 g3 0-0 11 .tg2

In a sense White has lost time with this bishop manoeuvre, but on the other hand the e4-pawn is welldefended and so the knight on c5 isn't doing much. Moreover, Black cannot play ... b6 easily and ... .td7 is met by b4, so he is going to have a lot of trouble developing his queenside.

11 'ifc7

12 .te3 I think that White already has a

significant advantage. 12 ••• J:lb8 13 f4 J:le8

Here I was about to play the standard plan g4-g5, when I noticed that it wasn't reaDy necessary as White can break through right away.

14 e5 dxe5 Forced, since 14 ... �fd7 15 exd6

.txd6 (or 15 ... 'li'xd6 16 b4) 16 1txib5 wins.

15 fxe5 tiJrd7 (D) 15 ... 'li'xe5 1oses to 16 .tf4.

16 l:.xf7! A very common tactic when the

rook is on e 8 and the f-file opens.

ANAND - I. SOKOWV, BRUSSE� SWIFT RAPID 1992 91

16 ... �xf7 If Black tries to decline the sacri

fice by 16...l0xe5 then 17 ..tf4 'llfd6 (17...�xf7 1 8 'iflh5+ �f8 19 �xe5 wins) and now:

I) 18�hl l0cd7 19 l:txe7 and after 19 ... lhe7 20 l0f3 'llfxdl + 21 .:txdl ilhl3 22 ..ixb8 .l:lf7 or 19 ... '1i'xe7 20 'l'e2 �c6 21 l0xc6 bxc6 22 �xb8 lill.b8 White has a positional advantage.

2) 18 .l:lxe 7 ! (an even stronger move) 18 ... %be7 19 l0db5 'li'xdl+ :ll niDI axb5 (20 .. .lllcd7 21 �xe5 lill.e5 22 :!d8+ <J;f7 23 l0d6+ wins) 21 i.xe5 l:ta8 22 �d6 .l:ld7 23 �xc5 lhdl + 24 lllxdl .l:lxa4 25 l0e3 .l:la2 26 .1d4 with a winning endgame.

After the text-move the attack crashes home.

17. 'i'h5+ �f8

18 .l:lfl+ lLlr6 19 exf6 ..txf6 20 tOdbS axbS 21 tOxbS

21 .1xc5+ .l:le7 22 l0xb5 also wins.

21 'i'd7 (D)

22 'li'xb7 A nice move. White can take the

c5-knight with check in two different ways, but ignores it. White threatens 23 .l:lxf6+ gxf6 24 �h6+, so Black has to waste more time.

22 'li'e7 23 .l:lxf6+ 'i'xf6 24 ..ixcS+ .l:te7 25 1Vh8+ �f7 26 �d6+ 1-0

I was eliminated by Adams in the semi-finals, so the above game was the highlight of the tournament for me.

In September I played a match with Ivanchuk in Linares. It was not part of any cycle, but simply a one-off event arranged by Rentero (the organizer of the Linares tournament) who wanted to hold an event in addition to the annual super-tournament. Ivanchuk and I were his first choice for the players, and we each had our own reasons for wanting to participate. Ivanchuk wanted to erase the memory of the loss to Yusupov in the Candidates quarter-finals in Brussels, and I wanted to keep my match experience going between world championship cycles. Subsequently, he organized two more matches: Ljubojevic-IIIescas and Lautier-Karpov.

Game 17

V. lvanchuk - V. Anand Match (1), Linares 1 992

Sici l ian, Richter-Ra uzer

1 e4 c5 2 lin d6 3 d4 cxd4 4 lL!xd4 lL!f6 5 ltJc3 lL!c6 6 .i.g5 e6 7 '6'd2 a6 8 0-0-0 h6

I had prepared this variation specially for the match. I hadn't played it before so I could hardly imagine Ivanchuk having prepared it deeply. Nevertheless, lvanchuk blitzed out his next few moves and in fact played the whole game at high speed!

9 .i.e3 lL!xd4 10 hd4 b5 ll f3

I I �bl is more promising. lvanchuk played it in the fifth game of the match and gained the advantage, although the game finally ended in a draw.

11 ..• '6'a5 12 &3

A new and quite good move. 12 �b1 was played in earlier games, while 12 'ilff2 b4 13 .i.b6 'ilfgS+ 14 i.e3 'ilfa5 is a popular variation for players who feel like an early dinner!

12 13 i.eJ 14 �b1

15 g4?!

eS .i.e6 i.e7 (D)

Playing lL!dS is an option which is available to White at virtually every move. However, without any knights White can hardly expect to oo any· thing against Black's slightly weak· ened queenside, so playing lLldS is an admission that White can no longer hope to gain the advantage.

11te problem with Ivanchuk's move is that he is soon forced to play 00 in any case, when the move g4 not only fails to benefit White but can even prove a weakening of his king· side.

IVANCHUK - ANAND, MATCH (1), LINARES 1992 93

After 15 h4 llb8 l6 lDd5 'I' xd2 17 /bxf6+ gxf6 18 llxd2 f5 the consequenceswould not be so serious for White as h4 does much less damage to White's kingside.

15 .•• llb8 15 ... b4!? was an interesting alter

native: I) 16 lD<l5 �xd5 17 exd5 llb8 is

unclear. 2) 16 axb4 'llfxb4 17 ttxl5 l0xd5

18 eXd5 'l'xd2 19 llxd2 �d7 with a roughly equal potision.

3) 16 li::la2 d5 17 axb4 'llfc7 with compensation for the pawn.

I wasn't feeling quite awake at this point and so I played the 'solid' 15 ... :0b8.

16 ll:ldS 'ifxd2 (D)

17 �xf'6+?

In conjuntion with 15 g4 this is a terrible move. I suspect Ivanchuk hadn't woken up either! He should have continued 17 llxd2, but after 17.. .�xd5 18 exd5 �d7 Black is

slightly better, because he has the .. .f5 break, whereas White has no comparable play on the queenside.

After the text-move I was about to make the routine capture 17 ... �xf6 (when White could perhaps be a little better after 18 llxd2 �e7 19 h4) when suddenly I noticed that 17 ... gxf6 might be a good move. After about ten minutes' thought, I decided to play it.

17 gxf6! ! Now w e were both wide awake!

18 llxd2 hS! (D)

At first glance White is better, or at least not worse, in view of Black's damaged pawn slrUcture. However, White is actually seriously worse. If White could consolidate his kingside pawn slrUcture by h3 then he would indeed be better, but just at the moment this is impossible.

19 llg1 Also after 1 9 �e2 hxg4 20 fxg4

llh3 Black's rook takes up residence


at a most inconvenient location for White.

19 20 fxg4

bxg4 .i.c4! ! (D)

Just in time to stop White setting up some sort of fortress on the kingside by h3 and l:tg3. Black's 'bad' e7-bishop will protect his pawns while he forces pawn exchanges eventually leading to connected central passed pawns.

21 b3 After 21 .i.xc4 bxc4 White again

has no time to consolidate his kingside by 22 l:tg3 (and 23 h3) because of 22 ... c3. If 22 l:td5, then 22 . .. l:tb5 and again Black is better.

21 ... .txn 22 l:txfi l:th3

Black appears to have committed a whole list of positional sins: allowing doubled f-pawns, giving White an outside passed h-pawn and exchanging his 'good' bishop with ... .i.c4; yet he is better. Paradoxical?

Yes, but this does not mean that the old positional rules have been suspended for the course of this game. Black's play depends on two things. First of all, his long-term aim is to exchange his d-pawn for White's epawn by ... d5 and to exchange his f6-pawn for White's g-pawn (either by ... f5 or by forcing White to play g5). Then he will be left witb two connected central passed pawns, supported by his king, whereas White will have pawns on c2 and h2 that aren't going anywhere. Secondly, he can only put his plan into action because he has the initiative, and especially as the rook on h3 disrupts White's whole position and leaves both g- and h-pawns vulnerable to attack. Had Black wasted even one move, White would have fortified his kingside and the old positional values would have reasserted themselves.

The text-move is more acurate than 22 .. .'�'d7, not because of 23 l:tr3 , when 23 ... :h4 24 h3 l:tbh8 wins a pawn, but owing to 23 g5!. which confuses the issue. Then after 23 ... fxg5 24 l:txf7 Black' overall plan has been disrupted, while 23...�e6 24 gxf6 .i.xf6 25 l:txd6+! �xd6 26 l:txf6+ �e7 27 �g5, fol· lowed by h4, gives White plenty of counterplay.

23 :e2 23 .i.g l would have been a bener

chance, but Black maintains the

IVANCHUK - ANAND, MATCH ( 1), LINARES 1992 95

advantage by 23 ... �d7 24 l:td3 l:th4! (24 . ..lhd3 25 cxd3 l:th8 26 l:tf3 d5 is unclear) 25 g3 (25 h3 l:tbh8 26 :m fS! 27 :Xf5 lhh3 28 l:txf7 :Xd3 29 cxd3 :hi 30 l:tfl ..i.g5 wins material,as in line below) 25 ... l:tg8 26 h3 J:lgh8 27 l:tff3 f5 ! and now White cannot maintain his fortress:

I) 28 J:lxf5 ltxh3 29 lhh3 l:txh3 30 i.f2 (30 lbfn l:th 1 3 1 :n ..i.g5 32 J:lel .li.d2 is winning for Black) 30 . .. >Pe6 and White's pawns on e4 and g4 are so weak that he might easily lose both of them.

2) 28 exf5 e4 29 l:te3 (29 ltc3? il.f6) 29 ... d5 and Black has strong pressure.

23 ... �d7 24 g5 (D)

24 ... �e6 Now this is good as White does

not have an exchange sacrifice (see note to Black's 22nd move).

25 gxf6 ..bf6 26 .td2

26 l:tef2 makes no sense as after 26 ... ..i.e7 White has to attend to his attacked bishop.

26 .•• ..i.e7! Simplest and best. The alterna

tives 26 .. ih4 27 ..i.b4 and 26.'.Jlg8 27 l:tef2 .i.e? 28 l:txf7 l:tg4 allow White more counter-chances.

27 ..i.el f6 28 ..i.g3

White has finally defended the weak h2-pawn, but Black has time for . . . d5.

28 ... dS 28 ... l:td8!? was also possible, al

though in this case Black would have to worry about 29 c4. I preferred to play ... d5 immediately.

29 exd5+ �xdS (D)

30 lUS! A good defence, forcing Black to

lose some time. 30 ... �c6

A forced move, as 30 ... �e6? 3 1 ..i.xe5 l:te8 3 2 lhf6+ and 30 ... J:lb7?


3 1 �xe5 �e6 (3 I .. .fxe5 32 l:tfxe5+ �d6 33 l:te6+ 'iti>d5 34 .l:xe7 is no better) 32 �xf6+ �xf5 33 �xe7 would lead to a draw.

31 l:tef2? 3 1 l:tf3 ! would have made life far

harder for Black as White threatens both 32 �xe5 and 32 .l:c3+ followed by 33 he5. There would be nothing better than 3 I ...l:th7 32 l:tc3+ �b7 (not 32...�d7 33 l:td2+ �6 34 .l:cd3 :l.b6 35 �f2 l:tc6 36 ..ltc5), but this would represent a success for White. The black king belongs on e6 and while the connected passed pawns guarantee Black an advantage, he would have a hard technical task ahead.

After the text-move White has a large disadvantage.

31 ... l:th6 (D) Not 3 1 ...�d5 32 l:txf6 �xf6 33

l:txf6 and White has complicated matters. The text-move prevents any sacrifice on f6.

32 �b2 33 l:te2 34 lU3

'iti>d7 �d6 llc8!

Very precise. 34 ... �e6 would al· low 35 l:tc3, intending llc6, and again White has some counterplay.

35 �el �e6 Mission accomplished!

36 l:td3 (D)

36 l:th7 37 l:tg3 �

Black doesn't even have to push the pawns immediately. He can play to improve the position of his pieces, or try to exchange a pair of rooks to reduce the chances d a blockade.

38 � l:td7 39 :cJ :CC7

39 ... l:td 1 ? 40 �f2 �xf2 41 :.XeS �d4 doesn't mate after 42 c3.

40 h4 l:tdl 41 .tf2 .td6 42 l:tg3 e4!

After all the fuss about Black's connected passed pawns, he gives

lvANCHUK - ANAND, MATCH (1), LINARES 1992 97

one of them up! However, it does win the exchange.

43 lbe4+ 43 llgl llxgl 44 i.xgl f5 wins

easily. 43 i.eS

44 lbe5+ Or 44 c3 :d2+ 45 �bl llxf2 win

ning. 44 rxeS 45 �b2 lld2

0-1

This match was an important point in my career as I gained a great deal of confidence as a result of beating Ivanchuk 5-3. He was the first really strong opponent I had beaten in a match and I took this as a promising sign for the next world championship cycle.

In November I took part in a very strong (and enjoyable) tournament in Moscow. I think it was only the second tournament to reach category 18, although this had been achieved by having only eight players. I lost the first game to Gelfand after allowing myself to be swindled in a favourable position, and after losing the first game in a seven-round event I couldn't really expect thatl would win. However, Moscow turned out to be one of my more succesful tournaments. I scored 4•h out of my remaining six games and shared first place with Gelfand. The following game is from the second round and was the first step towards recovering from my first-round defeat

Game 18

V. Anand - G. Kamsky Alekhine Memorial, Moscow 1 992

Ruy Lopez, Worral l Attack

1 e4 eS 2 ttJf3 �c6 3 .i.bS a6 4 .i.a4 ttJf6 5 0-0 .i.e7 6 'ife2

I played this move because I had beaten Kamsky with the same move the previous year in Tilburg, and I just decided to repeat it.

6 bS 7 .i.b3 0-0 8 c3 d6 9 d4 .i.g4

10 .l:l.d1 exd4 1 1 cxd4 dS 12 eS ltJe4 13 a4 (D)

A couple of months before Short's semi-final Candidates match vs Karpov (which was in April 1992), he asked if I would like to come to Athens to work with him. This visit was just a one-off arrangement and I wasn't Nigel's second. We analysed the 'ire2 Ruy Lopez during the week I stayed there, but I had no part in some of the other openings Nigel played in the Karpov match.

A strange coincidence occurred in the Linares tournament shortly

before the Short-Karpov match. I played the Queen's Gambit Accepted against Karpov, tried out something very unusual and drew without any difficulty. I had no idea that Nigel had prepared the QGA fer Karpov until Nigel revealed the fact after my game!

8

13 ••• l:tb8 One of the points of 13 a4 is the

line Short showed against Karpov: 13 . . . bxa4 14 .i.xa4 �b4 15 h3 ih5 16 �c3 .i.g6 17 .i.e3 .l:l.b8 18 �a2 This last move is one of the discoveries we made during our anaytical week in Athens: it is much more important to eliminate the knight on b4 than the one on e4. The reason is that

ANAND - KAMSKY, MOSCOW 1992 99

the knight on e4 is only potentially threatening (e.g if Black plays . . . c5 and . . . cxd4) while the knight on b4 is critica; ,as it stops White invading on c6.

Karnsky, not surprisingly, does not repeat the line that Jed to a Joss for Karpov

14 axbS axbS 15 b3 ..thS 16 ..te3 (D)

16 ••• 'l'd7 Not 16 ... lllg5 17 ..txg5 ..txg5 1 8

�3 �b4 1 9 g4, winning the d5-pawn

i7 .l:cl 17 liX:3 would have been interest

ing, e.g. 17 .. .lbb4 !? ( 1 7 ... lllxc3 1 8 bxc3 b4 19 c4 dxc4 20 ..txc4 is unclear) 18 .1f4 'l'f5 (18 ... ..tg6!? is also possible )and now both 19 lllxd5

llld5 20 ..txd5 lllxf2 and 19 'l'e3 lll xc3 20 bxc3 lllc2 21 ..txc2 'l'xc2 are utnclear.

17 .l:b6

Supporting the weak third rank and the c6-knight in particular. 17 ... llld8 was an alternative.

18 lllcJ lllb4 After 18 ... lllxc3 19 llxc3 White

has a small advantage. 19 l:ta7 lllgS

Again the correct choice. After 19 ... .1:g6 ( 1 9 ... lllc6? loses after 20 lllxd5 lllxa 7 2 1 .l:xc7) White replies simply 20 �h2 and Black has to deal with the threats of 21 'l'xb5, 2 1 lllxb5 and 2 1 .l:xc7 (20 ... lllc6 fails to 2 1 lllxd5 !).

20 ..bgS ..txgS 21 l:tca1?! (D)

I intended the exchange sacrifice given in the next note, but I had miscalculated one variation, so the solid 21 .l:d1 would have been better.

B

21 ..tf4? 21 ... lllc6! was the right move: 1 ) 22 -Ua8? lllxd4 23 .l:xf8+ �xf8

24 l:ta8+ ..td8! and Black wins, but not 24 .. .'�e7? 25 lllxd4 ..txe2 26


ll:lxd5+ 'l'xd5 ZJ .txd5 .tc4 28 .tc6! winning.

2) 22 �xd5 �xd4! (this is the move I had missed; after 22 ... �xa7 23 .l:l.xa7 White has good compensalion for the exchange) 23 'ilfd3 �+ 24 gxf3 'ilfxh3 and Black is better.

3) 22 'l'd3 .txf3 (22...�xa7 23 ..l:l.xa7 wins the d-pawn and is promising for White) 23 1fxf3 �d4 24 'l'xd5 'l'xd5 25 .1xd5 c6 26 .l:l.dl (26 .te4 is also equal) 26 . . . .tf4 27 l:xd4 .txe5 leads to a draw.

22 IL!xd5! This gives White a large advan

tage in every line. 22 ••• �xdS 23 'l'e4 (D)

.txf3?! Now White is winning. Black

could still have continued the f_\ght by 23 ... 1Lle7! 24 'ilrxf4 lLJg6 25 '1Ve3 .txf3 26'i'xf3 'i'xd4 27 .l:l.xc7 1Llxe5, but now either 28 .l:l.xf7 IL!xf3+ 29 l:.xf3+ 'iic4 30 .l:l.c3 or 28 .txf7+

Wh8 29 'iie3 11rxe3 30 fxe3 gives White a clear extra pawn.

w

24 11rxf3 .txeS 25 dxeS liJb4 (D)

26 ... e3?! After the game Gelfand pointed

out a much simpler (and quicker!) win by 26 ..l:l.aS! ..l:l.bbS (or 26 . ..1ilc6 27 e6) 27 .l:l.xbS lhbS 28 la8!.

26 ••• 1fc8 Or 26 ...... d3 27 .txf1+ lxn 28

.l:l.a8+ .l:l.f8 29 1hf8+ Wxf8 30 'lc5+ and White wins.

27 'li'e4 Once again missing an esdier win,

this time by 27 e6 fxe6 28 .l:l.cl, for example 28 ••. .l:l.f7 29 'l'xb6, 28 ... /ila6 29 'iixb6 cxb6 30 .l:l.xc8 1Jtc8 31 .txe6+ or 28 ... 1Lld5 29 .txd5 exd5 30 .I:I.cxc7.

27 ••• tba6 After 27 ... 1Llc6 28 .l:l.a8 'l'xa& 29

.l:l.xa8 .l:l.xa8 30 e6 fxe6 31 .ixe6+ Wh8 32 .td5 White wins material.

28 .tdS c6

ANAND - KAMSKY, MOSCOW 1 992 101

On 28 ... ltXS, 29 Wb4 lhd7 30 J:.aS finishes Black off.

29 'i'e3 l:.b7 More or less forced, because af

ter 29 ... c5 White's bishop is absolutely dominant, but now White has a forced win.

JO hf7+ l:.bxf7 31 J:.7xa6 l:.xf2 32 e6 (D) 32 .••

33 e7 34 l:.a8 35 �h2

l:.xb2 l:.e8 l:.b1+

White must still be careful. After 35 lhbl? Black would slip out by 35 . ..'ii'xa8 36 'i'e6+ �hS 37 l:.fl h6 38 'iWd7 �g8.

35 'i'c7+ 36 g3 l:.b2+ 37 �g1 'i'd7 38 lhe8+ 1-0

Since 38 ... 'il'xe8 39 'il'e6+ �h8 40 l:.fl leaves Black defenceless.

Game 19

V. Anand - V. lvanchuk Linares 1 993 Petroff Defence

I e4 eS 2 �fJ ll'lti 3 d4 �xe4 4 .td3 ciS S �xeS �d7 6 �xd7 .txd7 7 0-0 'ft4 8 c4 0-0-0 9 c5 g5

10 � .tg7 ll g3 'ft3 12 �xe4 dxe4 13 .txe4 .tbS (D)

This position had been reached umpteen times before the game and the path to equality for Black had been more or Jess worked out after 14 11'b3 and 14 .txg5, but I had a

new idea that I had been waiting to use.

14 .i.g2! After 14 .hg5 lhd4 15 .i.g2 •rs

16 .b3 c6 17 .te3 .hfl 18 bfl, Black can continue 18 ... .1:.hd8! 19 'ifa3 .l:ldl 20 11'xa7 l:bfl+ 21 .hfl .l:ld1 22 • a8+ Wc7 23 .i.d2 'I' xeS 24 .ta5+ b6 25 11'a7+ rJid6 26 hb6 ilb5 27 11rc7+ rJie6 28 Wc8+ � witb a draw.

14 ••• 'ilffS Ivanchuk looked surprised, be

cause it was well known that 15 dS is bad, but this was not White's idea.

IS .i.e3! (D) 15 'irb3 c6 leads nowhere.

IS •••

ANAND - IVANCHUK, LINARES /993 103

16 .btl White intends 'iia4 followed by

ldl-d3, swinging the major pieces intothequeenside attack. 1 6 �xfl ?! is inferior as the king is exposed on n.

16 l:the8 Black has several alternatives: 1) 16 ... i.xd4 (the tactics don't

work for Black) 17 .bd4 'l'e4 18 1xh8lb.d1 19 l:l.xdl with too much for the queen.

2) 16 ... .!:.xd4 17 i.xd4 l:d8 18 'lhS l:l.xd4 19 i.h3 wins the queen.

3) 16 ... 'i'd7 17 'i'b3 l:he8 18 l:l.dt, intending d5, with a strong initiative.

4) 16 .. . c6 17 'i'a4 h5 18 filxa7 lhd4 19 l:l.el l:l.hd8 20 i.xd4 1-0 Kharlov-T.Christensen, N lilrreSWldby Open 1993 was an abrupt finish.

S) 16...�b8 17 'i'a4 c6 18 l:l.dl l:l.d7 19 l:l.d3 l:l.hd8 20 l:l.a3, Zarnicki -Howell, Capablanca Memorial, Matanzas 1993, with fine compensation for White.

lvanchuk's move is good and allows him to develop counterplay against d4.

17 'i'a4 �b8 White wins after 17 ... l:txe3 1 8

fxe3 'i'e4 (18 ... 'i'f3 19 l:tel) 1 9 'i'b3! (not 19 let hd4 and White's queen is threatened) 19 ... l:te8 (19 ... l:txd4 20 ig2 also wins) 20 l:tel i.xd4 21 1g2 i.xe3+ 22 �h 1 and Black's position collapses.

18 l:l.d1 (D)

White's plan is to use the bishop on fl to support .l:d3-a3, and then to swing the bishop back to g2.

18 ...... c6 After 18 ... l:txe3 19 fxe3 1l'e4 (the

line 1 9 ... 'i'f3 20 l:l.d3 is similar) 20 l:l.d3 White consolidates his extra pawn.

19 l:td3 19 i.g2 is premature since after

1 9 ... l:te7 ! 20 d5 cxd5 21 c6 d4 the position is unclear.

19 ·-· 11'e4 20 l:ta3

In view of Black's improvement at move 21 , an interesting alternative here is 20 'i'd l !?, intending either i.g2 followed by d5, or simply b4, a4 and b5. The queen on dl overprotects d4, preventing a countersacrifice by Black on that square.

20 ••• a6 21 i.d3 (D)

After 21 i.xa6 l:txd4! Black gains enough counterplay to hold the balance: 22 i.xd4 i.xd4 (22 ... 11'xd4 is

104 VI SHY ANAND: MY BEST GAMES OF CHESS

also viable) 23 i.fl i.xc5 (but not 23 ... i.xb2? 24 "ira7+ �c7 25 l:tb3 and wins) 24 "ira8+ �c7 25 "ira5+ i.b6 26 'li'xg5 l:te5 ! and the pressure on f2 gives Black enough for the pawn.

21 "irg4? This move loses because it leaves

the c6-square weak. During the game, I realized that 21 . .."ird5 ! probably saves Black:

I ) 22 i.c4 "ire4 will be a repetition.

2) 22 l:tb3?! i.xd4 and now: 2a) 23 l:txb7+? �xb7 24 1rxa6+

�b8 25 "irb6+ �c8 26 "ira? (after 26 i.f5+ l:td7 27 i.xd7+ "irxd7 28 i.xd4 "irxd4 White has �o perpetual check) 26 ... l:td7 27 "ira8+ �c7 28 1rxe8 i.xe3 and Black wins.

2b) 23 i.xa6 l:te7 ! 24 l:td3 (24 i.xd4 Wfxd4 25 1rxc6? "irdl + 26 �g2 WVxb3 wins) 24 .. .l:te4 25 l:tb3 l:td7 and White's attack runs out of steam.

2c) 23 i.xd4 l:el+ (23 ... I' d4 24 l:txb7+ �xb7 25 "irxa6+ � 26 Wfb6+ �aS 27 •xc6+ �b8 leads to perpetual check) 24 .if! "i'xd4 25 l:txb7+ (25 WVxc6? l:txfl+ 26 �fl 'li'dl + 27 �g2 loses after 27 ... 'l'xb3) 25...�xb7 26 "irxa6+ �7 (26 ... �b8 27 1Vh6+ is a draw) 27 "i'a5+ �d7 28 Wfxe 1 1rxc5 and Black is slightly better.

3) 22 i.xa6 i.xd4 23 hb7 (23 l:tb3 is line 2b above) 23 ... hcS!! (D) (an amazing defence that I spor· ted w hi 1e waiting for his reply; not 23 ... i.xe3 24 Wfa7+ �7 25 b6# nor 23 ... �xb7 24 'li'b4+ �c8 25 l:ta7 winning) and now:

3a) 24l:tb3? Wfd l+25 �g2llxe3 26 "ira8+ (or 26 fxe3 l:td2+ 27 �h3 Wfh5+ and wins) 26 ... �c7 27'ia5+ �d7 28 fxe3 "ire2+ 29 � 'if!+ 30 �g4 f5+ and Black mates.

3b) 24 .. a8+ �c7 25 'lra5+�xb7 26 i.xc5 .. dl+ 27 �g2 •d5+ and White cannot gain the advantage :

ANAND - IVANCHUK, LINARES 1993 105

3bl) 28 f3? ::le2+ 29 �h3 'i'e6+ and Black wins.

3b2) 28 ::tf3 :as 29 Wb4+ (29 9b&+ "'c8 and White is lost be-cause of the doomed rook on f3) 29··· "'c8 is unclear, but White can-not have the advantage.

3b3) 28 "'h3 'i'f5+ is a draw. 3b4) 28 "'gl 'i'd l + is also a

draw 3c) 24 .txc6 (best) 24 ... 'i'd l + 25

ri>g2 'i'xa4 26 ::lxa4 ::lxe3 27 fxe3 :d2+ 28 � (28 "'h3 ::lxb2 is just a draw)28 ... ::lxb2 with slight winning chances for White. A subsequent game Gi.Hemandez-Howel , Cap blanca Memorial, Matanzas 1993 ended in a draw. It's certainly possible to repeat this and try to play for a win in the ending, but White's advantage isn't that great

22 ::lb3! .txd4? (D) This loses immediately, but the

alternatives are not much better: 1) 22 .. .::td7 23 .ba6 (not 23 ::lb4?

::le3 24 fxe3 'ilt'f3 and Black wins) 23. .. .bd4 24 'l'xc6 'I'd 1 + 25 "'g2 •xb3 26 'i'xd7 and wins.

2) 22 ... ::le7 is relatively best, but after 23 ::lb4!, threatening both 24 .ba6 and 24 d5, White has a clear advantage in any case.

23 lhb7+!! "'xb7 24 'lfxa6+ �b8 25 'ilfb6+ "'a8 26 11Vxc6+ �b8 27 'ilfb6+ "'a8 28 .i.bS 1-0

Because 28 ... ::lc8 29 .tc6+ ::lxc6 30 'i'xc6+ "'a7 3 1 'i'xe8 'i'dl+ 32 �g2 .txe3 33 'i'xe3 leaves White three pawns ahead.

Gurevich and I had the final position on the board during our work in 1 99 1 - an unusual experience for me! I consider myself reasonably well-prepared, but to have worked out the whole game in advance is rare. It is also quite unusual to catch Ivanchuk out in opening preparation. Several players later asked me: "Gee, you out-prepared Ivanchuk?"

If it hadn't been for the following game, which I played a couple of days later,I would have considered this my best game from Linares 1993.

Game 20

B. Gelfand - V. Anand Linares 1993

Queen's Gambit Accepted

1 d4 dS 2 c4 dxc4 3 e4 cS 4 dS l0£6 5 lDc3 bS

I had prepared this line for my match against Ivanchuk the previous year, when I was analysing with Wolff, but as Ivanchuk played 1 e4 throughout the match, I had to wait a year before it came up. During the interim I had realized there were some holes in the earlier analysis, but I hadn't really looked at it thoroughly, so to some extent the specific line played in this game was improvisation at the board.

6 .i.f4 'i'aS 7 eS �4 8 li:lge2 lila6! 9 f3

After 9 a3 .i.b7 10 f3 li:lxc3 1 1 li:lxc3 li:lc7! Black is slightly better because the d5-pawn is weak.

9 li:lb4! (D) In the second round I played

9 ... li:lxc3 10 li:lxc3 .i.f5 against Beliavsky, but he responded very accurately: 1 1 g4 i.g6 12 a4 ! and White gained the advantage. In my earlier analysis I had considered 9 ... lilb4,

but only with the idea of taking the rook on h1 (see the note to Black's 1 1th move). After the Beliavsky game I looked at 9 ... li:lb4 again, and discovered the move 1 1...g6! !. I did'nt imagine that I would face this line again, but a few days later Gelfand confidently went down the same variation. I was quite happy as I felt sure that he would not have seen l l ...g6! ! .

10 fxe4 lt:ld3+ 11 �d2 g6!! (D)

I played this and got up from the board. Gelfand sank into lenghtly thought, during which time he seems to have convinced himself that he was completely lost. This is an

GElFAND - ANAND, LINARES 1993 107

exaggeration, but I think that Black is already slightly better. White's best lines lead to positions in which Black has something like two pawns and a dangerous attack for the piece.

1 1 ...1Df2?! 12 'irel ltlxh 1 is completely wrong. The knight will be tmpped after 13 g3 and Black has surrendered all his pressure.

After the text-move Black's immediate compensation lies in the paralysing effect of his knight. However, in the longer term Black may play ... i.g7 and ... ltlxe5, when the queenside pawn-mass, supported by the g7-bishop, will become extremely dangerous, especially as White's king will be floating around in the centre.

12 b3?! This is just a mistake. The alterna

tives are: I ) 12 a4 b4 1 3 lbbl (13 ltJb5 a6

and the knight is trapped)13 ... i.g7 will fantastic positional compensation for the piece.

2) 12 d6 exd6 (12 ... e6!? is interesting, taking away the d5-square; Black plans ... i.g7 and ... i.b7 with great play, but maybe the d6-pawn improves White's prospects slightly) 13 a4 and now:

2a) 13 ... 1Dxf4 (this was our postmortem analysis, but there is a hole in it!) 14 ltlxf4 i.h6 15 g3 dxe5 16 Wc2! (the line 1 6 axb5 1Wd8+ 17 Wc2 1Wxdl + 18 lhdl exf4 favours Black) 16 ... exf4 (16 ... b4 17 1Wd6! is good for White) 17 'it'd6! (the flaw; after 17 'llt'd5 0-0! 18 'llfxa8 b4 Black has more than enough compensation) with an unclear position. One possible line is 17 ... i.e6 18 'irc6+ �e7 19 1Wxc5+ Wf6 20 1Wd4+ with a draw by perpetual check.

2b) 13 ... b4 14 ltJd5 i.g7 15 1Df6+ (15 exd6 0-0 is unclear) 15 ... i.xf6 16 exf6 i.e6 intending ... 0-0-0. Admittedly this is all a bit speculative, but Black has real compensation.

Both line 2b and 12 ... e6 offer Black good chances against 1 2 d6.

3) 12 g3 i.g7 13 i.g2 ltlxe5 and Black continues with his plan.

4) 12 We3 i.g7 13 g3 ( 1 3 1Dc1 ? 1Dxf4 1 4 Wxf4 he5+ 15 Wxe5 g5 ! leads to mate in six more moves by 16 d6 f6+ 17 Wd5 e6+ 18 Wxc5 'iib6+ 19 Wb4 a5+ 20 Wa3 11Fc5+) 13 ... ltJxe5 and again Black has good play. He will continue with ... 0-0 and possibly ... f5.

5) 12 i.g3 i.h6+ 13 Wc2 b4! ( 1 3 ... 'itb4 is met by 14 'irbl ! with

1 08 VlSHY ANAND: MY BEST GAMES OF CHESS

advantage to White, and not 14 .l:b1? 'i!Vb3+! 15 axb3 lDb4#) regaining the piece.

6) 1 2 �c2 can be met by 1 2 ... b4 or 12 ... �g7 13 �b1 0-0, intending . . . l:b8 and ... b4.

12 13 bxc4 14 lDxf4?

14 cxb5 �xeS 15 'ti'b3 �xe2 16 ..liLxe2 0-0 is excellent for Black, but the text-move loses even more rapidly.

14 IS lDfe2

w

16 1Wa4+

�xeS b4 (D)

White also loses after 16 :C 1 bxc3+ 17 lDxc3 j£4+ or 16 1Wb3 bxc3+ 17 lDxc3 l:b8.

16 1W:xa4 17 lDxa4 �xal 18 lDxcS 0-0!

I thought about 18 ... f5, trying to break up his pawn-chain, but then I realized that after 19 lDf4, the arrival

of a knight on e6 would only complicate the winning process.

19 lLld3 aS Black has a decisive material ad

vantage, but White still has some potential counterplay with his central pawns, so the technical phase is still quite interesting.

20 g3 � White must try to make something

of his pawn-mass, but Black will strike with both a left hook ( ... .ia6) and a right hook ( ... �h6).

21 .i.g2 h6! 22 cS

22 l:c 1 l:ac8 23 c5 will be a transposition.

22 ... :acs 23 c6 l:!.fd8

Threatening 24 ... �xd3 25 �xd3 .l:!.xc6.

24 .l:!.cl �h6+ 2S lDef4

25 lDdf4 e5 26 .i.h3 f5! wins for Black.

2S �xd3 26 �xd3 eS 27 �c4

Again, if 27 .i.h3 then Black replies 27 ... f5.

27 ... exf4 28 l:!.el fxg3

Black still has to becaref ul ; if White could obtain theee connected passed pawns then his counterplay could prove troublesome. For this reason 28 ... f6? is inferior since after 29 .l:!.fl ! g5 (29 ... fxg3 30 :xf6) 30

GElFAND - ANAND, LINARES 1993 109

h4 Black will have to be extremely careful.

29 eS Now 29 hxg3 f6 1eads to the block

ade d the pawns. 29 ••• .U4

Here the bishop is well-placed to hold back the pawns.

30 hxg3 Or 30 lte4 g5.

30 ... 31 .:eJ

.i.xg3 .i.f4! (D)

Such little finesses make life easier. After 3l....i.h2 White could play 32 d6.

32 .l:l.e4 .i.h2 32 ... g5 was also good.

33 .i.h3 Now White cannot play 33 d6 be

cause his rook is blocking the long diagonal.

33 34 .l:l.e2 35 .l:l.e3 36 .l:l.e4 37 'ificS

.l:l.c7 ..tg3 M4 g5 .l:l.e7

This move wins tacticaUy. 38 �d4

Or 38 d6 .l:l.xe5+ 39 .l:l.xe5 .i.xe5 40 .i.d7 h5! 41 c7 .l:l.fB 42 �d5 (42 .i.e8 .i.xd6+) 42 ... .i.f4 (or 42 ... .i.xd6 43 �xd6 g4 and after 44 �e7, with the threat of 45 .ie8, Black wins by 44 ... .l:l.a8!) winning as 43 .i.e8 .l:l.xe8 44 d7 fails to 44 ... .l:l.e5+ 45 �d4 .l:l.e4+ 46 �xe4 .i.xc7.

38 39 d6 40 .l:l.xeS

0-1

fti! .i.xeS+ .l:l.xd6+

I was extremely proud of this game and Gelfand was very sporting; he said that he didn't mind losing such a game and that I would have good chances to win both the best game and best novelty prizes in lnformator (in fact I won neither!).

The fOllowing game was played in the penultimate round (round 12). I was on +3 and havinga very good tournament by any standards, but after eight rounds I had been on +4, sharing the lead with Kasparov. However, Kasparov was m even better fonn and beat both myself and Karpov to take the lead; he went on to win the tournament in very convincing style.

Game 21

V. Anand - E. Bareev Linares 1 993 French Defence

1 e4 e6 2 d4 dS 3 lllc3 li:lf6 4 eS lllfd7 5 f4 cS 6 lllr3 a6 7 ..te3 bS

I'm not sure what the point of delaying ... lllc6 is (the same idea was played in Kamsky-Ivanchuk, Tilburg 1992). I just responded with natural moves.

8 'l'd2 ..te7 The Kamsky-lvanchuk game con

tinued 8 ... ..tb7 9 ..td3 b4 10 llldl lllc6 I I 0-0 cxd4 12 lllxd4 ..te7 1 3 :n 0-0 14 :h3 g6 15 lllf2 and was also won by White.

9 ..td3 g6 This is really asking a bit too

much of Black's position. Of course the position is somewhat closed but Black cannot completely neglect his development.

10 0-0 ..tb7 11 llld1

Forcing Black to commit himself. If White is allowed to play c3 then, having reinforced his centre he will be free to play lllf2 followed by either g4 or lllg4 as appropriate.

11 12 lllxd4

13 b4!

cxd4 lllcS (D)

A similar idea to White's 18th move in Game 14. White locks the queenside pawn structure and increases his dark -square control on that side of the board, too.

13 ••• llla4 1 3 ... lllxd3 14 cxd3 liJc6 15 llbl is

clearly better for White (see Game 14 for a similar type of position). while after 1 3 ... llle4 14 'l'el followed by a4, Black's queenside is undermined.

Bareev chooses the best plan. which is to manoeuvre his knight to c4.

ANAND - BAREEV, LINARES 1993 111

14 c3 15 i.f2?!

l0b6! his development by I 8...lLxl7 1 9 �b3 �b6 20 �a5 �bc4, and

Probably not the most accurate; White's queenside play has been as we will see, it allows Black a good stymied. White's problem here is chance at move 17. It would hallie poorly placed knight on b2; in been better to play 15 �b2 �c4tl6 game White manages to swap it <tlxc4 dxc4 17 i.c2, followed by a4, off for the well-placed c4-knight effectively gaining a tempo over the The blockading move ... �a3! is note to White's 17th move. rather unusual and I had simply

15 .•• 'flc7 missed the possibility. 16 �b2 �c4 18 a4 17 _.e2?! (D) Now White has a distinct advan-

Even here 17 �xc4 would be bet- tage. ter, for example 17 ... dxc4 18 i.e2 18 llJxd4 <tlc6 19a4 lbxd4 20i.xd4, followed 19 .hd4 (D) by 'l'e3 and i.f3, with a slight edge fc:r White.

17 �c6? A grave error after which Black's

poslmdeclines dramatically. It is essential to hold up a4 and 17 ... �a3! seems to do the job for a few moves. After 18 lbc2 Black just returns to c4 (not 18 ... 'i'xc3? 19 i.d4), while after 18 :ac1 Black just continues

19 i.c6? A loss of tempo when Black can

ill afford it. 19 ... 0-0 was a better chance, although White has various lines that preserve his advantage:

l ) 20 a5 �xb2 (20 ... f6!? 21 exf6 .i.xf6 22 'i'xe6+ �g7 gives Black some counterplay for the pawn) 21 'i'xb2 f5 although here it would be hard to make progress.


2) 20 axb5 axb5 2l l0xc4 dxc4 22 .i.e4 with an edge.

3) 20 lOdl (perhaps the strongest line; White preserves his knight for the kingside attack) 20 ... bxa4 (oth· erwise White plays a5, and then feeds his knight to the kingside) 2 1 l0f2l f6 (21....i.c6 2 2 �g4 is even more dangerous) 22 exf6 .i.xf6 23 .i.xf6 l:txf6 24 lLlg4 with advantage. Note that 24 . . . l:l.xf4 loses to 25 1Wxe6+ �h8 26 l:.xf4 1ixf4 27 l:tfl 'ifd6 28 'iff7, with decisive threats.

20 axbS axbS 21 l:.xa8+

The correct move-order. After 21 lLlxc4 dxc4 22 .i.e4? Black is not forced to exchange as he can just play 22 ... 0-0.

21 .i.xa& (D)

22 l0xc4! bxc4 Forced, as 22 ... dxc4 23 .i.e4 0-0

(23 . . . i..xe4 24 'ifxe4 0-0 25 l:tal is similar) 24l:tal wins; this may seem a strong evaluation but I don't see

any hope for Black. Although in terms of structure Black has the

'good' bishops, in fact his bishop has no squares at all. White can pene trate down the a-file and Black will soon lose the pawns on b5 and c4. It's an unusual case, but here White's 'bad' bishop is far better than Black's 'good' bishop, which is se· verely restricted by White's pawn chains.

23 i..c2 .i.c6 24 'ilre3

Here White doesn't have a decisive entry on the a-file, but he has both a middlegame advantage (attack on the kingside) and an end game advantage (protected passed b-pawn).

The move played prepares a strong reply in case Black castles.

24 .•. O.O (D) Mter 24 ... �d7 White just contin

ues 25 g4, followed by f5.

25 fS! exfS

ANAND - BAREEV, LINARES 1993 113

After 25 . .lla8 26 f6 �f8 White can either start an attack with 27 h4 or simply seal Black's bishop in by g4-g5 , continuing with both an extra pieces and extra kings.

26.tf5 (D)

Now White has the possibility of e6, opening up the kingside. Once again we can see that White's bishops are much more effective Black's.

%6 ••• 'li'd8 Or 26 .. Jia8 (26 ... gxf5?? loses the

queen after 2-rfk g3+ �h8 28 e6+) and now 27 e6 gxf5 28 exn+ �f7 29 'l'h6 :tl!! is not conclusive, so White should play 27ig4, followed by e6, much as in the game.

27 .ig4! To help with e6. The bishop has

no more work to do on the bl-h7 diagonal.

27 ... �gS 28 1i'e2 hS (D)

Black was almost in zugzwang: l) 28 . . . .i,d7 29 e6 wins. 2) 28 .. :ife7 29 ,tc5 picks up the

exchange. 3) 28 ... l:l.e8 29 e6 (the simplest)

29 .. .f6 (now that the rook has left f8 White can meet 29 .. .f5 by 30 .ixf5 gxf5 31 l:l.xf5 winning) 30 g3, followed by h4, and wins.

29 .ixbS gxhS 30 'i!lxbS .ie8

Or 30 . .. �7 3 1 h4! winning after 31 ... .ixh4 32 l:l.f4 or 3 1 ... id2 32 l:l.f3.

31 l:l.f6! 1-0 The end might be 3 1 ....id2 32 e6

.ie3+ 33 �hl fxe6 34 J:g6+. A finish reminiscent of the fa

mous encounter Fischer-Benko, US Championship 1963/4.

This win put me in joint second place with Karpov on +4 - it was an exceptionally good result.


At the Madrid tournament in June I started with five draw. No matter what I did I would either have to defend a worse position and hang on for a draw, or not be able to break through. At this point there was an article in the tournament bulletin, which said that I was Kon vacation". It infuriated me. I don't know if there was any connection, but I won my last four games and finished in joint first place. The following game was the second of the series.

Game 22

V. Anand - F. lzeta Madrid 1993

Pirc Defence

1 e4 d6 2 d4 ltlf6 3 lt:lc3 c6

I was happy with this line as it game me the chance to play something reaDy sharp.

4 f3!? eS S .ie3 i.e7 6 ... d2 0-0 7 0-0-0 bS 8 g4 exd4

8 ... b4 9 lt:lce2 exd4 10 lt:lxd4 c5 I I �5 ! gives White a good attack, while after 8 ...... a5 !? 9 �b1 b4 10 lbc:e2 i.e6 1 1 lLlcl c5 12 d5 i.d7 1 3 �ge2, foUowed by lt:lg3, White is also slightly better.

9 ... xd4!? The right recapture. You shouldn't

take with the queen if Black still has the option of ... lt:lc6, but in all the lines with .. .c6, taking with the queen is the normal capture. The logic is the same as in the Sicilian after 1 e4 c5 2 li.:lf3 lt:lc6 3 d4 cxd4 4 lDxd4; Black mustn't play ... lDxd4 too early,as the queen occupies a dominating post in the centre.

After 9 il.xd4 Black could continue 9 ... b4 10 lDce2 c5, and grabbing a pawn by 1 1 i.xf6? i.xf6 12

11'xd6 1l'a5 would give Black excellent compensation.

9 i.e6 10 gS (D)

10 For a time I was worried about

1 O ... c5, but it turns out that White can gain the advantage by an accurate sequence of moves: 1 1 1l'd3 ! (not 1 1 11'd2 b4 1 2 gxf6 bxc3 1 3 11'xc3 i.xf6 and Black i s a little better) 1 1 ... c4 ( I I .. . i.c4 12 1l'd2 favours White) 12 1l'e2! (12 1l'd4 lDc6 1 3 gxf6 lDxd4 1 4 fxe7 11Vxe7 1 5 i.xd4 is risky as ... b4 and ... c3 might rip open the white king position before his minor pieces can come into play) 12 ... lDe8 1 3 lDxb5 and White stands


very well, since 13 ... .txg5 may be met by 14 f4 and 15 e5.

11 f4 WFas 12 a3 fS?

Black had two better lines: 1) 12 ... c5!? 13 1ird2 b4 and now: la) 14 �d5 .i.xd5 15 exd5 ( 15

11rxd5 bxa3 ! i s good for Black) and after 15 ... �d7 or 15 ... .i.d8 Black has good counterplay.

lb) 14 �b1 ! �c6 15 li:lf3 l:tb8 16 f5 ! .i.a2! (l6 ... .i.d7 17 .i.c4 bxa3 18 �xa3 with advantage to White) 17 b3! .i.xb1 18 �xb1 bxa3 (18 . . . 11rxa3 19 .i.c4 �aS 20 1ird3 is also good for White) 19 11rxa5 �xa5 20 �a2 and White has the two bishops and a structural advantage.

2) 12 ... f6 13 h4 �c7 is rather unclear.

After the game !zeta explained that he did not like 1 2 .. f6 because White is not obliged to take on f6, but can play 1 3 h4 (as in line 2 above). He therefore preferred 12 ... f5 , because he thought that it would force 13 gxf6; otherwise White would be left with a backward f-pawn and not much chance of an attack. However, this argument is flawed because after the exchange on f5 Black cannot maintain his bishop on that square, and once it has gone, the white queen can occupy the powerful central square e4, ready for moves such as .i.d3 and f5.

13 exfS! .txrs 14 �ge2

Now White has a definite advantage.

14 ••• fi:A:7 15 li:lg3! .i.g4 (D)

If 15 ... lt:le6, then 16 'ilfd2 i.g4 17 J:te1 followed by .i.d3 with a promising attacking position.

16 .i.d3! Although this is not bad, White

misses the most efficient continua· tion, 16 Wfe4!:

I) 1 6 . . . .i.xd1 17 .td3 g6 18 f5! (18 Wxe7 is also good) 18 ... .if3 19 Wfxe7! J:te8 (19 ... D.f7 20 "i'd8+ llf8 21 Wfxd6 .i.xh1 22 fxg6 wins) 20 Wfxd6 J:txe3 2 1 fxg6 with a decisive attack.

2) 16 .. .D.e8 17 .i.d3 g6 18 f5 .i.xd1 (18 ... .i.xg5 19 Wfxg4 .i.xe3+ 20 �b1 is hopeless) 19 fxg6! (19 �xdl .i.xg5 is unclear; Black will win a piece, but meanwhile White will get in fxg6) 19 ... .i.f8 (19 ... .ixg5 20 gxh7+ �h8 21 ._d4+ J:te5 22 .ixg5 wins) 20 gxh7+ �h8 (20 . .. -i>g? 21

ANAND - /ZETA, MADRID 1993 117

hB'if+ �xhB 22 'Wb7#) 21 �d4+ iLg7 22 iLxg7+ 'il;>xg7 23 'ifg6+ �f8 24 'iff6+

16 .•. d5 (D) 16 ... iLxd1 17 'it'e4 transposes into

the preceding note.

17 fS! Of course White cannot play 17

:del c5 18 'l'e5 �6 and his queen is trapped ,but I always intended to sacrifice the exchange here.

17 ... �xd1 18 lt:lxd1 (D)

This is again a slight inaccuracy. The alternatives are:

I) IB:Xd1?'i'b6 19 'i'e5 (19 'i'f4 ot\e6! 20 'l'e5 'l'xe3+! and Black wins)l9 ... iJ..d6 20 'it'xd6 (20 �xb6 1xeS is lost for White) 20 ... 'ii'xe3+ 21 �bl lCJeB 22 'i'dB (22 'iVb4 'l'xg5 iswinning)22 ... lt:lf6 ! 23 iic7 'it'xg5 24 l'b7 lCJbd7 25 iixc6 and Black has a clear advantage.

2) 18 f6! :txf6! (18 ... gxf6 19 lt:lf5 and l8 ... iLxf6 19 gxf6 :Xf6 20 :Xd1

are hopeless fer Black) 19 gxf6 hf6 20 'ilff4 �xc3 21 l:lxdl and White has an extremely strong attack.

18 �d6? Black should have hied 18 ... c5!

19 'l1Vg4 ( 1 9 'l'h4 c4 20 f6 cxd3 21 fxe7 l:leB 22 lLlf5 'it'a4! is unclear) 1 9 . . . c4 and now:

l ) 20 1Vh5? cxd3 21 g6 h6 22 �xh6 1Va4 ! 23 lDe3 1Vh4 24 �xg7 1Vxh5 (24 ... �xg7?? 25 f6+! wins for White) 25 lDxh5 �g5! and Black wins.

2) 20 �2 (it is not very attractive to have to retreat the bishop, but White retains a very dangerous attack) 20 ... �d6 (20 ... �h8 2 1 f6 gxf6 22 lDf5 lDc6 23 g6 wins) 21 f6 g6 22 h4 and the position is still very difficult fer Black.

19 iih4 Now White's attack is decisive.

19 ... iDeS Alternatively, 19 ... lDd7 20 f6 g6

21 1Vh6 l:lf7 22 �xg6 hxg6 23


'fkxg6+ �fB 24 lilf5 and White's attack breaks through.

20 .i.d4 'fkc7 20 ... c5 loses immediately after 21

.i.xg7 ! lilxg7 22 f6. 21 f6 (D)

21 ••• gxf6 After 2l ...g6 the most convincing

line is 22 l:l.el ! (22 hg6 hxg6 23 f7+ �xf7 24 'l1Fh7+ lilg7 is less convincing) and now:

I) 22 ... 'fkf7 23 lile3 1ie6 24 lilgf5! wins.

2) 22 ... c5 23 .i.xg6 hxg6 (White also wins after 23 ... cxd4 1A De7 'fixe? 25 fxe7 hxg6 26 exf81i+) 24 f7+�xf7 25 'fkh7+ lilg7 26 1ixg7t.

22 gxlti l:l.f7 23 l:l.g1 �h8 24 .i.xh7! lhh7 25 f7+ li:Jg7 26 .hg7+ 1..0

As 26 ... �xg7 27 li:Jf5+ �f8 28 'fk xh 7 leads to mate.

The summer of 1993 was Interzonal time again. I started with two draws, and the following game was played in the third round.

Game 23

V. Anand - L. Ftacnik Biel Interzonal 1993

Sicil ian, Najdorf

1 e4 cS 2 lbt'3 d6 3 d4 cxd4 4 li}xd4 lbt'6 S M a6 6 ..te3

I had quite a few interesting ideas in the .te3 line against the Najdorf, but this is a very difficult and complex variation to analyse. It has taken many years for theory to converge oo what are now considered 'main lines' . The positions are so tricky that you can never be sure that your ideas are correct; the advantage is that your opponent has the same problem!

6 e6 7 f3 bS 8 g4 h6 9 'i'd2 i.b7

This game showed for the first time that 9 ... .tb7 is just a mistake with this move-order. Black has to play 9 . ..ll:lbd7, when we reach the main line (after lO 0-0-0 .tb7). In this case White would have no time for 10 h4 because of lO ... b4 1 1 li.:lce2 d5, and White cannot push his e·pawn since the e5-square is covered

10 h4 Here Ftal!nik thought for some

time and realized that the attempt to transpose to the main line with 10 ... lilbd7 is bad after 1 1 l:lg1, with the immediate threat of g5-g6.

10 •.• b4 (D) After 10 ... i.e7 1 1 0-0-0 White is

slightly better because Black has been forced to spend a tempo on ... 1J..e7, which he can normally avoid in this system.

11 �el dS 1 l .. .e5 1 2 'i'xb4! gives White a

clear advantage after 12 ... 'i'd7 13 lDb3 d5 14 lDc5 or 12 . . . 'i'c7 l 3 'i'a4+ lDbd7 14 lDf5.

12 eS lDfd7


13 f4! We have reached a kind of super

French position, but in the French Black would not weaken his pawn structure by playing .... h6.

13 .•• ll:lcS

14 .i.g2! (D) Better than 14 ll:lg3 .i.e 7 - the

knight might go to f4 later.

14 ll:lbd7 After 14 ... ll:le4 White can start the

attack against e6 with 15 ixe4 dxe4 16 f5!, followed by ll:lf4.

15 0·0·0 .i.e7 The alternatives are also favour

able for White: 1) 15 ... ll:lb6 16 b3 a5 (16 ... ll:le4

17 .i.xe4 dxe4 18 f5 .i.d5 19 ll:lf4 and again e6 is the weak point) 1 7 f5 a4 1 8 fxe6 axb3 1 9 cxb3 fxe6 20 ll:lf4 'iWc7 21 �b1 is good for White.

2) 15...a5 16 g5 (16 �bl is also possible, with the idea of ll:lb5) 16 ... h5 17 f5 ll:lxe5 1 8 ll:lf4 ll:lc4 19 'l'e2 (a typica1 1ine; the weakness of

e6 proves fatal) 19 ... e5 20 �xdS .i.xd5 2 1 ll:lc6 1ic7 22 llxd5 winning.

3) 1 5 ... ll:le4 16 ixe4 dxe4 17 f5 ll:lxe5 18 fxe6 ll:lc4 (18...�0 19 exf7+ �xn 20 ll:lxf3 exf3 21 'l'xd8 .:.Xd8 22 .:.xd8 fxe2 23 %ld7+ wins) and now both 19 'I' e1 ll:lxe3 20 'lf2 fxe6 21 '1'xe3 and 19 exf7+ �xf720 'l'el ll:lxe3 21 'l'f2+ �g8 22 'lxe3 are promising for White.

4) 15 ... '1'b6 1 6 g5 h5 17 g6! fxg6 18 .i.h3 with a strong attack.

16 gS! White has the advantage, but he

mustn't waste too much time! 16 ... bS (D)

Black is distinctly worse after 16 ... hxg5 17 hxg5 l:lg8 18 g6! or 16 ... g6 1 7 h5! gxh5 18 l:lxh5.

17 fS! The best way 1o soften up the e6-

square. After 1 7 g6 fxg6 18 .i.h3 ll:lf8 or 1 7 �bl 'l'b6 18 g6 fxg6 19 .i.h3 ll:lf8 the situation is less clear.

ANAND - FrAtNJK, BIEL INTERZONAL 1993 121

None of this was home preparation. I had been working on the English Attack with Patrick Wolff, and he mentioned that 9 ... J.b7 is bad because of 10 h4, and we left it at that. However, all White's moves appear perfectly natural, so it was not difficult to play. In such messy and complex position, think it is better not to calculate too much - the tree of variations can get enormously dense. I prefer to wait to see what my opponent plays, and that immediately removes a large percentage of possible branches.

17 •.. lln:eS (D) Forced, as 17 ... exf5 1oses to 18

/i)xfS �xeS 19 lbxg7+.

w

18 lbf4! 18 fxe6! lbc4 (18 .. .fxe6? 19 lbf4

wins) 19 exf7+ is also strong, e.g.: I) 19 ... Wd7 (Fta�nik) 20 J.h3+

Wc7 21 .tf4+ wins. 2) 19 ... Wxf7 20 1Wel lieS (the

line 20 . . . �xe3 21 1Wf2+ Wg8 22

1rxe3 is also good for White) 2I lbf5 (21 J.f4 Wg8 is Jess clear) 21...llc8 (2l. ..lbxe3 22 lbxe3 and d5 hangs) 22 J.d4 J.xg5+ (or else White has a very strong attack in any case) 23 hxg5 1rxg5+ 24 �bl 1rxg2 25 llgl with a decisive attack.

3) 1 9 ... Wf8 20 1i'el 1i'a5 (White wins the queen after 20 ... lbxe3 21 lbf4 lbxg2 22lbfe6+) 21 Wbl lbxe3 22 lbf4 (D) with the two possibilities:

3a) 22...�xg2 23 lbg6+ Wxf7 24 'ifxe7+ Wxg6 25 llhfl ! �4 (the lines 25 . . . llhe8 26 'l'f7+ Wh7 27 'l'xh5+ Wg8 28 g6 and 25 ... llhf8 26 llxf8 llxf8 27 'ifxf8 are no better) 26 'l'f7+ Wh7 (26 ... wxg5 27 llgl+) 27 'l'xh5+ Wg8 28 'l'f7+ Wh7 29 llhl mating.

3b) 22...�xdl 23 lbg6+ Wxf7 24 'l'xe7+ Wxg6 (24 ... Wg8 25 �f5 llh7 26 llfl �c3+ 27 bxc3 bxc3 28 'l'f8+ llxf8 29 lbfe7#) 25 'ii'd6+ with a final branch:


3bl ) 25...�h7 26 g6+ �g8 (or 26 .. . �h6 27 lllf5#) 27 .bd5+ .ixd5 28 'i'xd5+ �f8 29 'ikf7#.

3b2) 25...�f7 26 g6+ �g8 (mate is also inevitable after 26 ... �e8 27 lllf5) 27 .ixd5+ .ixd5 28 'i'xd5+ �f8 29 'i'f7#.

18 .•• lllc4 18 ... exf5 19 lllxf5 is clearly very

good for White. 19 'i'el (D)

19 .•. 'il'aS Going for queenside counterplay.

The alternative is 1 9 ... e5: I) 20 lllxd5 and now: I a) 20 ... lllxe3 2 1 lllc6! (21 1Wxe3

.ixd5 22 lllc6 .ixg2 23 lllxd8 .ixh1 24 'il'xe5 transposes to the unclear line I b below) 21.. .lllxg2 (2l....ixc6 22 lllf6+ gxf6 23 .ixc6+ �f8 24 .l:l.xd8+ .l:l.xd8 25 'ifxe3 also wins) 22 lllxd8 .l:l.xd8 23 lllxe7 winning.

!b) 20 ... .ixd5 and now: lbl) 21 lllb3 lllxe3 ! (2I ...lllxb3+

22 axb3 "illa5 23 bxc4 .ixg2 24

1Wxg2 1t"a1+ 25 'it>d2 l%d8+ 26 �e2 'i'xb2 27 'i'c6+ oMs 28 :xd8+ .ixd8 29 .ic5+ �g8 30 "ille8+wins for White) 22 "illxe3 lllxb3+23 axb3 .ixg2 24 .l:l.xd8+ .!hd8 favours Black

I b2) 2 1 lllc6 .ixg2 (2l...fue3 22 .ixd5 1Wc7 23 'ilfxe3 and White wins) 22 lllxd8 (22 .l:l.xd8+ :xd8 23 lllxd8 lllxe3 24 .l:l.el �xd8 25 'l'xe3 .ie4 is also unclear) 22 ... lllxe3! (not

22 ... .ixhl? 23 .ixc5) 23 'iWxe3 ixhi

24 'i'xe5 (24 lllxf7 �xf7 25 :xhl is again unclear) 24 ... f6! (24 ... llxd8 25 .l:l.xd8+ 'it>xd8 26 'i'b8+ �7 27 "illxh8 g6 28 'ilfh7! should win for White) and the position remains unclear.

2) 20 f6!? gxf6 21 lllxdS .hdS (21 ...lllxe3 22 lllc6 lllxd5 23 I!Jxd8 .l:l.xd8 24 gxf6 .ixf6 25 'i'c4 and White wins) 22 lllf5 lllxe3 23 1!Jxe3 with advantage to White.

3) 20 lllc6! .ixc6 21 lllxdS (21 'llhc4 dxc4 22 .ixc6+ �8 23 .l:l.xd8+ .l:l.xd8 24 llld5 llld7 is good for Black) 2 1 ...'llra5 22 lllxe7! (after 22 .ixc5 'i'xa2 23 lllc7+ ¢>f8 24 .ixe7+ �xe7 or 22 'ifxc4 .ib5 23 'ikxb5+ axb5 24 lllxe7 llld7 25 ixa8 'llfxa8 26 llld5 11Vxa2 Black has the advantage) and now:

3a) 22 . . . .ib5 23 .ic6+ .ixc6 (or

23 ... � 24 .ixc5 'i'xa2 25 lllc8+1 and mates) 24 'llfxc4 .ixhl 25 .ixc5 .if3 26 g6! .l:l.f8 27 gxf7+ libn 28 lllc6 and wins.

3b) 22 ... .ixg2 23 'i'xc4! lllb3+ (23 ... .ixh I 24 .ixc5 n-ansposes to

ANAND - FrACNIK, BIEL INTERZONAL 1993 123

line 3 a) 24 cxb3 and White is winning.

3c) 22 ... '1'xa2 23 �xc6+ <M8 24 �g6+ �g8 25 'i'xc4 'i'xc4 26 �xa8!

wins for White.

2 J ...lilxb3 (2I ...ea3 22 �xc5 li:la4 23 �d4 li:lc3+ 24 �xc3 bxc3 25 li:ld3 wins) 22 axb3 li:lxdl 23 l:txdl 0-0-0 24 fxe6 I would prefer to be White, but in the end 21 fxe6

20 �bl tt:lxb2 (D) seemed even stronger! The alternatives are winning for Note that 21 �xb2? is bad in view

White: of 2 l...li:la4+ 22 �c l tt:lc3 23 16'd3 l) 20 ... tt:la4 21 fxe6 tt:lcxb2 22 e5.

exf7+ �f7 23 �d2 with a decisive 21 attack ,foc example 23 ... li:lxdl 24 :XdJ IDc3+ 25 �xc3 bxc3 26 'i'e6+ �27 hd5 !.

2)20 ... e5 21 tt:lxd5 tt:lxe3 (White wins after21...exd4 22 �xd4 �xd5 23 .ixd5) 22 'i'xe3 �xd5 23 �xd5 exd4 24 lbd4! l:tc8 25 f6 finishes Black.

w

21 fxe6!! White should just ignore every

thing, except mate itself, and just hack away, I was on the verge of playing the reallyamazing move 21

�b3!?, if only because it eliminates all threats to White's king. After

Or: 1 ) 2 1 ...0-0 (Ftaenik) 22 11Vxb5

li:lxdl 23 l:t:xdl tt:la4 24 �d2 li:lc3+ 25 �xc3 bxc3 26 exf7+ l:txf7 27 g6 irb4+ 28 li:lb3 and Black's position collapses.

2) 2 l . ..tt:lxdl 22 exf7+ �xf7 (or 22...�d7 23 l:txd l ) 23 l:txdl (D) and now:

2a) 23 ... tt:le4 24 �xe4 dxe4 25 'llfc4+ �e8 26 �de6 l:tc8 27 tt:lxg7+ �8 28 'Llg6+ �xg7 29 �d4+ leads to mate.

2b) 23...lila4 24 �d2 li:lc3+ 25 �xc3 bxc3 26 'llfe6+ �e8 27 lt:lxd5


.i.xdS 28 .txdS 'i'b4+ 29 .i.b3 winning.

2c) 23 ... l:af8 24 �xdS winning. 3) 2 l ...fxe6 22 �xe6! (22 �xb2

lba4+ 23 �cl �c3 is less clear-cut) 22...�xdl 23 �xg7+ �f7 24 .lhdl with a decisive attack.

22 �xb2! �4+ 23 �1 b3 (D)

23...�c3 24 1Wd3 1Wxa2 25 exf7 wins as Black has no real threats.

24 �xb3! Not 24 cxb3? .ta3+ 25 �c2 1fc7+

26 �bl �c3+ and Black wins. 24 .•. .ta3+ 2S �bl �3+

Or 25 ... 1Wc3 26 .i.d4. 26 �al Wa4

Black's compensation is inadequate.

21 Wd3 .tb4 28 �cl �b8 (D)

Other lines are also hopeless: I) 28...�xdl 29 llxdl fxe6 30

�xe6 lid? 3 1 .i.d4 winning.

2) 28 •.• .tc6 29 exf7 .tbs JO .th3+ �b8 3 1 trd4 also wins.

3) 28 ... .tas 29 .id4 (29 exf7 'iWb4 30 �b3 Wa3 31 � I 'ib4 repeats moves) 29 ... 'iWb4 30 hc3 consolidates the extra piece.

29 .td4 .:.cs 30 .teS+

Even the calm 30 exf7 would have been sufficient.

30 ••• �1 (D) 30 ... �a8 loses to 31 fudS.

ANAND - FTAtNIK, BIEL iNTERZONAL 1 993 125

31 'i'e3+ J:l.cS 32 J:l.d3 'i'xc2

32 ... d4 33 .bd4 'il'xc2 34 .i.xc5+ �aS 35 .i.xb7 'Otxb7 36 J:l.d7+ Wc8 37 'iW d3 wins.

33 . .bc3 34 J:l.xc3 35 'l'xc3 36 exf7

..txc3+ 'il'xc3+ J:l.xc3 J:l.f8

37 g6 1-0 I spent only about 30-35 minutes

on this game. Afterwards, I was very proud of my cold-bloodedness during the game; I hadn't prepared it at home, but even so I played the critical and best moves with very little effort. This game won me the prestigious Best Game Prize in lnformator.

The following game was played in the very next round, and was part of my best streak of the event.

Game 24

L. 011 - V. Anand Biel lnterzonal 1993

Slav Defence

1 d4 dS 2 c4 c6 3 lD£3 llJr6 4 �3 a6

I played this simply because it was fashionable.

S e3 bS 6 b3

I don't think this is the most dangerous system for Black as his reply is quite straightforward: he just takes his bishop outside the pawn chain to g4, swaps it off and then plays ... e6. It is difficult for White to make anything of his two bishops.

6 ... .i.g4 7 h3

More recently, the refinement 7 'ifc2 has been introduced. White doesn't mind the exchange on f3, and by playing very precisely he may gain a slight advantage.

7 ... .bf3 8 'ifxf3 e6

The alternative is the sharp line 8 ... bxc4 9 bxc4 e5, but I didn't see the need for such extreme measures. In particular, I didn't like the idea of opening the position by exchanging on c4. White has no particular threats on the queenside, so there seems no

reason for Black to clarify the pawn structure.

9 .id2 9 cxd5 cxd5 10 i.d3 llX:6is equal.

9 .ib4! (D)

10 'ifdl 10 a3 i.xc3 1 1 i.xc3 bxc4 12

bxc4 �4 13 Acl ll:lxc3 14 llxc3 'if aS 15 �d2 is certainly not better for White.

10 ... 0.0 11 i.e2

After 1 1 cxd5 ( 1 1 i.d3 dxc4 12 bxc4 c5 ! is also fine for Black l l ...cxd5 12 a4 bxa4 13 1Llxa4 (13 Axa4 a5 is also comfortable for Black) 1 3 ... ll:lc6 14 i.xb4 1Llxb4 15 i.e2 'ifd6 the position is level.

OU - ANAND, BIEL INTERZONAL 1993 127

11 ... bxc4 1l bxc4 cS!

At this point I felt that I had equalized completely.

13 dxcS? After this error White is strug

gling. The alternatives are: I) 13 0-0 cxd4 14 lllxd5 lllxd5 15

cxd5 .ixd2 16 'ifxd2 dxe3 17 'ifxe3 exd5 18 .if.3 and now 18 ... �c6 19 :ad! fiJe7 20 :Ce1 l:te8 2 1 i.xd5 lZlxd5 22 'l'xe8+ 'ifxe8 23 l:txe8+ l:txe8 24 :xd5 is a dead draw,while !8 ... d4!, followed by ... l:ta7-d7, is a possible way foc Black to play for the advantage.

2) 13 cxd5 (the soundest line) 13...cxd4 14 exd4 i.xc3 (14 ... lllxd5 15 fiJxd5 i.xd2+ 16 'Wxd2 exd5 17 0..0 is a little better for White) 1 5 ixc3 fiJxd5 with equality. Even though White has an isolated pawn, the two bishops and the potentially weak pawn on a6 will ensure that he is not worse.

13 d4! (D)

14 exd4 Forced, as 14 llla4 'ifa5 favours

Black, while 14 llle4? lllxe4 15 i.xb4 dxe3 is just lost for White.

14 ... 'ifxd4 15 ii'c2

15 J:c 1 lld8 prevents castling and is very awkward for White.

15 lllc6 16 0.0 (D)

16 'ifeS 16 ... l:tad8 was also good, for ex

ample 17 l:tadl 'ife5 1 8 i.d3 i.xc5 or 17 :Cd 1 '1We5 1 8 i.d3 ( 1 8 i.f3 l:txd2!) 18 ... lt:ld4 1 9 ii'b2 i.xc5 with a clear advantage to Black in either case.

17 1Wa4 Alternatively: I) 17 :Ce1 l:tad8 18 i.f3 ( 18

l:tad I llld4 is also good for Black) 18 ... llld4! (not 18 ... l:txd2 19 ii'xd2 .txc3 20 l:txe5) 19 .lhe5 lfu.c2 20 l:td 1 l:txd2 2 1 l:txd2 i.xc3 and Black wins.


2) 17 l:tael l:tad8 (the simple 17 .. .'ihc5 is also good) 18 i.f3 lbd4 1 9 l:.xe5 lbxc2 20 lbb 1 i.xd2 2 1 l:te2 lbd4 22 l:txd2 lbxf3+ 23 gxf3 l:tc8 with a clear endgame advantage.

3) 17 i.dl ! is relatively best, but Black retains the advantage after 1 7 ... l:tad8 or 17 ... lbd4.

17 l:tad8 18 i.e1 lbd4 (D)

19 'l1Vxb4 19 i.d3 i.xc3 20 i.xc3 00+ 21

gxf3 'ibc3 is winning fa Blade, while after 19 i.dl i.xc5, intending ... l£Jf5-g3, White has a miserable possition.

19 ••• �el+ 20 lbxe2?!

Losing quickly. Tite last chance was 20�hl l:td3 (20 .•• lbe4 21 �xe4 11Fxa1 22 lbd6) 21 l:tdt .lhc322.hc3 lbxc3, when Black is much better although the passed c-pawn offer some hope of counterplay.

20 'i"xal 21 lbc3 'i"cl! 22 1WaS .4 23 11'xa6 :as 24 1i'd6 'i"xc4

0·1 As White will soon lose the c5-

pawn.

In the fifth round I won against Khalifman, and was in joint first place Then I lost recklessly against Gelfand; had I defended well I could have held the position, but I was careless and suddenly it was all over. This put me on +2, below the level necessary for qualifying. I tried very hard to improve me score but despite reaching a number of winning positions I failed to score the full point in any of them. I had given up hope when I met Korchnoi in round 12. In this game Korchnoi completely outplayed me and I felt a huge weight off my shoulders. It was suddenly clear that I wasn't going to qualify and I felt enormously relieved at the lifting of the burden. Then Korchnoi started to have difficulty winning and I began to regain my interest in drawing the game. Korchnoi gradually went completely awry and even lost the game. Suddenly I had moved from absolutely nowhere to having a theoretical chance of qualifying.

A win in the last round would have guaranteed qualification, but I could only draw with Epishin. Then it required five different games to have the

BIEL INTERZIJNAL 1993 129

right result for me to qualify.ln the end everything work out perfectly; all the people with worse tie-break than me fi.nished level with me on+3 and all the people withbetter tie-breaks ended up on +2. It wasn't very convincing, but at least I had reached the Candidates again.

The following game is from the final European Clubs Cup,in which I was playing for Lyons.Curiously,the only French player in Lyons team was Lautier. Micheal had horrible event; he lostall three game (ofwhich this was the first).He then went to Gronigen for the PCA Qualifier and almost lost the first game there. However, he salvage a draw and, typically for Micheal,then went on to win the tournament!

Game 25

M. Adams - V. Anand European Clubs Cup Final, Hilversum 1993

Sici l ian, Kan

1 e4 cS 2 lDc3 e6 3 m

While White gains some advantages with this move-order, it allows Black to adopt a system in which ... ll:\f6 is delayed.

3 4 d4 5 ll:\xd4 6 £4 7 ..td3

a6 cxd4 d6 bS .b7

Thanks to the omission of ... ll:\f6, Black is able to accelerate his queenside development.

8 0-0 One advantage of Black's system

is that 8 1i'f3?! doesn't really work when Black hasn't played ... ll:\f6 since g4-g5 threatens nothing! The game Anand-Wojtkiewicz, Manila Olympiad 1992 continued 8 ... ll:\d7 9 J.e3l:[ c8 10 g4 ll:\c5 1 1 g5 lt\e7 1 2 ll:\b3 ll:\a4 13 ll:\xa4 bxa4 1 4 lLld2 d5 and Black was slightly better. If White wants to adopt the 1i'f3 and g4-g5 system then he shouldn't play 2 ll:\c3.

8 ... l0t'6 Now that White has effectively

abandoned the 1i'f3 system (because

the advance of the g-pawn doesn't fit in with kingside castling) I decided to play ... ll:\f6. Black canstill tinker with his move-older by 8 ... ll:\d7!?, but after 9 �1 (1101 9f5? 1lrb6!) he probably bas nothing better than 9 ... ll:\gf6 .

9 a3 Meeting Black's threat to the e

pawn, but after this loss r:i tempo Black has no problems.

9 ••• li:Jbd7 10 'i'b1 i.e7 (D)

IO ... .Zlc8 and 10 ... g6 are playable alternatives.

11 b4?! 1bis plan can be effective, but

only when Black's knight has been

ADAMS - ANAND, HILVERSI.IM 1993 131

developed to c6. Here Black has a ready counter in the fonn of ... .l:l.c8 and . ..liJb6-c4.

11 •.• 0-0 12 i.b2 l:1c8! 13 'i'e2

Preparing e5, which Black meets by providing the f6-knight with a well supported square on d5.

w

13 lDb6! (D)

14 eS? If White tries gueenside pla)' by

14 a4 lt:lc4 15 i:c 1 'i'b6 16 lDf3 :fd8, then Black has a promising position; he has played normal moves but White has done disgusting things on the queenside. The continuation might be 17 axb5 axb5 18 e5 and now:

I) 18 ... dxe5?! 19 fxe5: Ia) 19 ... i.xb4 20 lDa2 i.xf3 2 1

.l:l.xf3 00 22 lllxb4 lllxb4 23 bh7+ �xh7 24 .l:l.h3+ �g8 25 'llfh5 wins for ,White

1b) \9 . ..liJd5 20 li:lg5 i.xg5 21 i.xg5 1tlxc3 22 i.xh7+ �xh7 23

ifh5+ 'itrg8 24 11'xn+ �h8 25 .l:l.f6 lDe4 (25 ... hg2+ 26 'ifrxg2 irc6+ 27 'ifrh3 lDe4 28 i.h6! and White wins) 26 i.h6 lDf2+ 27 ..l:l.xf2 gxh6 (27 ... l:.g8 28 lfh5! g6 29 ifh4 wins) with a likely draw.

lc) 19...�d7 20 lDg5 i.xg5 2 1 i.xg5 l:.f8 22 i.f4 i s unclear.

2) 18...�d5! 19 lDxd5 bd5 is fine for Black.

14 ... dxeS 15 fxeS lDrdS! (D)

Not i5 ... Wxd4?? 16 exf6 i.xf6 17 �4! (17 � I ? ifh4! 1 8 i.xf6 gxf6 favours Black) 17 ... 11'd8 1 8 hf6 gxf6 19 �xb6 11'xb6 20 •h5 f5 2 1 1l'g5+ �h8 22 11'f6+ �g8 2 3 ..l:l.xf5 and White wins.

16 �xdS I spent a long time on the alterna

tives, but I eventually decided that they were not worrying for Black:

I ) 16 lDxe6 fxe6 17 1Wh5 .l:l.f5! (17 ... g6 18 i.xg6 hxg6 19 11'xg6+ �h8 20 11'h6+ �g8 21 11'xe6+ is


only dangerous for Black) 18 l:!.xf5 g6! 19 l:!.g5 .bg5 20 .bg6 l:!.c7 ! would be winning for Black.

2) 16 �e4 and now: 2a) 16 ... �a4?! 17 "ifh5 with a

further branch: 2al) 17 ... h6?! 18 i.c l with very

dangerous threats. 2a2) 17...�xb2?! 18 �g5 ! (18

�f6+ �xf6 19 exf6 �xd3 wins for Black) 18 ... h6 (D) and now:

2a21) 19 l:!.xf7 �4 20 l:txg7+ �xg7 21 �gxe6+ �g8 22 "ifxh6 i.xg2+ 23 �gl "ifxd4+ 24 �xd4 �bxd3 25 cxd3 i.d5 is unclear.

2a22) 19 i.h7+ �h8 20 �xf7+ :Z.xf7 21 :Z.xf7 �4 22 :1xf4 �xh7 and Black defends.

2a23) 19 �xf7 l:!.xf7 (but not 19 ... "ife8 20 i.g6 and now 20 ... "ifd7 2 1 �xh6+ gxh6 22 i.f7+ :Z.xf7 23 :Z.xf7 �4 24 l:!.xf4 i.g5 25 "irg6+ "ifg7 26 "ifxe6+ wins for White) 20 "ifxf7+ �h8 2 1 �xe6 "ifg8 22 "ifg6 �xd3 with a murky position.

2a3) 17 ... g6! 18 i'h6 li:lxb2 19 :Z.f3 �f4! 20 irxf4 (20 lhf4 i.xe4 21 :Xe4 �xd3 22 cxd3 i.gS with a clear advantage foc Black) 20 ... li:lxd3 21 cxd3 i.xe4 and Black has the same type of advantage as in the game.

2b) 16 .. lDc4! (even stronger than line 2a3 above) 17 'ifh5 (17 i.xc4 :Xc4 followed by ... 'i'a8 gives Black strong pressure) 17 ... h6! (17 ... tJ 18 11Fh6�xb2is line 2a3) 18i.cl �de3! with a large advantage fer Black.

16 ... .bdS Now Black has a safe advantage.

The exchange ofknights has extin· guished White's hopes of a successful kingside attack, and his position is sncturally much worse.

17 :1ae1 i.c4 (D)

18 00 The sacrifice 18 i.xh7+? 'i>xh7

19 "ifh5+ �g8 is not dangerous: 1 ) 20 :Z.f3 g6 21 l:lg3 �g7 22.:fl

(22 �f5+ exf5 23 e6+ i.f6 24 e7

ADAMS - ANAND. HILVERSUM 1993 133

1l'd6 25 exf8'i'+ lbf8 consolidating the extra material)22 ... 11'e8 23 J:lf6 J:l b8 24 'lg5 li:ld7 and Black wins.

2) 20 lZe3! forces a::curate defence:

2a) 20 ... g6? 21 J:lg3 'ireS (the line 2I. .. ixfl 221lxg6+ also draws) 22

J:lxf7 (22 :C6 �g7) 22 ... 11'xf7 23 J:lxg6 'lxg6 24 'l'xg6+ �h8 with a draw.

2b) 20 ... ig5? 21 J:lh3 .lth6 22 lf6! with a dangerous attack.

2c) 20 ... ixfl ! 21 J:lh3 .ltxg2+! 22 'i>xg2 'l'd5+ 23 �gl f5 ! (utiliz-

!he pin) 24 'l'h7+ �7 25 J:lg3 (25 'l'h5+ g6 26 'ilfh7+ �e8 27

'ilfxg6+ �d7 and the king escapes) 25 ... 'i>e8 26 lbg7 J:lc7 is hopeless roc White.

18 .ltxdJ 19 cxdJ (DJ

19 li:ldS? A clear error. Black has too many

pieces wanting to occupy d5, but whereas the queen does not have a

good alternative post, the knight would also be well-placed on a4. Therefore the correct arrangement is queen on d5 and knight on a4, which Black could have achieved by either 19 ... 11'd5! or 19 ... li:la4!, with a distinct advantage.

20 1l'd2! Now I realized my error. I consid

ered 20 ... a5!? 21 bxa5 b4, but after 22 J:lal I decided that the complete dissolution of the queenside would not help Black's winning chances. However, this line might have been objectively best in that Black would preserve a slight edge.

20 •.• J:la8 Perhaps 20 ... li:lb6!? is best, re

verting to the correct plan, although this would be psychologically difficult to play.

21 lL!d4 .ltgS 22 'irr2 (DJ

22 .lth4 23 g3


This may prove to be weakening in the future, but one cannot pretend that it is a serious matter.

23 it.gS 24 h4 it.h6 25 �h2 llc8 26 ..li.cl?? (D)

But this is a serious mistake. 26 lle4? is also bad after 26 ... f5! 27 exf6 (27 l::.e2 "it'd? should also win for Black, because 28 .. .f4 is a threat) 27 ... tt:lxf6, but 26 tt:lb3! is the correct move. White has a good outpost on c5 and this compensates for Black's theoretical structural advantage; the position would be roughly equal.

26 llxcl! 27 llxcl tt:le3 28 Wh3?

28 tt:lc6! was the best chance: 1) 28 ... "it'xd3 29 "it'f3 ! (not 29

"it'xf7+ .:l.xf7 30 tt:le7+ Wf8! nor 29 tt:le7+ Wh8 30 "it'xf7 lllxfl+, as Black wins in both cases) 29 .. . "ifd2+ (29 ... tt:lxfl + 30 .:l.xfl "it'xf3 3 1 :Xf3

is only slightly better for Black) 30 <t>h3 tt:lxfl 31 .l:l.xfl and White has defensive chances.

2) 28 ... tt:lg4+ 29 �g2 (forced,has 29 Wg1 'lrxd3 wins) 29 ... wrxd3 (not 29 ... tt:lxf2? 30 ti:lxd8 liUt.d3 31 lL:B! defending) 30 'llrf3 and again the win for Black is not guaranteed

The text-move plays fer a trap, but if Black avoids this the games is over.

28 ••• ti:lxfl! Avoiding 28 ... 'ilrxd4?? 29 "l'xflt

.l:l.xf7 30 .l:tc8+ lUI! 3 I l:tfld'B#. 29 lhrt 11fd5! (D)

A dream position for Black. White has several weak pawns,an exposed king and passeively placed pieces. Black only needs kJ bring his rook into play and White will start shed· ding pawns.

30 g4 g6 31 .:tel llc8 32 ti:lfJ llc3 33 gs .trs

ADAMS - ANAND, HILVERSUM 1993 135

34 J:l.e3 :xa3 39 'ii'cl 'l'c4 35 d4 :xe3 40 'ifbl a3 36 'ilbe3 .. bb4 0-1

37 �g3 aS The finish might be 41 'l'e4 a2 42 38 �f4 a4 iVaB+ �fB.

I haven't played much club chess, the total being the one year I played for Lyons. That year was quite pleasant because I just turned out for the big matches.However, I wasn't especially excited by club chess,so I wasn't too dissapointed when the club just folded up - at least I didn't have to resign from the team! I might play for a club again in the future, but there is no im mediate prospect of this.

The next game is from the PCA Qualifier event held in Groningen during December 1993.

Game 26

V. Anand - A. Beliavsky PCA Qualifier, Groningen 1 993

Sicil ian, Najdorf

1 e4 cS 2 �f3 d6 3 d4 cxd4 4 �xd4 �6 S M a6

Round about this time Beliavsky had started playing the Najdorf and it was quite clear why - he had been one of Kasparov's seconds for his match against Short. Since Short was one of the players who popularized 6 .i.e3, they must have looked at it very deeply. Unfortunately, I didn't have anything else against the Najdorf, so I decided to play it in any case.

6 .i.e3 e6 7 f3 bS 8 g4 h6 9 'ird2 �bd7

Beliavsky doesn't repeat Ftafnik's imprecise move-order - see Game 23.

10 0-0-0 .i.b7 1 1 h4 b4 12 �ce2 dS

The big question was whether to play 13 .i.h3 or 13 .!Llg3. The former seemed rather foolhardy, as it involves various pawn sacrifices. However, when we had the position after

12 ... d5 on the board, I just decided to gamble.

13 .i.b3!? White is committed because quite

play doesn't work, fc:r example 13 exd5 �xd5 14 �4 �e3 15 'ltxe3 'ilt'b6 16 .i.c4 0-0-0 17 ll'ld3 1fc7 was slightly better fc:r Black in SaxAnand, Philadelphia 1986.

13 dxe4 14 g,5 (D)

14 '" hxgS Beliavsky played all these moves

more or less instantaneously. text-move is the best move-order, as after 14 ... exf3 the reply 15 gxf6! fxe2 16 'irxe2 is very dangerous for Black, for example 16 .. . i.xhl 17

ANAND - BEUAVSKY, PCA QUAUFIER, GRONINGEN 1993 137

ifue6 or l6 ... 'i'xf6 17 l:thfl. Preparation foca World Championship match needs to be extremely thorough - one must not only take the existing theory a little bit further, but in fact almost reinvent the lines you eJqJett to use since it must withstand several months of scrutiny by a team of grandmasters. I had noticed this linle detail when I looked at the line, but didn't expect that players such as Kasparov and Beliavsky would have missed something like this when preparing for a world championship match!

15 hxgS exfJ 16 �4

Now 16 gxf6 fxe2 17 'i'xe2 'i'xf6 18 :hfl 'l'e5 ! is fine for Black, as the bishop on h3 is hanging.

16 �g3 is a major alternative, but in this book I am avoiding getting embroiled in opening theory.

8

16 � 17 'i'e1 (D)

This was all theory and I was wondering when his novelty was going to appear.

17 ••• f2 This was it. but it became clear the

following year that it was not best. The alternative 17 ... ll'lxg5? is bad:

1 8 ll'ldxe6! fxe6 1 9ll'lxe6! ( 19 .ixe6 is also possible: 19 ... ll'lxe6 20 l:txh8 'i'f6 2 1 l:td6! 0-0-0 22 ll'lxe6 l:te8 23 "ifxb4 f2 24 "ifc4+ �b8 25 "ifc7+ �aS 26 :Xa6+ 1-0 Romero HolmesTukmakov, Wijk aan Zee 1991) 19 ... ll'lxe6 20 .be6 :Xhl 21 'i'xhl ll'lr6 (2l..."ifa5 22 .ixd7+ � 23 "ifh2 is clearly winning for White) 22 l:txd8+ l:txd8 23 'i' gl and White has a clear advantage.

After my success in the Beliavsky game, I continued playing the line as I felt that if I had won against Kasparov's preparation, it must be good for White. However, the following year I played the variation once too often and Ljubojevic brilliantly refuted White's play by 17 ... :Xh3! 1 8 ll'lxh3 e5 1 9 ll'lb3 ( 1 9 ll'lf5? "ifa5 20 �b1 .ids 21 a3 l:tb8, threatening 22 ... ll'lc3+, and 19 ll'!xf3? "ifa5 20 �b1 ll'lc3+ 21 bxc3 .ixf3 are both very good for Black) 19 ... a5 with a clear advantage for Black, AnandLjubojevic, Sicilian theme tournament. Buenos Aires 1994. I won this game anyway, but only due to Ljubo's blunders in time pressure. At the present time the line remains unplayable for White.


Actually the Ljubo near-disaster was the result of a lapse of memory on my part. Ljubo had come to Groningen and after the Beliavsky game he mentioned 17 ... l:lxh3 ! to me, but later I imagined that it was Van Wely who suggested it. I therefore felt it was safe to play the line against Ljubo several months later, but not surprisingly Ljubo was baffled as to why I had played the line against him when he had already told me the refutation.

These days I note down not only the ideas, but also who told them to me!

18 .bf2 'ifxgS Not 18 ... lt.lxf2? and White has a

pleasant choice: I) 19 .i.xe6! fxe6 (19 ... l:lxhl ? 20

..11Lxf7+! �xf7 21 'iVe6#) 20 il.ldxe6 'We7 21 lhh8 lt.lxdl 22 lt.lg6 winning.

2) 19 lt.ldxe6 fxe6 20 'l!lxe6+ 'l!le7 2 1 'l!lg6+ �d8 22 l:lxd7+ 'ii'xd7 23 'l!lb6+ �e7 24 lt.lg6+ is also deci-sive.

19 .l1Le3 .h4 After 17 ... f2 I was of course wor

ried about my preparation, as if this was indeed part of Kasparov's world championship preparation, then I could expect that it would be very well analysed. However, this move really surprised me, because I saw that I could reply 20 lL\dxe6. The more I looked at it, the better it seemed, and it appears that the Kasparov

team must have overlooked some thing in their analysis.

19 ... 1Ve7 20 il.ldxe6 l:l.xh3 21 lt.lxg7+ .i.xg7 22 .l:.xh3 was a possible alternative, with an unclear potion.

20 �e6! (D) Not 20 'l!le2 lt.lg3!.

20 ••• 'libel! Beliavsky played this move very

quickly. The alternative is 20 .. fxe6 21 .i.xe6 'ifxhl (2L'irxel 22 hd7+ �n 23 iLe6+! should win for While after 23...�6 24 .i.d4+ IPg5 25 llhxel �xf4 26 l:tfl+ IPg5 27 l:l.f5+ or 23 ... �e8 24 l:lhxel) 22 .ixd7+ �f7 (22.. .�d8 23 .i.b6+ 1Pe7 24 W'xb4+ and 22 ... �e7 23 'i'xb4+ are hopeless for Black) and now:

I) 23 W'xhl l:lxhl 24 1hhl �5 25 l:ld I l:ld8 is unclear.

2) 23 .i.e6+ and now: 2a) 23...�f6 24 .i.d4+ lttgS 25

'il'e3 1Wxdl + (25 . . . l:l.h2 26 lLldS+ �h5 27 .i.xg7! 'ifxd I+ 28 lftxdl

ANAND - BEllAVSKY, PCA QUAliFIER. GRONINGEN 1993 139

llhl+ 29 �e2 and White wins) 26 'ifadl :hl + 27 'itte2 llh2+ 28 lLlg2+ 'ittg6 29 �f5+ 'ittxf5 30 'ili'f4+followed by 3 1 'l'xh2, winning.

2b) 23...�e8 24 'ili'xhl l:txhl 25 :xhl is slightly better for White.

3) 23 'i'e2! and now: 3a) 23..."ii'h6 24 'illc4+ 'itte7 25

lc7 wins. 3b) 23...�d6 24 'ii'c4+ 'itte7 25

lilg6+ �xd7 26lLle5+ 'ittd8 27 �b6+ �e8 28 Wif7#.

3c) 23 ... 'i'xd1+ 24 'ili'xdl �d6 25 If! ! and the two rooks are not a match for White's queen, for example 25 ... ..txf4 26 'ili'xf4+ 'itte7 27 �h3 with a very strong attack.

21 lLlxg7+! Not 21 lLlc7+? 'ittd8 22 :hxel

.:xh3, which favours Black after 23 lilxa8 :Xe3 ! 24 llxe3 �xa8 or 23 1b6 :h6.

21 ••• �d8 (D) 2l...i.xg7 22 �xd7+ �f8 23

.:hxel leads to a slightly better ending foc White.

22 llhxel! After 22 �b6+?! 'itte7! 23 lLlf5+

(23 llhxe1 lLlxb6 wins for Black) 23 ... 'ittf6 (not 23 ... 'itte8 24 llhxe l lLlxb6 25 �g2 and White retains some advantage) 24 �d4+ 'ittg5 25 lldxe1 llxh3 26 lLlxh3+ 'ittxf3 27 l:thfl + 'ittg6 28 llg 1 + the position resolves to perpetual check.

22 llxh3 23 lLlxh3 �xg7 24 lLlgS! (D)

Now I was really happy-White is going to get Black's last kingside pawn and as a result can play for a win at absolutely no risk. Did Beliavsky miss that 24 ... lLlxg5 is met by 25 �b6+?

24 ••• �e8? (D) Not the best square for the king.

The alternatives are: 1) 24 ... lLlxg5?? 25 �b6+ �c8 26

lle8#. 2) 24 ... �e7?? 25 lLlxe4 .i.xe4 26

�g5+ wins.


3) 24...�6 25 �xf7+ rj;c7 26 ..if4+ Wb6 27 lte6+ is clearly better for White.

4) 24 ... Wc8! 25 �xf7 a5 was the correct choice, with just a minimal advantage for White. Black should draw but White can still poke around for a few moves.

25 �! Now I felt that I was going to win

this game. Black is in considerable difficulties.

25 ••• �e5 Here the alternatives make grim

reading: 1) 25 ... ltc8? 26 �6+ �xd6 27

lbd6 wins. 2) 25 . .. Wxf7 26 ltxd7+ is also

hopeless for Black. 3) 25 . .. ..tf8 26 �g5 �df6 27

�xe4 �xe4 28 ltd4 ..ie7 29 ..if4 �f6 30 ..id6 �g8 3 1 l:.xb4 ..ic6 32 l:.g4 with excellent winning chances for White, much as occurs in the game.

4) 25...�6 26 �6+ lDxd6 27 .:.Xd6 and White has a clear plus.

Or:

26 �d6+! lbxd6 27 :xd6 'itnl?! (D)

1) 27 ... 'it;>f7? 28 ..td4:e8 (White wins after 28...�f3 '19 :d7+) 29 ..ixe5 ..ixe5 (or 29 ... 1lxe5 30 :XeS ..ixe5 31 l:ld7+) 30 l:td7+ �6 31 l:lfl +! We6 32 .:.Xb7 and wins.

2) 27 ... �f3? 28 l:e2is very good for White.

3) 27 ... ltd8! (probably Black's best chance) 28 l:lb6 gives White distinct advantage.

28 ..ih6! Very strong. IfWhite can exchange

bishops and win the b-pawn, then Black will be facing defeat, since Black's king willbe too far away to defend the queenside. 28 .i.d41 is less accurate on account of28 .... l:.e8.

28 ••• .i.xb6+ 28 ... l:le8 loses to 29 ltf6+ �g8 30

ltgl l:le7 3 l llb6.

ANAND - BEUAVSKY, PCA QUAUFJER. GRONINGEN 1993 141

29 :xh6 Ci'Jf7 30 l:.b6 il.dS 31 l:.xb4 l:.c8

3 1 ...a5 might have made things slightly more difficult, but would not fundamentally change the position.

32 l:.b6 aS 32 ... l:.c6 33 l:.xc6 il.xc6 34 �d2

is a simple win. 33 a4! (D)

Not 33 l:.b5? .ba2 34 b3 a4 35 �b2 axb3 36 cxb3 il.xb3 drawing.

33 •••

34 b3 �c6 il.d7?!

Once again Black could have dragged the game out by defending more accurately with 34 ... lt:ld8, but in view of White's material advantage and Black's poorly placed king. White should win in the long run. I was just going to play quietly with 35 �b2, but White still has to break Black's resistance.

35 �b2 l:l.cS 35 ... lt:ld8 36 l:.f6+ drives the king

even further away. 36 .l:lbB+ �g7 37 l:l.b7 il.cB

Or 37 ... il.f5 (Black also loses after 37 ... .1:ld5 38 .l:le7 and 37 ... �c6 38 .l:lc7) 38 c4 �6 39 .l:lb5 lt:le5 40 �c3 lt:ld7 41 �d4 and the queenside pawns will decide.

38 l:l.bS 1-0

The importance of this game, which was played in round five, cannot be overestimated; it was the encounter which put me on the road to qualification for the PCA Candidates. Before it, Beliavsky was on +4 and I was on +2. This win propelled me towards the lead while Beliavsky, who had managed a fantastic start,subsequently collapsed and failed to qualify. The following game was played the round after the Beliavsky game.

Game 27

J . Benjamin - V. Anand PCA Qualifier, Groningen 1993

Sicilian, Richter-Rauzer

1 e4 c5 2 �f3 d6 3 d4 cxd4 4 .fud4 liJr6 5 li)c3 llk6 6 j_gs e6 7 'i!Vd2 j,e7 8 0-0-0 0-0 9 .'Llb3 'MI6

I had prepared the Classical Sicilian to be my main defence for this tournament, and in view of Game 4 in this book, it was ironic that we entered the same variation, but with colours reversed.

10 f3 l:td8 11 �b1 (D)

11 flc7

An interesting move which forces White to decide how he is going to prevent ... d5.

12 .hr6! When Benjamin took on f6 I just

couldn't believe it - I hadn't considered this move at all, believing that nobody would voluntarily give up the dark-squared bishop! However, it is a reasonable move; in return for surrendering an important minor piece, White gains time to push his kingside pawns. There are a number of alternatives, for example 12 1Llb5, 12 j,£4 or 1 2 h4, but we will leave the relative merits of these to a book

on opening theory. 12 ..• .hr6 13 g4 g6

This is a critical moment for Black There is an argument for 13 ... g5, for example 14 h4 h6 1 5 hxg5 hxg5 16 l:th5 fle7, followed by ... 1Lle5-g6, dominating the dark squares on the kingside. However, this involves a certain amount of risk, as if the position opens up. the exposed state of Black's king may be more important than control of a few dark squares. Black can continue with 13. .. a6!? 14 g5 j,e7, but after 1 5 f4White

BENJAMIN - ANAND, PCA QUAliFIER, GRONINGEN 1993 143

prpbably has a slight advantage • I preffered the text-move because it seemed to me that Black's queenside attack, supported by the bishop on the long diagonal, would be very dangerous. I still hadn't taken Benjamin's idea seriously, but the next few moves show that it is not so easy for Black.

14 h4 a6 1S gS!

Now 15 h5?! would be bad, as 15 ... g5! 16 h6 'ile7 seizes the dark squares without opening the h-file.

15 i..g7 16 hS bS 17 hxg6 hxg6 (D)

18 f4! After 18 i..d3 lbe5 19 f4 lbxd3 !

(not 19 ... .!2x4 20 i..xc4 'l'xc4 21 'l'h2 with a dangerous attack) 20 cxd3 (20 lxd3 b4! 21 li)e2 a5 is similar) 20 .. . b4 2l lbe2 a5 the position is unclear,with both sides having attacking chances·

The text-move threatens 19 f5, which would win as Black cannot reply 19 ... exf5 because of 20 li)d5. Thus Black is forced to drive the knight away from c3.

18 ••. b4 19 �4! (D)

A very comfonable square for the knight, blocking Black's queenside attack. 19 li)e2 a5! would be much weaker.

19 ... :b8 After 1 9 ... e5 20 fxe5 ! (20 i..c4

exf4 21 'l'xf4 lbe5 22 i..d5 i..b7 is unclear) 20 ... li)xe5 (both 20 ... dxe5 21 i..d3 and 20 ... i..xe5 21 i..c4! favour White) 21 11fxb4 Black has insufficient compensation, but at least he doesn't have to worry about his king ! In fact this drastic remedy might be necessary; I just hadn't realized how critical my position was.

20 1fb2 This is one way to build up on the

h-file, but it was also very dangerous

144 VISHY ANAND: MY BESI' GAMES OF CHESS

to play 20 i.d3 !?, followed by J:[h4, 'ifh2 and J:[b I .

The queen on h2 i s not only effective down the h-file; in some lines it can act along the h2-b8 diagonal, for example by f5 followed by lLlac5.

20 WfS (D) Time to get the king out! The al

ternatives are: I) 20 ... e5? 21 f5 gxf5 22 'l'h7+

�f8 23 exf5 and White wins. 2) 20 ... i.d7 21 l:l.d3 lLle77 (the

best move is 2 l ... �f8, transposing to the following note) 22 l:l.h3 �f8 23 l:l.h7 ! lLlg8 24 f5 !? (24 l:l.xg7 �xg7 25 'l'h8+ �8 26 l:l.h7 is less clear after 26 ... �e7) and now:

2a) 24 ... i.e5 25 J:[xf7+ �xf7 26 .. h7+ i.g7 27 fxg6+ � 28 i.d3 with a decisive attack.

2b) 24 ... i.xa4 25 l:l.xg7 �xg7 26 'ilfh8+ �8 27 l:l.h7 i.e8 28 f6 forcing mate.

2c) 24 ... exf5 25 lLlac5 i.c8 26 l:l.xg7 �xg7 27 'iWh8+ �8 28 l:l.h7 winning for White.

21 l:l.d3 eS 2l ... i.d7 was also possible, for ex

ample 22 l:l.h3 (22 f5 exf5 23 liJacS is unclear) 22 .. .'�e7 ! 23 'l'e2 (23 l:l.h7 l:l.h8 24 'l'f2 lh-lh Landenbergue-Georges, Swiss Team Championship 1994) 23 ... e5! (23...<i'f824 l:l.h7 lLle7 25 "i'h2 and 23...ltla5 24 f5 i.xa4 25 f6+ i.xf6 26 gxf6+ �xf6 27 .. e3 i.xb3 28 l:.f3+! are good for White) 24 l:l.b7 l:l.h8withan unclear position.

22 rs 22 i.h3 exf 4 23 'l'xf 4 lDe5 i!i fine

for Black. 22 ••• gxfS 23 l:l.h3! liJe7?! (D)

23 .. .f4? 24 :b7! and 23 .. .fxe4? 24 l:l.h7 i.e6 25 l:l.xg7 win for White but 23 ... �e7! 24 l:th7 l:l.g8 was a better defence.

24 l:l.hB+ A critical moment At first sight

24 l:l.h7 appears very dangerous.but Black can defend:

BENJAMIN - ANAND, PCA QUAUF/ER, GRON/NGEN 1993 145

I) 24 ... �g8? 25 .:l.xg7 �xg7 26 'l'h8+ �f8 27 :h7 ..i.e6 28 exf5 :deS 29 ex;3 bxc3 30 fxe6! and wins.

2) 24...�g6! 25 l:l.xg7 �xg7 26 'l'h6+ �g8 and now Black is better. The continuation might be 27 ..i.e2 l'e7 28 ..i.h5 "ill'f8 29 ..i.xg6 'ifxh6 30 :xh6 fxg6 3 1 J:xg6+ �h7 with a winning ending.

24 ... �g8 Not 24 ... ..txh8? 25 1i'xh8+ lLlgS

26 lh7 transposing to line 1 of the previous note.

25 :Xg8+! After 25 1i'h 7 ..i.xh8 26 1i' xh8

l'e7! the attack peters out.

8

l5 •.• �xg8 26 1Wh7+ � 27 exf5 (D)

17 ••• ..txrst 27 ... '1We7!'? is possible, but White

is slightly bener after 28 f6 ..i.xf629 gxf6 'llrxf6 30 .th3.

18 l'xf5 'it'c6

29 g6! l:tb7 30 l:th7?!

Later on the computer Deep Blue suggested 30 "ill'g5!. Black's best appears to be 30 ... l:tc8 31 1i'g2 l:bc7 (3l ... e4? 32 lLla5 really does win) 32 1i'xc6 J:xc6 33 ..i.d3 d5!. White is better, but his minor pieces are not well placed and so Black might be able to generate a lot of counterplay.

Coincidentally, Benjamin later ended up working for the Deep Blue team in their matches against Kaspa-rov!

30 1i'xa4 (D)

31 'ti'g5?! Here is Wbite's last chance to

force a clear-cut draw, for example: l ) 3 l ..i.c4 d5 (3l...'it'd7 32 'it'h5

is unclear, but 3 1 ...ltdd7 32 1i'f3 is probably good for White) 32 0c5 'it'c6 33 lllxb7 'it'xb7 34 .ixa6 'it'd? leads to a drawn ending.

2) 31 l:lxg7 �xg7 32 gxf7 and now:

U6 VISHY ANAND: MY BEST GAMES OF CHESS

2a) 32 ... l:xf7? 33 ... g5+ �h7 34 �d3+ wins for White.

2b) 32 ... l:lf8? 33 •g4+ �xf7 34 �c4+ �6 35 ... f3+ is also winning.

2c) 32...�f8 33 ... g5 'it'd? (not 33 ... l:bb8 34 �c4 d5 35 'it'g8+ 'l;e7 36 Wg7 llfil 37 &5 and Black loses) 34 �c4 d5 35 'it'g8+ �e7 36 'it'g5+ 'l;xf7 37 "ihD+ �g8 38111 g5+ draws.

2d) 32 ..• 'Wd7 33 ... g5+ 1Pxf7 34 �c5 dxc5 (34 ... 'it'e7? 35 'it'h5+) 35 �c4+ with perpetual check.

31 ..• 'it'e8 White still has an amazing number

of attacking possibilities, but Black seems to be able to hold out with accurate defence. Here he must avoid 3 I.. .llc8 32 llxg7 �xg7 33 gxf7+ �xf7 34 'it'f5+ and 3 J ...:e8 32 :Xg7 �xg7 33 gxf7+ �xf7 34 �c4+, with a win for White in either case.

32 �xa6 (D)

After 32 �g2 !? (32 <t:la5 llc7 33 <t:lc6? f6 wins for Black) Black again must be very careful:

1) 32 ... 'it'e7 33 11fb5 wins. 2) 32 ... llc7? 33 :Xg7 �7 34

gxf7+ �xf7 35 �d5+ also wins. 3) 32 ... e4? 33 l:lxg7 �xg7 34

gxf7+ �xf7 35 'Wd5+ gaining mate. rial.

4) 32 ... d5! 33 <t:lc5 'ii'e7! defends. 32 l%e7! (D)

33 ..i.d3 Threatening to win by 34 lhg7

�xg7 35 gxf7+ 1ixf7 36 �g6+. 33 ... e4! 34 �bS

White has nothing better, fix example:

1) 34 �xe4 l:lxe4 35 gxf7 l'lel+ 36 <t:lcl llxcl+! 37 'ilfxcl 'i'e5 defends.

2) 34 .i.c4 d5 35 hd5 :e5 36 gxf7 :Xg5 37 fxe8'i'+ :Xe8 wins for Black.

3) 34 nxg7 Wxg7 35 gxf7+ lftxf7 36 �c4+ d5 37 hd5+ :xdS 38 11Vxd5+ �f8 with good winning chances for Black.

BENJAMIN - ANAND, PCA QUAUFIER, GRONINGEN 1993 147

34 •••

Forcing the following liquidation. 35 gxf7 %bg5 36 fxe8'if+ %:lxe8 37 i.xe8 l:tg1+! 38 li:lcl <t>xe8 (D)

39 a4?? A blunder cause by time-trouble.

White had to try 39 c3 (getting rid of the c-pawn to release the king) 39 ... bxc3 (39 ... b3? 40 axb3 ii.f6 4 1 <t>c2 i.g5 42 ttle2 J:.el i s ingenious. but leads to less than nothing after 43 %:lh2!) 40 bxc3 ii.xc3 41 <t>c2 ii.f6, when Black is clearly better, but the reduced material gives White some drawing chances.

39 .•. bxaJ 40 bxaJ ii.cJ!

With total paralysis. 41 %:lh4 dS

0-1

This game gave me the clear lead and. unlike at Biel, I reached the necessary score very comfortably.

At the beginning of 1994 I played my first-round match in the ADE Candidates cycle against Artur Yusupov, which I won 41h-21h. In the draw for the second-round matches I was paired against Gata Karnsky. Before the match took place, Gata and I participated in the Linares tournament and we met in the very first round. In view of the forthcoming match. the game had more importance than a normal tournament game. In fact we were to spend a lot of time with each other in the following years ...

Game 28

V. Anand - G. Kamsky Linares 1994

Sici l ian, Najdorf

1 e4 2 lff3 3 d4 4 llJxd4 5 llJcJ

c5 d6 cxd4 llJf6 a6

At the time Kamsky played many openings, but I wasn't sure whether he had really studied them or whether he gave priority to being difficult to prepare for. Later it became clear that he does study a lot and understands a lot of different openings, but not too well! I often encountered holes in his repertoire.

6 ..te3 e5 7 llJb3 ..te6 8 f3 ..te7

The thematic Sicilian thrust 8 ... d5 does not equalize here: 9 exd5 llJxd5 10 llJxd5 ..txd5 1 1 c4 ..tb4+ 12 Wf2 ..te6 13 11'xd8+ Wxd8 14 l:1dl+ gives White an advantageous ending.

9 1l'd2 llJbd7 10 g4 h6

JO . . . b5 is less accurate as White may continue I I a4! b4 12 llJd5 ..txd5 13 exd5, when 1 3 ... llJb6 fails to 14 a5! llJbxd5 15 g5 llJxe3 16 gxf6 llJxfl 17 fxe7 V/xe7 18 J:lxfl and White wins a piece.

11 h4 bS (D)

w

12 J:lg1 This idea, which came to me over

the board, is borrowed from the game against FtaCnik (Game 23 ). Although the position is completely different the idea is the same - White save a tempo by missing out 0-0-0 and uses it to push through g5 as quickly as possible.

12 ••• b4 12 ... llJb6!? 13 g5 hxg5 14 hxg5

llJfd7 is another possible continua-tion .

13 llJa4 Later it turned out that 12 :gl had

been played before, only to be followed up by the weak 13 ti:Je2?, when 1 3 ... a5 14 g5 llJh5 gave Black

ANAND - KAMSKY, LINARES 1994 149

a fine position in Los-De Boer, Groningen Open 1990.

13 ... d5 (D) l3...a5 14 g5 hxg5 15 hxg5 ltili5

is unclear.

14 g5 d4 It is hard to judge how the disap

pearance of the h-pawns affects the position. After l4 ... hxg5 15 hxg5 d4 16 hd4 (16 gxf6 dxe3 17 'l'xe3 10xf6! and now White has to play 18 .td3 in order to castle; the resulting position is unclear) Black can try:

I) 16 . .. .txb3 17 gxf6 .txf6 18 axb3 exd4 19 0-0-0 �5 with the same position as in the game, minus the b-pawns. Certainly the h-file is bad for Black if he castles, but in some other lines it is useful. Still, 20 f4 looks good for White anyway after 20 . ..Ii�f3 21 'l'g2 lllxgl 22 1i'xgl !

(here 22 eS allows 22 ... .i.xe5 ! 23 ne5 li:lb3 and the knight escapes) and I think Whitehas great longterm com pen satim.

2) l 6 ... fue4 17 fxe4 .i.xb3 18 axb3 exd4 (here the interpolation of ... hxg5 favours Black -compare the note to Black's 15th move in the game) 19 0-0-0 �e5 is unclear.

15 .txd4! Not 15 .i.f2? �h5 and Black is

better. 15 ••. .txb3 (D)

After l 5 ... �xe4 16 fxe4 .txb3 17 axb3 (not 1 7 .tb6 �xb6 and Black has no problems) l 7 ... exd4 18 1i'xd4 hxg5 19 0-0-0! ( 19 hxg5 .i.xg5 20 1Wxg7 .i.h4+ 21 �e2 .i.f6 22 1i'g3 is unclear) White seems to have a pleasant edge, for example after l9 ... .l:xh4 both 20 1Wxg7 .i.f6 21 'l'g8+ �e7 22 l:lxd7+ �xd7 23 1i'xf7+ and the simple 20 .i.c4 are very good for White.

16 gxf6 1 6 axb3 �xe4 17 fxe4 is also

promising, transposing into the note to Black's 15th move.

16 .bf6


16 ... .i.xa4? 17 fxe7 W'xe7 18 il.f2 is good for White.

17 axb3 exd4 18 0-0-0

Now White threatens simply f4 and e5, with an overwhelming position, so Black has to take counter-measures.

18 .•• li)eS (D) The critical line as ! 8 ... i.xh4? 19

Wxd4, 18 ... 0-0 19 'liVxh6 and 18 ... Wa5 19 f4 (followed by e5) are all very good for White.

19 f4!! The most forcing continuation.

The alternatives are: I) 19 i.g2 i.xh4 20 f4 0-0! is not

very dangerous for Black. 2) 19 'l'g2 �g6 20 h5 �4 21

'l'g4 i.e5 22 ltlc5 0-0 23 '1'£5 with an edge for White.

3) 19 .te2. intending 20 f4, is also promising as 19 . .. d3 may be met by 20 'l'e3!.

19 �f3

20 ... g2 �gl 21 e5 0-0

Mter 2 l ...i.xh4 22 ifxg7 llf8 23 ... xg I , followed by i.g2, :Xd4, etc. White has an enormous attack.

22 i.d3! White wants to take on gl with his

rook, so as to tie Black down to the defence of g7. After 22 exf6 1Wu6

23 'liVxgl 'l'xf4+ 24 rt>bl '1'xh4 White is not as well place to attack Black's king as after 22 .id3.

22 ••• i.xeS! White wins after 22 ... .ixh4 23

llxgl g6 (or 23 ... g5 24 1i'h3!) 24 i.xg6 �h8 25 i.h7! i.g5 26 fxg5 'ifi>xh 7 27 'lfe4+ 'ifi>h8 28 'i'h4 wgB 29 •xh6, followed by :l:thl or g6.

23 fxeS 'ii'xh4 24 l:xg1 (D)

24 .4+ The ending after 24 ... l'g5t 25

'ii'xg5 hxg5 26 l:xg5 is favourable for White.

25 �b1!

ANAND - KAMSKY, UNARES 1994 151

25 1ld2 was also possible, but the text is more incisive.

25 1i'xeS (D)

Threatening 28 _.xh6. 27 ••• 1i'e3

Or 27...�h8 (27 ... l:e8 28 li:ld7 'ife3 29 li:lf6+ wins) 28 li:ld7 _.e3 29 li:lxf8 ! 'ifxg1 + 30 'iPa2 (the threat is 3 1 _.c8) 30 ... g6 31 'ifc5 :as 32 'ife5+ 'iPg8 33 li:lxg6 fxg6 34 'ifd5+ and wins.

28 l:g2 29 l:e2 30 �a2 (D)

'it>h8 1i'gl+

Now the threat is 3 1 l:e8 'iPg8 32 l:xf8+ Wxf8 33 '1fd6+ mating. Black has still not been able to coordinate his pieces and, indeed, in order to meet White's threat he is obliged to

In this position Black has a rook retract his 26th move. and three pawns for a bishop and a knight, and there isno obvious way • • • • for White's attack to break through, M m � fB & '�" "

so it might appear favourable for B • R m • • Black. However, it turns out that the A ••• • • most important factor is the initia- • � • • live. White can keep harassing Black E E fB • before hecan coordinate his rooks. • • � m

26 llkS! :a' m £3:. m..tm m Black could have removed the �9 £3:. • J: rl •

danger to his king by jettisoning a • • • �� couple of pawns: 26 ... l:ad8 (26 ... l:fd8 27 10d7 •g5 28 'i'hl ! is also very good for White) 27 li:lxa6 and now:

I) 21 .. .:reS 2s li:lxb4 g5 29 :n (not 29 liJc6?? 'ife1 + mating), followed by �c6.

2) 27 .. .:as 28 �xb4, followld by �c6 and b4-b5, when White should win.

rt 'i'c6

30 Now White aims to lransfer his

knight to e5, when the weakness of f7 will tell.

31 li:ld7 l:ac8 Black loses after 3l ....:t'c8 32 'ii'B

or 31 . ..J:Ud8 32 ttJe5. 32 'l!fn rs


After 32 ... 1:!.fe8 33 ll:le5 the f7-square is fatally weak.

33 lLJxf8 lieS (D) Or 33 .. .J:%xf8 34 1i'b7 and White's

initiative is too strong. The textmove is a dangerous try, but White has a good reply ready.

34 ll:lg6+! The simplest method, although 34

1i'xf5 1:!.xf5 35 .ixf5 1i'fl (35 ... g5 36 l:te7 and 35 ... �g8 36 .ig6! also win for White) 36 ll:lg6+ �h7 37 �h4+! would also have b��en effective.

34 - 'l'xg6 After 34 ... �h7 White prevents

the mate by 35 .:.e5. 35 l:!.el 'i!ff6 36 .. a8f. Wh7 37 .ic4

White's attack is too strong. 37 ••• :a; 38 1i'g8+ Wg6 39 l:!.gt+ 1-0

As 39 ... �h5 40 .ie2+ �h4 41 .. d5 leads to mate.

Although the above game was a good start, the rest of the tournament didn't go so well -I barely made 50%, which amounted to a fairly lousy result. The following game was the only other high spot - but it was quite pleasant!

Game 29

V. Anand - J. Polgar Linares 1 994 Sicilian, Najdorf

1 e4 cS 2 lill'3 e6 3 d4 cxd4 4 lillcd4 li:lf6 s li:lc3 d6

Giving White the chance to play a Scheveningen, but I decided to transpose into the English Attack.

6 ..te3 a6 7 f3 li:lbd7

I was surprised that she chose the ... <t:lbd7 system, because an earlier game d Kasparov had seemed to show that White could gain a strong initiative. However, Judit comes up with an improvement.

8 g4 h6 9 l:tg1 bS!

Better than 9 ... 11Vb6?!, when the continuation 10 a3 li:le5 1 1 .tf2 ilc7 12 f4 <t:lc4 13 i..xc4 'ifxc4 14 'iVf3 e5 15 <t:lfs i.xf5 16 gxf5 dS 17 fxe5 <t:lxe4 18 l:tg4 was very good for White in Kasparov-Kamsky, Linares 1993.

10 h4 li:lb6 IO ... g6 has been played before, so

this is the innovation. 11 � lill'd7!

After l l . .. hxg5 12 hxg5 lill'd7 13 1!6, with the idea 13 ... .:th2 1 4 gxf7+

�xf7 15 i..f4!, White has some advantage. After the text, however, 12 g6 is impossible because the h4-pawn is hanging with check.

12 1fe2!? 12 ild2 may be met by 12 ... b4 13

li:ldl dS or 12 ... li:le5. 12 ••• hxg5

After 12 ... li:lc4 13 0-0-0 li:lxe3 14 'l'xe3 hxg5 15 hxg5 g6 16 f4 White's lead in development compensates for the two bishops.

13 hxgS g6 (D) Not 13 ... b4? 14li:lc6 ilc7 15 li:lxb4

dS 16 li:ld3 and White wins a pawn, but 13...ltx:4 14� li:lxe3 15 'iVxe3 g6 16f4 is still playable, transposing to the preceding note.

w


14 0-0-0 lLlc4 15 ..tr2 ..tb7 16 f4 (D)

16 'ilfaS After 16 ... .l:tc8 White also plays

1 7 'iii' g4, with the plan of eliminating the dangerous knight by .ixc4.

17 'ili'g4! Once again White's priority is to

eliminate the knight. Black's queenside play appears dangerous, but she cannot land a ... lllxb2 blow before the knight is swapped off.

Note that 17 li:lb3 'i'c7 is inferior, since next move 18 . . . .ig7 really will threaten to take on b2.

17 ··- 0-0-0? This allows an unusual combina

tion, so 17 ... .l:tc8 would have been better:

I) 18 lllxe6? li:lxb2 is good for Black.

2) 18 .ixc4 .l:txc4 19 lllxe6! fxe6 (19 ... .1:txc3 20 bxc3 fxe6 21 "ifxe6+ �d8 22 .id4 .l:th2 23 .l:thl ! with

advantage to White) 20 11Vxe6+wd8 21 .id4 is unclear.

3) t8.!bb3!? 'irc7 19 .ixc4'1xc4 (19 ... bxc4 20 llld4 and Black's cen tralized king is a ready target) 20 llla5 'irc7 21 lllxb7 'l'xb7 22 id4 leading to a double-edged position.

18 .ixc4 bxc4 (D)

19 li:lxe6! While this sacrifice is perfectly

normal when Black's king is still on e8, it is unusual when Black has already castled queenside. However, here White is not aiming for direct attack but for positional compesa-tion.

19 .•. fxe6 20 'li'xe6 �b8

After 20 ... .ig7 (the continuation 20 .. . d5 2 1 li:lxd5 is also favourable for White) 2 1 .id4! .ixd4 22 :Xd4 'li'b6 23 .l:tgd 1 ! Black will shed some more pawns (note that 23 ... :hi?? fails to 24 .l:txc4+ ).

21 'li'xg6!

ANAND - J. POLGAR, LINARES 1994 155

Gaining a third pawn for the piece. Unless Black develops counterplay quickly the passed g-pawn will decide the game.

21 ... J:l.b3 (D) 21 ...�c5? loses to 22 .txcS 'lixcS

23 'lf6!.

w

22. �bl! The most straightforward way to

nullify Black's threat of . . . J:I.xc3. Instead, foc example, 22 .td4? J:l.xc3 ! 23 ixc3 'l'xa2 would allow unnecessary complications. 22 'i'fS 1i' xf5 23 exf5 J:l.f3 24 J:l.gfl is playable, but more complicated after 24 ... .tg7 ! 25 l0e2 (25 ffi li:lxf6 26 gxf6 .txf6 with just anedgefoc White) 25...�c8! 26 id4 J:l.e8. White is still better, but if Black returns the piece, then the ending would be difficult, so I preffered to keep the queens on.

22. ••• J:l.f3 White wins after 22 ... J:I.xc3 23

.tel 23 .td4

23 J:l.gfl ! would have been more precise - there was no need to give

up one of the pawns.

B

23 lhf4 24 1i'h7! (D)

24 ... dS? Black's best chance was 24 ... lLic5

25 g6 li:lxe4 26 g7 .txg7 27 J:l.xg7 li:lxc3+ 28 �xc3 (28 bxc3 �e4 1eads to an unclear position) 28 ... 1i'd5 29 �c I , although I think that White has a clear advantage. His king is ultrasafe, while Black's is very weak. Opposite-coloured bishops only begin to exert their dra wish tendency once the major pieces have been exchanged, but here all the major pieces are still on the board so Black is going to come under a very strong attack.

After the text-move White wins comfortably.

25 g6 �cs Now 25 ... dxe4 26 g7 �xg7 is in

effective and White wins easily by


continuing Zl :txg7 .ic6 28 1Vh6 'ilc7 29 .ib6.

26 ..b:cS lilll:cS Or 26 ... 'ffxc5 27 g7 ltJf6 28 g8'ff

:Xg8 (28 ... lbxg8 29 :Xg8) 29 :Xg8+ lDxg8 30 'ffxg8+ <tJa7 3 1 exd5 with a simple win.

27 g7 �a7 28 g8'ff

28 :txd5 was also very good. 28 •.. lhg8 29 'ffxg8 lilll:e4 30 lilll:e4 :txe4 31 'ffg7 'ffcS 32 :tgel :tf4 33 :te7 'ffb6 34 'ffgS c3 35 b3 '6'b4 36 :tc7 :te4

36...�b6 37 :txc3.

37 'ffxdS <tJb8 (D)

38 '6'dll+ �a7 39 a3 irxa3

1-0 White mates in five: 40 lbb7t

�xb7 41 :td7+ �6 42 'ffc7+ �b5 43 :td5t �b4 44 'ffb61#.

Game 30

V. Anand - G. Kamsky PCA Candidates (3), Las Palmas 1995

Ruy Lopez, Arkhangelsk

I e4 e5 2 lDf'3 l2Jc6 J .i.bS a6 4 .i.a4 l2Jr6 S 0-0 bS (j .i.b3 .i.b7

Kamsky plays a lot of different openings, and here he decides to play the Arkhangelsk. However, this was no surprise; he played it in the 1994 FIDE Candidates match in Sanghi Nagar, which I lost, and subsequently played it against Short in their PCA Candidates match.

7 :et .i.cS 8 c3 d6 9 d4 .i.b6

10 .i.e3 0-0 Not IO ... lDxe4? I I d5 and White

wins a piece. 11 l2Jbd2 h6 12 hJ

I was following the game ShortKarnsky, PCA Candidates (6), Linares 1994, which continued 12 ... �d7 13 aJ lDe7 14 .i.a2 �h8 1 5 b4 and was eventually won by Black. I had prepared some new ideas in this game but in fact Kamsky was the fisrt to vary.

12 .. b8!? (D)

This looks an odd move, but Black is trying to batter down d4 by ..... a7. The problem is that it allows White to play d5.

13 dS! �e7 The tactical variation 1 3 ... .i.xe3

14 dxc6 .i.xd2 1 5 cxb7 .i.xel 1 6 bxas• .i.xf2+ 1 7 �xf2 •xa8 1 8 l2Jd2 l2Jxe4+ 19 l2Jxe4 •xe4 20 •d5! leads to an ending in which White has a slight advantage. Gata played 13 ... l2Je7 very quickly so he obviously didn't believe this line.

14 .i.xb6 cxb6 IS .i.c2!

After 1 5 a4 bxa4 16 .i.xa4 .. d8 Black has a satisfactory position.

IS �d7

}58 VISHY ANAND: MY BEST GAMES OF CHESS

A critical moment. One of the points of 12 .. Ji'b8 (as opposed to the similar line 12 . . . l:le8 13 d5 li:Je7 14 i&.xb6 cxb6 15 i&.c2 li:Jd7) is that by leaving the rook on f8 Black sets up the immediate threat of ... f5. It is also helpful for Black that his queen defends d6, since after .. .f5 a white knight might easily come to e4.

I decided that it is imperative to stop ... f5, and to this end I was able to use the negative side of .. .'ii"b8, namely that on b8 the queen does not exert a latent influence on the d8-h4 diagonal.

16 li:Jh4! 1i'd8 (D) 16 ... f5 is bad in view of 17 li:Jxf5

li:Jxf5 1 8 exf5 i&.xd5 1 9 li:Je4, so Gata decides simply to return the queen to d8. However, this in itself shows that White's strategy was correct.

17 ljjf1 White can hardly play a weaken

ing move such as g3, so the knight

has to remain undefended on b4 Fortunately, after the text-move Black is not able to exploit Ibis.

17 ••• gS? After 17 ... li:Jxd5 1 8 li:Jf5,followed

by 19 li:Jxd6, White gains the advantage. However, this variation helps us to find the best line for Black, namely 1 7 ... b4! 18 cxb4 ttlxd5. Even in this case the continuation 19ll:lg6! (19 li:Jf5 li:Jxb4 20 li:Jxd6 ttlxc2 21 'ilhc2 is also slightly better for

White) 19 .. ixg6 20 i.b3 gives White some advantage. It is odd that in this variation the doubled b-pawns are liquidated, but Black gets doubled g-pawns instead!

In this match Kamskyhad a tendency to take really drastic measures when faced by minor positional problems, and the text-move is a good example. Pushing the g- pawn clarifies the position, but at the cost of seriously weakening Black's kingside.

18 00! (D) A difficult choice as 18ttlf5ll:lxf5

19 exf5 li:Jf6 was also tempting. After, for example, 20 i.e4 ltc8 21 f3 l:lc5 22 li:Je3 �g7, White has ce-mented everything but the position becomes rather closed and I was'nt sure that I would be able to break through on the kingside later.

The text-move is more complicated, but promises a larger advantage if everything works out. One factor in the decision was that at this

ANAND - KAMSKY. PCA CAND. (3). LAS PAIMAS 1995 159

point of the match I was a point down and badly need a win.

1s ... rs After 18 .. .'it>g7 19 ltle3 White

again prevents .. .f5. 19 exrs ll:lxrs 20 �h2!

A key move, with the idea of regrouping the knight via g4. White intends using all the light squares and the weakness of the bl-h7 diagonal. Black is unable to mount a counter-attack against d5 quickly enough to deflect White from his plan.

20 ••• 1i'f6 After 20 . .. ltle7? 21 ltle3 White

consolidates his grip. 21 lCe4 'ii'g7

The queen comes across to support the weakened kingside.

22 ltlge3! After 22 �fe3 li:Je7 the knight on

g4 hasnowhere to go. 22 lbxe3

Now, however, 22 ... ltle7 23 ltlg3 leaves both knights ideally posted.

23 �e3 Black is in a very bad way be

cause it is almost impossible to defend f5.

23 ••• .IH4 (D) After 23 ... :£6 24 1i'h5 1Wf7 25

1We2, threatening 26 ltlg4, White also has a clear advantage.

24 a4!! White would like to occupy f5,

but neither 24 .i.f5 ltf8 25 .i.e6+ �h8 nor 24 g3 l:l.f6 25 .i.f5 l:l.af8 is really good for White. However, 24 ltlf5 'lif6 25 ltlg3! was a good alternative to the text-move, with ideas of ltlh5/ltle4 and 'l!lh5.

24 ... :ars The point of White's play is that if

one pair of rooks can be exchanged, then the occupation of f5 will be permanent, e.g. 24 ... bxa4 (24 ... b4 25 g3 l:l.ff8 26 cxb4 is just a clear extra pawn) 25 l:txa4 and now:


1) 25 .. . J:af8 26 J:xf4 J:xf4 (both 26 ... gxf4 27 .!i:lf5 1Wf6 28 _.h5 and 26 . . . exf4 27 .!i:lf5 1rf6 28 J:e6 are also excellent for White) 27 .i.f5 ! and Black has no defence against g3.

2) 25 ... J:xa4 26 .ba4 b5 27 .ic2 .!i:lf6 28 b3 when White has cona-ol of f5 and can continue with c4.

Gata decides simply to abandon a queenside pawn.

25 axbS aS Forced, because Black cannot al

low White to create a passed pawn on a6, for example 25 ... J:xf2 26 bxa6 .i.a8 27 J:fl 1t'f7 28 1t'el forcing exchanges.

26 :n .i.c8 27 g3 (D)

27 J:4f7 After 27 .. . .!i:lf6 White can win

with the complicated 28 gxf4 gxf4+ 29 �h2 fxe3 30 fxe3 ! 16'g5 3 1 :a4! or the simple 28 �h2!, which just leaves the rook trapped.

28 b4

I decided that it was time to stop ... .!i:lc5, but it would have been safer to play 28 .ie4! .!i:lc5 29 .ig2, de fending the slightly weak kingside. After 29 ... e4 30 .!i:lc4! White starts to exploit Black's weak pawns.

28 ••• e4! A good try, activating the knight

and creating some kingside counter play. One of Kamsky's strengths is that once he realizes his position is critical, he doesn't hesitate to make the necessary sacrifices tostirup complications. Here one wasted tempo would be too late: if White could play .i.e4, Black's position would be hopeless.

29 .he4 � (D) After 29 ... 1rxc3 30 bxa5 bxa5 31

l2lf5 ! White wins, e.g. 3I...llxf5 32 i.xf5 J:xf5 33 :C I or 3I ...'ef6 32 1rh5 l:h7 33 f4.

30 .i.g2 axb4 After 30 ... ffi+ 31 �hl g4 32

bxa5 gxh3 33 i.xf3 llxf3 34 axb6

ANAND - KAMSKY, PCA CAND. (3), lAS PAIMAS 1995 161

l:txf2 35 l:txf2 l:txf2 36 .J:la8! Black's counterplaycomes to end.

31 cxb4 lilf3+ 32 LfJ

Necessary; White can't continue with his knight stuck in the middle of his position.

32 ... l:l.xfJ 33 l:ta8 .ixh3

33. .. 'l'b7 34 'ifal !, threatening 35 D.a7, is hopeless for Black.

34 'ifxf3 l:txa8 Despite the inaccuracy commit

ted a move 28, White retains a large advantage.

35 l:tcl .J:lf8 (D) 35 ... '1'b2 is met by 36 J:c2.

36 'l'e2 37 .:.c7 38 .:.b7 39 M

.i.d7 :n 'ifa1+

The ending after 39 ... 'ifa8 40 lhb6 'ifxd5 41 'ifd2 'ifxd2 42 lilxd2 is winning for White.

40 lhb6 'ifd4 41 .J:lb8 'ifxb4

Or 4 I...'ifxd5 42 'ltb2+ 'ife5 (the line 42 ... .J:lf6 43 .J:lf8! 'ife5 44 1lrxe5 dxe5 45 lhf6 �xf6 46 b6 .i.c6 47 lile3 �e7 48 lLlc4 also wins) 43 'ifxe5+ dxe5 44 lLle3 and the b5-pawn advances.

42 lLle3 hS 43 b6

Not 43 'ifxh5? 'ifel + 44 �h2 .J:lxf2+ 45 l£Jg2 l:txg2+! 46 �xg2 'ifd2+ 47 �fl 'ifcl+ 48 �e2 'ifc4+ with perpetual check.

43 ... h4 (D) 43 ... .ib5 loses instantly after 44

l£Jf5+ !.

44 g4 After 39 �h2 'il'f6 Black will de- This pawn is very important since

velop more counterplay than in the it shuts Black's bishop out of the game. game. It may appear weakening, but

39 �g7 Black cannot exploit this because of


White's dangerous passed pawn and the fact that Black's king is also exposed.

44 ••. �bS 45 1\Vdl

From the practical point of view, 45 �f5+ '>1i>h7 46 'iid 1 would be better as it cuts out some tactical ideas by Black, but White has not gone wrong yet.

45 .•• 'iib2 46 fi)fS+ :xrs

46 ... 'it>h7 loses to 47 11Vd4 1Wxd4 48 <'Dxd4 �a6 49 �e6 :b7 50 <'Dxg5+.

47 gxfS �e2 48 'Wa4?!

Unnecessarily complicating matters. After48 WVel ! h3 49 :eS White would win comfortably.

48 :i.f3! (D)

w

A brilliant resource, which forces White to play very accurately. After 48 .. .'i'bl+ 49 �h2 'iVxfS 50 'iVd4+ m 5 1 b7 White wins far more easily.

49 'iVd7+ 49 f6+ is simply met by49 ... 'ilfxf6.

49 ••• Wh6 Incredibly, there is no mate for

White. 50 'ilre6r �hS {D)

51 'iVeS+! An important finesse, which forces

Black to block the g4-square with his king. After 51 'iVel 'llrd4 the p®tion would be a draw.

51 ••• �g4 52 liel

Now everything is under control again.

52 ·- hdS 53 :eS .if3

53 . . . 1ixb6 fails to 54 'ifdl + .if3 55 :e4+ while 53 .. . �xf5 opens the position up and allows White to resume his attack by 54 l::tf8+ � SS 'iVe8+ �g7 56 �h2!, when Black is helpless.

54 f6 There are just too many pawns.

ANAND - KAMSKY, PCA CAND. (3), LAs PAIMAS 1995 163

54 ... <;tJb5 54 ... '1'xf6 allows 55 "ire6+ exchanging queens.

55 n "i!Vd4 (DJ

56 lte4 This interference move finishes

Black's resistance. 56 ••• 1Wf6

56 . .. .ixe4 57 f81i'. 57 b7 .ixe4 58 1i'xe4

58 ll'd I+ .if3 59 b81i' .ixdl 60 f81i' is much more elegant; White loses all his original pieces but he has produced two queens.

1-0 Gata had seen enough. 58 ... 'i'al+

59 <;tJh2 Wf6 60 <;tJh3 would be a possible finish.

After this struggle, which equalized the scores, there was a series of five draws in which I gradually gained the initiative. In game seven I gained a decisive advantage, but although I failed to convert this into a win it was nevertheless an important game because it dented his main opening as Black - the Aohr-Zaitsev line of the Ruy Lopez. The drawing run was finally broken in the following game.

Game 31

V. Anand - G. Kamsky PCA Candidates (9), Las Pa/mas 1995

Ruy Lopez, Flohr-Zaitsev

1 e4 eS 2 lU£3 �c6 3 .ibS a6 4 .ia4 016 s 0-0 .ie7 6 .l:el bS 7 .ib3 d6 8 c3 0-0 9 h3 :es

10 d4 .ib7 11 �bd2 .if8 12 a4 h6 13 .ic2 exd4 14 cxd4 �b4 IS .ibl 'i'd7 (D)

An unexpected and rather unusual move. In game seven he had played the more normal l 5 ... c5.

16 b3! I hadn't prepared anything espe

cially against 15 ..... d7, althought I knew all the theory. Since Kamsky must have carefully prepared this line, I decided to try to fmd a continuation which would take the game out of theory and after a time I noticed the move 16 b3 (16 e5 and 16

.l:a3 had been tried before). While there is some risk in leaving the known paths, I had spent so much time on the Aohr-Zaitsev that I felt

confident in my general understanding of this type of position.

16 b3 is quite a difficult move to make if you are used to the standard theme of swinging the rook over to the kingside via a3,but blocking the third rank doesn't mean that I am abandoning the rooks to its fate. can later have an influence along the a-file, but for this White needs to delay axb5 until it is really effective..

16 ••• g6 If 16 . . . c5, then 17 .ib2 and White

prevents the usual re-deployment of the bishop by ... g6 and ... lg7.

17 .ib2! (D) 1he earlier game Van der Wiel-

Karpov, Amsterdam 1991 continued

ANAND - KAMSKY, PCA CAND. (9), lAS PAIMAS 1995 165

17axb5 1' xb5 18 d5 c6 19 i&.b2 llTh5 20 dxc6 cilxc6 with an unclear position. However, in this example White play axb5 too soon; it is a useful threat and shouldn't be executed too early.

17 ... J.g7 18 'ii'cl!

This is really the new idea: White intend to play .tc3 and 'it'b2, building up pressure on the long diagonal. Van der Wielhad analysed 18 d5 c6, which is satisfactory for Black, but the alternative 18 .i.c3 c5 19 axb5 (not 19 d5? l0xe4 20 .txg7 lbxd2 and Black wins) 19 ... axb5 20 J:txa8 .txa8 21 i.xb4 cxb4 22 .td3 also merited attention.

18 Lc8 White gains the advantage after

either 18 ... d5 19 e5 lbe4 20 lbxe4 dxe4 21 i.xe4 .i.xe4 22 l:xe4 l:ac 8 or 18. . . c5 19 e5 dxe5 20 dxe5.

19 i.c3 cS 20 dS

The structure is now very simila1 to the Benoni Defence. White's dream is to achieve the push e4-e5 under favourable circumstances. Obviously not 20 ixb4?! cxb4 21 'li'b2 J:tc3 with very active play.

20 _ 'flle7 Preparing ... lbd7.

21 � (D) Now 21 "i1Pb2 is not especially

effective because Black can reply 2 1 ...lbh5.

21 ••• tt:lb7?! Black changes his plan and de

cides to retreat the knight to h 7 instead. 2 1 ...lLid7 would have been better, although White has a slight advantage after 22 axb5! (22 .txg7 �xg7 23 tt:le3 h5 followed by ... tt:le5 is safer for Black) 22 ... axb5 23 J:ta7 .txc3 24 11hc3 J:tb8! (24 ... tt:lb6 25 'it'cl �h7 26 e5 J:ta8 27 exd6 'it'xel 28 'it'f4! .i.xd5 29 J:txf7+ �g8 30 ixg6 wins for White) 25 11'd2 �g7. This line again emphasizes the point


that White should delay axb5 until it results in a concrete gain.

22 .bg7 lt>xg7 23 liJe3! (D)

Threatening tllg4. 23 ••• hS

This weakens the kingside but the alternative defences also favour White:

1 ) 23 ... liJg5 and now: 1a) 24 lbxg5 hxg5 (24 ... 'ihg5 25

axb5 axb5 26 f4! is very promising for White) 25 axb5 axb5 26 .:ta5 ! 'Wc7 27 .:ta7 when Black is in trouble:

1al) 27 .. . 'ilfh6 28 lDf5+ gxf5 29 'it'xg5+ � 30 'Wh6+ �e7 (30 ... �g8 3 1 .:te3 f4 32 e5 wins) 3 1 e5 ! with a winning attack.

1a2) 27 ... .:ta8 28 lDf5+! gxf5 (White also wins after 28 ... �6 29 .:txb7 'Wxb7 30 lbxd6) 29 'Wxg5+ �f8 30 'il'h6+ �e7 3 1 e5! �d8 32 exd6! .:txel+ 33 �h2 is an attractive finish.

lb) 24 'il'c3+ (also very strong) 24 .. .'i'f6 (24 ... 'it>h7 25 liJxg5+ 'lxgS 26 liJg4 and the f6-square is a horri ble weakness) 25 'ifxf6+ �6 26 liJxg5 hxg5 27 axb5 axb5 28 :as with a very promising ending for White.

2) 23 ... 1l'f6 24 liJg4! •xal 25 11Vxh6+ �g8 26 e5) (D) and now:

2a) 26 ... 1l'c3 27 .hg6 (not 27 liJg5? 'ffxel + 28 �h2 'llfxe5+ 19 liJxe5 liJxg5 30 liJxg6 f6 and Black defends) 27 .. .fxg6 28 lili6+! 10xf6 29 11Vxg6+ �f8 30 'Wxf6+ �g8 31 .:te4 ! winning for White.

2b) 26 ... dxe5 27 liJg5 llc7 28 lbxh7 also wins.

2c) 26 ... lbxd5 27 .ixg6 'l'xel+ 28 liJxel fxg6 29 'ifxg6+ �h8 30 lDh6 .:tf8 31 iDn+ .:txf7 32 'l'xf7 with a clear advantage foc White .

2d) 26 ... :Xe5 27 l/)gxe5! (not 27 .:txe5? dxe5! 28 liJg5 'llfxbl+ 29 �h2 11Vf5) 27 ... dxe5 28 .ixg6 'l'xel+ 29 liJxel fxg6 30 'ffxg6+ strongly

ANAND - KAMSKY, PCA CAND. (9), LAS PMMAS 1995 167

favours White. The queen and passed pawns are more effective than Black's scattered forces.

24 "i'd2!? White starts to play against the

knight on b4. One idea is to play llld4 at some stage,and if ... exd4 then l'xd4+ and 'ii'xb4.

24 �g8 Simply stepping off the danger

ous long diagonal. After 24 ... •f6 25 :aJ bxa4 26 lha4 or 24 ... bxa4!? 25 lba4 White has a positional advantage, while 24 ... lt:lf6 meets with the tactical refutation 25 axb5 axb5 26 IL\r5t! gxf5 27 'ii'g5+ �h8 28 e5 and now:

1) 28...�h7 29 •xh5 f6 30 .ixf5 1xd5 31 �4! 'fin 32 lllg6+ �g7 33 e6 1xe6 34 .i.xe6 J:l.xe6 35 J:l.xe6 'i'xe6 36 J:l.a7+, winning.

2) 28 ... dxe5 29 J:l.xe5 'fkd6 30 'ih6t i>g8 31 J:l.xf5 lt:le4 32 J:l.g5+ �xg5 33 "i'xd6 with a decisive material advantage.

25 axbS axbS (D)

26 llldl!! I spent a long time on this move,

because if White delays then Black can play ... J:I.a8 and relieve the pressure. My main problem against Kamsky has always been messing up winning positions (as in the Sanghi Nagar match), so I put a special effort into being precise.

The point of the move is both to prevent ... lt:lg5 and to prepare lt:lc3, striking at the weak b5-pawn. It turns out that to save the pawn Black has to retreat his knight from b4, but then his most active piece disappears.

26 ••• lt:la6 Or 26 .. .1:1&8 (26 ... ll:lf6 27 lt:lc3 is

very awkward for Black) 27 J:l.xa8 J:l.xa8 28 lt:lc3 •d7 29 e5 dxe5 30 lt:lxe5 •d6 31 "i'f4 "i'f6 32 'li'xf6 lt:lxf6 33 d6 with a clear endgame advantage for White.

27 lLlc3 b4 (D)

28 lt:lb5

168 VTSHY ANAND: MY BEST GAMES OF CHESS

28 lba4!?, followed by �b6-c4, is also very strong.

28 ... l:Oc7 After 28 ... :b8 29 'l'f4 l:.ed8 30

..id3 White again consolidates his positional advantage.

29 ..id3 �xbS 30 ..ltxbS :eds 31 ..ltc:4!

Most Benoni players would have fainted by this point! White has a dream position: the bishop on c4 supports d5 in preparation for e4-e5, Black's minor pieces are ineffective and his kingside is weak.

31 ... �f6 (D) Trying to bring the knight back

into play.

32 'il'h6! This creates the tactical threat of

e5, followed by d6 and 'il'xg6+, so Black is forced to retreat his queen. The effect is that White activates his queen with gain of tempo.

32 'l'fB

Black has no time for 32 ... :a8 (32...�xe4 33 :a2! wins) owing to 33 :xa8 :xa8 34 e5 ltlh7 (34 ... dxe5 35 d6 'l'f8 36 'l'xg6+ and 34 ... � xd5 35 �g5 are dead lost for Black) 35 e6 driving a wedge into Black's position.

33 'l'gS 1l'g7 33...�7 34 'l'f4 and 33 .. ."1'e7 34

:a7 ! are no better for Black.

8

34 :S1 (D)

34 :c:7 There is no defence: 34 ... :117 (if

34 ... ..ia8, then 35 e5 �h7 36 lg3 puts Black's position under intolerable presswe) 35 e5 lt:\e8 (35 ... /2lxd5 36 ..ltxd5 and 35 ... dxe5 36 <tlxe5 :dc7 37 d6 also win for White) 36 e6 :dc7 37 exf7+ Wlxf7 38 :e6 wins for White.

35 h6! :bB Or 35...:dd7 36 ..ltxb7 .l:txb7 37

:aS+! �h7 38 11'f4 li:lg8 39 e5 dxe5 40 :xe5 !, followed by .l:tee8.

36 e5!

ANAND - KAMSKY, PCA CAND. (9). lAS PAIMAS 1995 169

The thematic breakthrough comes just when Black's pieces are tied up and unable to meet the new threats generated by this move.

36 ••• l0e8

Forced, as 36 ... dxe5 37 d6 :d7 38 li:lxe5 and 36 ... l0xd5 37 exd6 are hopeless.

Y1 lhb7 J:.cxb7 38 .ixb7 J:.xb7 39 'i'd8 'i'f8 40 J:.al!

40 e6 is less accurate as Black can still resist by 40 ... fxe6 41 J:.xe6 (41 dxe6 'fle7) 4I...l0g7 42 'i'xf8+ �xf8 43 :xg6 c4.

40 ••• l0c7 41 'i'd7 'i'b8

White can win more easily after 4!...1le8? 42 'i'c6! 'i'xc6 43 dxc6 Zlb6 44 exd6 or 4l. ..dxe5 42 d6 'iWe8 43 dxc7! 'i'xd7 44 l:a8+ Wg7 45 c81.

42 'i'xd6 c4 The only chance.

43 bxc4 b3 (D)

44 l:bl I thought for some time about this

move in order to find a clear-cut win. In fact 44 lt:lg5 b2 45 J:.b1 l:a7 would also have won, but White needs to find 46 Wh2! (after 46 J:.xb2 'iWxb2 47 'i'd8+ lLle8! 48 'i'xe8+ Wg7 Black is saved because White's king is too exposed) 46 ... l:a1 47 J:.xb2 'i'xb2 48 'i'd8+ �g7 (or 48 ... l0e8 49 'i'xe8+ Wg7 50 'i'xfl+ �h6 5 1 'i'f8+! �xg5 5 2 h4+ �g4 53 'i'f3+ �xh4 54 'ii'f4#) 49 'ii'f6+ Wh6 50 lt:lxfl+ �h7 5 1 'ii'h8#.

44 ••• b2 45 'ireS! (D)

The idea is to bring the queen back to d4. Then Black's pieces will be tied down to defending the b2-pawn, and White can exploit his cenlral pawns.

45 ••• llb3 45 ... lla7 46 'i'd4 l:a2 Ieaves Black

equally pinned down, when White wins by 47 'Llg5 followed by 48 e6.

1 70 VISHY ANAND: MY BEST GAMES OF CHESS

46 'l'd4 Wfb4 (D)

47 lLlgS

White could also win by 47 e6 :c3 48 exf7+ � 49 lt:lg5+ �7 50 d6+ (but not 50l:lel+ 'itd7 51 'lg7+ 'itc8 52 d6! lbc4 ! and fight con . tinues as 53 J:.e8+ loses to 53...<i'b7!) so ... 'itd7 (50 ... �d8 5t lt:ln+�d752 I.Lle5+) 51 _.g7+! 'itxd6 52 lik4+.

47 ... l:lcl 48 ... f4! f6

Or 48 ... :ci + ( 48 . . .'ilff8 49 :Xb2) 49 l:txcl ( 49 'ith2 is also winning) 49 ... bxcl ... + 50 ... xcl with three ex-tra pawn

49 exf6 so f7+

lt:lxdS 1-0

Game 32

V. Anand - G. Kamsky PCA Candidates (1 1), Las Pa/mas 1 995

Sici l ian, Najdorf

1 e4 2 liJf3 3 d4 4 liJxd4 5 liJc3

c5 d6 cxd4 liJf6 a6

I had held the I Oth game with Black fairly easily, so Gata was getting a bit desperate; hence his choice of the N ajdorf.

6 ..te3 e6 7 ..te2 ..te7 8 f4 llic6 9 'ii'd2 llixd4

10 'l'xd4 0-0 11 0-0-0 'ii'aS? (D)

A serious mistake, all the more surprising in that similar positions arise in the Richter-Rauzer, an opening with which Kamsky is very familiar.

l l . . .b5 or I I ...'ii'c7 would have been better.

12 'l'b6! 13 ..txb6

'ii'xb6

13 l2Je8 Black also has a poor position af

ter l 3 ... ..td7 (13 . . .lbd7 drops a pawn to 14 ..tc7 while l3 ... e5 14 f5 does not help Black) 14 e5 dxe5 (14 . . . lbe8 15 ..tc5 ! with a decisive advantage) 15 fxe5 llid5 16 liJxd5 exd5 17 ..tfJ l:tac8 18 ..txd5 ( 1 8 ..teJ ..tf5 19 c3 d4 20 l:txd4 ..tc5 21 ..txb7 l:tc7 is less clear, for example 22 ..txa6 l:ta8) 1 8 ... ..tf5 19 ..tbJ ..tg5+ 20 �bl .

A dream position for White, espe- White seems to keep his pawn and dally in view of the match situation though Black has a temporary initia(1 needed one point from the last two live, White should be able to weather games k> win). White has a clear end- it. game advantage and can press hard for the win without the slightest risk of losing.

14 eS! Immobilizing the knight on e8.

14 dS


After 14 .. 16 1 5 iLf3! ( 15 exf6 .i.x.f6 16 �e4 is also good for White White has strong pressure 14.d.xe5 15 fxe5 f6 16 1f3 is similar.

15 fS! (D) The most forceful method. After

15 1g4 g6\, followed by h4,

Blackhas better defensive chances.

Or: 15 .id7

I) 15 ... exf5 16 lLlxd5 .ig5+ 17 �bl .ie6 18 .if3, followed by h4, and White stands very well because of the e8-knight.

2) 15 ... .ig5+ 16 �bl .if4! 17 .id4! (17 .ic5 .ixe5 18 .ixf8 �xf8 is only slightly better for White) 17 . . . b5 ( 17...�7 1 8 g3 i.g5 1 9 f6 ! is similar) 18 l:l.dfl .ig5 19 f6 with a large advantage.

16 .ig4! (D) White must be accurate. After 16

fxe6 fxe6 17 .ig4 .ic8, followed by ... g6 and ... lLlg7, Black might wriggle out.

B

16 ·- .ic8? After this Black is in dire trouble

The alternatives were: I) 16 ... l:l.c8? 17 ffi! gxf6 (17 ... 1c5

I8 ll'lxd5 gxf6 19 exf6 and 17 ... 1d8 18 ltlxd5 are also winning) 18 li)xd5 .idS 19 .ixd8 exd5 (19 .. Jbd8 20 ltlb6) 20 .ixd7 l:l.xd8 21 e6 winning either a clear pawn or the exchange for a pawn.

2) 16 ... .ib4! 17 fxe6 .be6 18

evxd5! ( 1 8 .if3 .ixc3 19 bxc3 :c

20 .ixd5 .ixd5 21 l:l.xd5 !i:Y:7 is not clear) 1 8 ... i.xg4 19 l:l.d4 winning pawn.

3) 16 ... .ig5+ 17 �bl :C8 (after 17 ... exf5 1 8 .if3 .ic6 19 li.:lxdS the e8-knight is again a serious

cap) 1 8 fxe6 .ixe6 (or 18 .. .fxe6 19 lL!xd5 and wins) 1 9 i.f3 with a clear

advantage to White. 4) I6 ... exf5 is most simply met

by 17 .if3! as in line 3 above. 17 lL!xd5?! is less clear after 17 ... 1gS+ 18 .ie3 f4! ( 1 8 .. .fxg4 19 i.xg5 and I8 ... .ixe3+ I9 lLlxe3 i.e6 20 .ixfS

ANAND - KAMSKY, PCA CAND. (ll), lAS PAIMAS 1995 1 73

should win for White) 19 lbe7+ (19 ixd7 fxe3 20 ll:d3 ll:d8 is unclear) 19.. .�h8 (19 ... he7 20 .:.xd7 fxe3 21 :x e7 strongly favours White) 20 :Xd7 f xe3, although 21 lDd5 retains some advantage for White.

17 llhfi! aS 18 �4! (D)

From here the knight can move to c5 or b6, as appropriate.

18 f6 There is no good move, for exam-

ple 18 ... exf5 19 i.xf5 i.e6 20 lbcS or 18 ... :a6 19 .te2 ll:a8 20 ..te3 and Black has serious problems in either case. The text-move is more com-plex, but White's advantage persists.

19 fxe6 fxeS 20 00

The knight has achieved its task and now heads for .the excellent square d5. Round about here I was suddenly worried that I had allowed the position to become unnecessarily complicated, but this was only

nerves due to the exceptional importance of the game -in fact White has everything under control.

20 -·· .tgS+ If 20 ... 1lxfl 21 llxfl lDf6, White

has a nice win: 22 ll:xf6!! hf6 (or 22 ... gxf6 23 lbxd5 Wf8 24 tt:lxe7 !)

23 tt:lxd5 a4 24 tt:lc7 picking up material.

B

21 �b1 (D)

21 ••• tt:\1'6 Or 21. ..1lxfl 22 ll:xfl tt:lf6 (after

22 ... 1la6 White has the pleasant choice between 23 tt:lxd5 and 23 e7 .tf6 24 .txc8 ll:xb6 25 tt:lxd5) 23 ll:xf6 (23 i.h3 d4 24 tt:lb5 tt:ld5 is less effective) 23 ... .txf6 24 tt:lxd5 h5 25 .th3 e4 26 tt:lc7 llb8 27 i.a7 e3 28 �c 1 and wins.

22 tt:lxdS tt:lxg4 22 ... tt:lxd5 23 llxf8+ �xf8 24

ll:xd5 is a simple win. 23 llxf8+

This wins the exchange and leads to a technically winning ending. 23


e7 would have had a similar result: 23 ... l:te8 (23 ... l:txfl 24 e8"6'+ l:tf8 25 Vb5) 24 {jjc7 {j)xh2 25 {j)xe8 .be? 26 :hl .ig4 (26...{j)g4 27 {j)c?) 27 {j)c? l:l.c8 28 l:l.de I .idS 29 l:l.xe5 .ixc7 30 .ixc7 lhc7 3 l l:l.xh2.

Z3 ... 'it>xf8 Z4 li'Jc7

Not 24 .ic5+ 'it>g8 25 {j)c? b6!. Z4 .•• l:l.a6

On 24 ... l:l.b8, 25 .ic5+ 'it>g8 26 .ia7 traps the rook.

ZS J..cS+! A necessary intermediate check.

zs ... 'it>g8 25 ... .ie7?? allows mate in one.

Z6 {jjxa6 J..xe6 26 ... bxa6 27 e7 �f7 28 �d8 also

leaves White the exchange up. z1 &7 ..trs Z8 h3 {j)f6

Or 28...�e3 29 .ixe3 J..xe3 30 g4 .ig6 3 1 �e6 and White wins another pawn.

Z9 g4! (D)

Of course the position is winning for White as Black doesn't even have a pawn for the exchange, but this forcing sequence of moves gives Black no chance to develop counter play.

Z9 -· J..e4 29 ... .ig6 30 � .ih4 31 .i.b6 is

no better. 30 {j)e6 .i.h4

Or 30 . . . .i.f4 31 .ib6 . 31 gS! lDd5 3Z :r1 h6 33 gxh6 gxh6 34 :t"8+ �h7 (D)

35 J..d6! lllb4 The alternatives are 35 .. . �6 36

lL!c5 .ig6 37 lLixb7 and 3S: . .ig2 36 .ixe5 .ixh3 37 lili4 t0xf4 38 :.x£4 .ig5 39 l:l.n+ �g6 40 l:txb7. lnboth cases White gains another pawn.

36 .ixeS .ixcZ+

White wins easily after36...ltlxc2 37 l:l.f4 .id3 38 l:l.xh4 tt:\d4+ 39�cl lllxe6. simplifYing the position.

ANAND - KAMSKY, PCA CAND. (11), LAS PAIMAS 1995 175

37 �cl .ie4 38 lt1'4 liJd3+

Otherwise Black is mated on h8. � liJxd3 .bd3 40 :h8+

White could have taken the pawn by 40 lf7+ �g6 4l lhb7, but forcing the exchange of bishops leaves White with a trivial technical task.

40 ••• �g6 41 .i.t'4! .i.gS 4Z .i.xgS �xgS 43 �d2 .ibS 44 :as .ia6 45 l:l.c8 �h4 46 l:l.cS a4 47 l:l.aS hS (D) 48 �

Of course it doesn't really matter, but48 �1! was a little more accurate

in that 48 ... .i.d3? 49 lba4+ �xh3 loses � bishop after 5_8_lfa3.

49 :04+ �xh3 50 'iW2 1-0

... and I had won the right to challenge Kasparov.

My success in this match was mainly due to not underestimating Kamsk:y's fighting qualities. Already in Sanghi Nagar I had him beaten, but I just didn't finish the job. This time I was more careful in winning positions and didn't relax until the match was actually over.

In the Riga tournament during May I finished second. My play was very convincing, apart from the loss to Kasparov in the Evans Gambit, and I felt on form This was the first time that I encountered Kasparov after becoming the challenger and although I lost the individual game, I was happy with my performance.Both the tournament and the city were very pleasant, so I had every reason to be in a good mood as I started my preparations for the World

Championship match.

Game 33

V. Anand - J . Timman Tal Memorial, Riga 1 995 Ruy Lopez, Worrall Attack

1 e4 2 lt:\f3 3 ..tbS 4 ..ta4 5 0-0 6 'l'e2

e5 lllc6 a6 ll\f6 ..te7

It's perhaps a bit flippant to say that I played 'l'e2 because I was sick of l:el , but it is useful to vary your openings a bit. Even chess professionals are human beings, and if you have seen a position hundreds of times then it is possible to become stale. A bit of variety helps to keep one's interest alive.

Another point is that I had just played 33 games against top-level grandmasters (I I in Las Palmas and 22 in Monaco) and no one has an inexhaustible fund of new ideas. I still had a few left in this "it'e2 variation, which was another reason to try it in this game.

6 7 ..tb3 8 c3 I) l:d1

10 h3

bS 0-0 d6 ..tg4

h's best to force Black to decide right away whether he is going to take on f3. For this reason I regard

9 ... ..tg4 as being premature. Tiviakov, who is an expert on this line, is of the same opinion.

10 •.• i.h5?! (D) IO . . . ..txf3 I I 'l'xf3 ll\a5 1 2 .ic2

c5 is probably the lesser evil.

w

ll d3 White could already play 11 g4,

but given that Black dido 't take on f.3 last move, he is hardly going to do so now.

11 ltJa5 12 ..tel c5 13 lllbd2 00

Now 14 lt:\fl can be answered by 14 . . . ..txf3 15 "tfxf3 i.gS,so it is time to break the pin.

14 g4! ig6

ANAND - TIMMAN, TAL MEMORIAL, RIGA 1995 1 77

IS IZ:lfl White has opted for a very solid

formation in the centre and will conduct all his play on the kingside.

IS 't:lb6 Black has many possibilities, but

White retains a slight advantage in any case, for example after 15 ... h6 16 �g3 .i.g5 17 't:lxg5 hxg5 1 8 �5. In tbis line the pawn on g5 doesn't really block White's kingside play because he can often continue with h4, meeting ... gxh4 by g5 and regaining the pawn on h4 at his leisure.

Black's best idea may be 15 ... 't:lc6 16 �g3 l:e8 17 lDr5 ILlfB, heading for e6, when the slight weakness of f4 might be relevant later.

16 't:lg3 lik6 (D)

I ll ••• w ll • • • •

• ••• B.i.B • • • • • �. � 8'·�-• m BttJ� � � Di.&W� B R � _: - � : - - - �____J

17 �S! A very awkward knight. If Black

ever plays ... J.xf5, White will take back with the g-pawn, opening the g -tile for his attack.

17 l:e8?!

18 h4! This threatens 19 lt:lxe7+ Wxe7

20 h5, so now Black is forced to take. Black can, it is true, win a pawn, but White's attacking chances more than compensate.

18 .•. .hrS After 18 .. 16 19 .i.b3+! .i.f7 (not

19 ... d5? 20 exd5 't:lxd5 21 c4! winning for White, and 19 ... �h8? 20 h5 gives White a clear advantage) 20 .i.xf7+ �xf7 21 g5 Wbitehasexce1-lent attacking prospects.

19 gxiS (D)

19 dS?! Black could have transposed into

the game by 19 ... .i.xh4! 20 �h1 (20 't:lxh4 Wxh4 21 �g2 We7 22 l:h 1 is also feasible) 20 ... d5. This moveorder would have been more accurate, as it denies White the opportunity mentioned in the next note.

20 �hi?! Not bad in itself, but Black's pre

vious move gave White an additional


option which he didn't exploit: 20 itJg5 ! (not20 h5 h6 2 1 �h2 .i.g5 22 lgl f6 and Black sets up a blockade) 20 ... h6 (otherwise White proceeds with his attack without sacrificing a pawn) 21 'il'b5! hxg5 22 hxg5 .i.d6 (22 .. :iid6 23 �g2 g6 24 "eh4 gxf5 25 :l.hl winning) 23 g6 fxg6 (23...�f8 24 .i.g5 fxg6 25 fxg6 and Black is lost) 24 fxg6 &7 (24 ... 11'f6 25 .i.g5 11'e6 26 exd5! itJxd5 27 i..b3 with a decisive attack) 25 "inl7+ �f8 26 i..h6 itJf5 27 .i.xg7+! followed by mate.

20 ... i.:xh4 21 :l.gl (D)

21 tllxh4 11'xh4+ 22 �g2 d4 is 1ess clear.

21 i.f6 White's attacking chances along

the open kingside files offer very good compensation for the pawn. In addition, White's bishops are very effective; the one on c 1 is ready for a sacrifice on h6, while the c2-bishop

can go to b3 and in the long run with play an important role in exploiting Black's weak light squares.

22 itJbl g6 23 WFf3 dxe4

1f 23 ... d4, White plays simply 24 lDg4.

24 dxe4 25 '1Vh3!

00 -.e7

Mter 25 . . . c4 26 .i.e3, followed by Ldl, the combined pressure oo d7 from d 1 and h3 would be very awk ward.

26 i.e3 .i.g7 27 itJg4 tDr6

27 ... gxf5'!'! loses immediately after 28 lt:\h6+ �h8 29 itJxf5.

28 lthl6+! 1f 28 .llg5'!, then Black escapes

from the pin by 28 . .. itJxg4 29 lhg4 f6.

28 ... 11'x1'6 (D) After 28 ... .i.xf6 29 lg2,followed

by :tag1 and lh2, White has an enormous attack.

ANAND _ T/MMAN, TAL MEMORIAL. RIGA 1995 179

29 l:lg3? The obvious 29 �xc5 ! would

have been the logical culmination of White's play. This not only regains the pawn, but also prevents ... c4, which imprisons the c2-bishop. After 29 ... g5 30 .i.e3 h6 31 l:lg2 'fie? 32 ib3 Black's position would come under steadily increasing pressure.

After the careless text-move Black gains a new lease of !if e.

13 �h2 lbe?!

Now 30 �xc5? is impossible because of 30 ... ll:Jxf5 31 exf5 'ii'c6+.

JO c4 ...

31 l:lhl (D)

33 'iWxh6 l:lad8 Not 33 ... 'ii'h8? 34 'ii'xh8+ �xh8

35 f6 ll:Jc6 36 �g2+ �g8 37 l:lgh3, winning.

34 'iii>g2 'il'g7 35 'iWe3 l:ld6

White still has a clear advantage, but he cannot win by playing solely on the kingside. The next move starts the opening of the queenside with the aim of activating the c2-bishop.

36 b3! .l:lc8 (D)

37 bxc4 bxc4 After this the bishop can become

active at a4, denying critical squares to Black's pieces. The alternative was 37 ... l:lxc4, but then the bishop

31 h6 becomes active on b3 instead and More a less forced to prevent White can gradually step up the

igS. If Black tties to keep the pawn pressure, much as in the game: then 3J ... l:lad8 32 'it>gl h5 33 �g51) 38 �b3 l:lc7 39 Wfg5 Wff6 40

'ffc6 34 'i'h4 f6 (34 ... 'itf8 35 f6 'il'h6 'ii'g7 is not entirely clear. wins) 35 J.e3 g5 36 11fxh5 and White 2) 38 ila7 ! Wff6 (38 ... ll:Jc8 39 is very much better. 'iWb7 l:lc5 40 �b3, etc.) 39 �b3 l:lc8

32 .1xh6 �xh6 40 fxg6 lZ'lxg6 41 .1xf7+ 'ifxf7 42


�h8+ �xh8 43 'ii'xt7 �4+ 44 lti>f3 and White will win.

38 �hh3 White doesn't have a straightfor

ward winning plan. He is just probing, altering his plan of attack according to how Black arranges his defensive pieces. Winning is often as much a matter of putting pressure on the opponent and making life unpleasant for him as it is about specific moves.

38 ••• �f8 39 'il'cl

Now the threat is 'ii'h l , so Black's king has to return.

39 �g8 (D)

40 �f3 Just at this moment White can

switch his rooks to f3 and g3, since 40 .. . gxf5+ loses to 41 �fg3 ! �g6 (or 4l ... �g6 42 exf5 �4+ 43 �h2) 42 exf5. This will eventually cause f7 to come under strong pressure.

40 ... �cd8 41 �hg3

Now White threatens a gradual penetration on the queenside by 1lf a3, .ia4 and 1!ic5. Then c4 will be attacked and White will also be threatening 11fc7. Timrnan decides that there is no real defence against this plan, and so goes for desperate counterplay. The result is only to accelerate his loss.

41 - fti?! 42 fxg6 �d2

42 ... �xg6? loses after 43 lbg6 11fxg6+ 44 �g3.

43 ft1 1!if8 (D) 43 ... �e2 loses to 44 'i!i'b6, but

Black might have lasted longer with 43 ... Wf8. However, after 44 J:l.g4 �xg6 45 1Wb4+ :&16 46 •c5 the win is just a matter of time.

44 'il'h 1! 'fig? 45 lhf6 1-0

In view of 45 ... 11fxf6 (45 ... 1hc2 46 �f7 and 45 ... llf8 46 'l'h 7 + are also cataslrophic) 46 'ifh7+ � 47 g7+.

MATCH WITH KASPAROV 1 995 181

The months leading up to the Kasparov match were a very exciting time, with winning the World Championship a real possibility. I had been competing in the FIDE and PCA cycles for roughly two years and there had been many tense and exciting moments. When I finally reached the Kasparov match there was a feeling of anticlimax, as if I was already spent from the earlier efforts. I had the feeling that, having played so many matches, I was fairly exposed because I had already shown most of my best ideas trying to reach the world championship itself. The champion can be much better prepared, as he only has to play when and where he chooses and can just wait to

see how the cycle develops. Of course, every cha1lenger says much the same thing!

Certainly, I would have preferred seven or eight months to prepare for Kasparov instead of just under six. Moreover, I had agreed to play tournaments in Monaco, Riga and Moscow (these were arranged before I knew I would play Kasparov) which ate into the possible preparation time. However, it was perhaps no bad thing that I played in these events, as six months is a long time to sit ana1ysing without any tournament activity, a1though it did mean that my preparation only really began in May, and the first game eventually started on September I I th.

I had to assemble a team in a hurry, and it was a unique experience suddenly from having at most two seconds to having four. These were Ubilava, who had been with me since the beginning of the cycle; Yusupov, who had already helped me earlier in the Kamsky match; Wolff, who had been my second in the Ivanchuk match (see Game 17), and Speelman, who was completely new. I had worked with Ubilava and Yusupov before, and was very happy with them. I felt that Patrick Wolff would be a help as he is very well organized and has great experience against the Sicilian. Speelman had been Short's second in his match against Kasparov and I felt that he might be able to offer me some insight into world championship chess.

It was amazing how much more you could accomplish with such help, but it was a1so much more confusing comparing the results of one person's analysis with another's. Trying to get everybody to work together in the most efficient manner was a major task in itself. You can't have all five people analysing on one board - it's just too many heads and hands. On the other hand, splitting into groups analysing the same position often leads to the groups heading off in entirely different directions and then it can be hard to decide which line you are actually going to play. Thanks to the information explosion, the amount of material you have to deal with is gigantic. Facing

Jill VISHY ANAND: MY BEST GAMES OF CHESS

Kasparov, analysis of the Sicilian was a priority, but this is one of the largest of all opening complexes. I had decided that my main defence to I e4 would be l ... e5, and this was also an enormous amount of work. Kasparov had shown that he was prepared to play openings other than the Ruy Lopez,so we had to spend some time on the Evans Gambit, Scotch and Irregular Open Games. He also plays I d4 regularly, so one can imagine the amount of work we had to do.

With hindsight, and given that we were new to the job, I think that we did a reasonably good job with our preparation. I am sure that we would do better next time, based on our experience in 1995. Still, this was an area where Kasparov had an advantage due to his vast experience preparing for world champpionship matches. His preparation was able to survive the close scrutiny of a world championship match while mine took some heavy blows.

It was exciting finally to be given the chance to play against Kasparov for the world championship, but I look back on it now with a fair amount of disillusionment. The organization of the match verged on the ludicrous. First the venue for the match was changed from Cologne to New York without consulting or even informing me! I was still preparing foc Cologne and making hotel reservations there when it was known within the PCA that New York would be the venue. I should perhaps explain that under the PCA system you are left largely to make all your own arrangements for the match, so a switch of venue is quite an inconvenience. Then, in late July or August, Bob Rice suddenly called to announce that the prize fund had been reduced from $1.5 million to $ 1 .35 million. And so on. The only response from the PCA to the various problems regarding the organization of the match was "We are doing our best, but ... ". I didn't find this argument particularly convincing. I had to try to shut myself away from all these other problems to concentrate on the chess, but I didn't have total success. At some level, it kept bothering me.

They tended to take the most optimistic interpretation of any good news. On the other hand, bad news would be parcelled out bit by bit in small doses, so that you would not realize the full import straight away. If it hadn't been for Frederic Friedel, who was in charge of player relations foc the PCA, I probably wouldn't have been kept infonned at all. A few days before the match we were suddenly told that we had to write a daily column foc USA Today. Towards the end I completely lost interest in this and produced just the barest minimum - deep notes along the lines of "He played the Sicilian."

The dealings with the PCA leading up to the match were thoroughly de pressing and, by the time I got to New York, I was just sick of the whole

MATCH WITH KAsPAROV 1995 183

thing. This was not a factor in my defeat but my inability to deal with it and

take it in my stride was! Kasparov put up with a lot during his 1984 match against Karpov. I have a much better idea now of how one should just develop a thick hide in these matters. Nevertheless, a positive consequence of the whole affair was that when it was over, I was quite happy to forget about both it and the events leading up to it.

Game 34

V. Anand - G . Kasparov PCA World Championship, New York (9) 1995

Sicil ian, Scheveningen

1 e4 The World Championship had be

gun with eight draws. The absolute record, 17 consecutive draws, still belongs to Karpov vs Kasparov from Moscow 1 984/5. However, in 1 995 the next six games had five decisive results !

1 cS 2 l£Jf3 d6 3 d4 cxd4 4 l£lxd4 l£lf6 5 l£lc3 a6 6 h2

This had become one of the main battlegrounds of the match. Neither player wanted to blink first, so another Classical Scheveningen appeared on the board.

6 e6 7 0-0 JJ..e7 8 a4 l£lc6 9 JJ..e3 0-0

10 f4 'it'c7 1 1 �h1 .l:e8 12 Jif3 (D)

Our discussion of the Scheveningen continues into my fifth White. I had tried 12 'iW d2 in the first game and then 12 ..td3 three times. In games five and seven Garry had

managed to find a solution to 12 ..i.d3, so it was time to switch variations.

12 ..td7 Kasparov had played 12 ... :b8

consistently in the matches against Karpov, but then switched to this move against Van der Wiel in Am· sterdam 1 987. I can't really say that I was surprise by ... ..td7 , because I had studied the Van der Wiel game but it wasn't uppermost in my mind. I was waiting for 12 ... l:b8 and was ready to whip out 1 3 g4, when this move was playe. I had 1D search my memory to remember what we had found in the Van der Wiel game.

13 l£lb3 ltlaS

ANAND - KAsPAROV. PCA WORLD CH., NEW YORK (9) 1995 185

14 lilxaS 'iWxaS 15 'l'd3 :adS 16 ladl!

Almost all the interesting games in this line were played by Van der W1e� against Polugaevsky and Kasparov himself. Van der Wiel's notes were a good starting point for my own analysis. We had prepared both 16 � (Van der Wiel-Polugaevsky, Haninge 1989) and the text-move, which was a recommendation of Van der Wiel (a third move, 16 'iWd2, was played in Van der Wiel-Kasparov). My decision to prefer 16 :Cd l was made at the board.

16 .•. .i.c6 Mter 16 ... e5 White simply plays

17 f5. 17 b4 'fic7 18 b5 .i.d7 (D)

Not 18 ... axb5? 19 axb5 i.d7 20 /Oa4 with advantage to White.

I was surprised that he was prepared to go down this line so blithely

but later it turned out that there was a good reason: an earlier game Cuijpers-De Boer, Dutch Championship 1988 had continued 1 9 1lle2 :cs 20 bxa6 bxa6 21 1fxa6 :as 22 1fd3 :xa4 23 :xa4 .i.xa4 with equality. It was lucky I didn't know about this game, or I might have abandoned the whole line!

19 J:abl! Clearly stronger than 1 9 1lle2.

19 ••• axbS (DJ After 1 9 ... .1:c8 20 e5 dxe5 2 1 fxe5

fixe5 (21...� 22 .i.xd5 exd5 23 �xd5 fixe5 24 .i.f4 is also good for White) 22 ..i.d4 fie? 23 .i.xf6 (after 23 b6 1fb8 24 ..i.xf6 .i.xf6 25 fixd7 :e 7 Black will regain the pi ICe) 23 . . . .i.xf6 24 1fxd7 White will create dangerous passed pawns as Black hasn't exchanged the a-pawns.

20 llJxbS! 20 axb5 .l:[cS offers White less

than the previous note, now that the a-pawns have gone: 21 llJa4 1fxc2


22 li)b6 (after 22 1Vxc2 1hc2 23 li)b6 White's compensation for the pawn is nebulous) 22 ... 'ii'xd3 23 l:l.xd3 l:l.c7 24 e5 dxe5 25 fxe5 li)d5 26 �xd5 exd5 27 li)xd5 �f5! 28 !Dxc7 �xd3 29 !Dxe8 (or 29 l:!.d1 l:l.d8) 29...�xb1 30 li)d6 �xd6 (or 30 ... �d3) 31 exd6 �f5 with a draw.

20 �xbS This surprised me since I was ex

pecting20 ... 1Va5 2 1 li)xd6 (anything else allows Black to play ... �c6 with a fine position) 2 1 ...�xa4 22 �b6 (22 e5 �xd6 23 exd6 lbd5 24 �xd5 l:!.xd6 25 1t'a3 1:1xd5 26 l:!.xd5 exd5 is, if anything, slightly better for Black) 22 ... l:l.xd6 and now:

1) 23 'ii'xd6 bd6 24.i.xa5 .hf4 (24 ... �xc2? loses to 25 e5) 25 l:l.xb7 �xc2 26 l:l.d8 l:l.xd8 27 �xd8 �xe4! (27 ... li)xe4 28 i.c7 gives White more chances) 28 l:l.b4 �xf3 29 l:!.xf4 .i.d5 30 �xf6 gxf6 31 l:l.xf6 is a draw.

2) 23 �xa5 ! l:l.xd3 24 cxd3 �xd1 and in the resulting ending White keeps a slight advantage due to his two bishops and Black's weak bpawn.

21 'ii'xbS 21 l:l.xb5 !? may be even stronger

than the game continuation, e.g. 21. . .lbd7 (2 1...l:l.c8 22 l:!.dbl 'ifxc2 23 'ifxc2 l:!.xc2 24 l:l.xb7 definitely favours White; note that 24 ... d5? loses to 25 e5) 22 l:l.db1 !Dc5 23 'ifc4 with a clear edge for White.

21 ••• l:l.a8 22 c4 eS (D)

w

23 �b6! A fine move, f01eing the queen to

go to c8. 1be exchange 23 fxe5? dxe5 is premature since then Black can meet 24 �b6 by 24 ... ifc6!.

l3 ... 'l'c8 Now 23 ... 'ii'c6 is bad owing to 24

'ifxc6 bxc6 25 c5! , when 25 ... dxc5 loses a piece after 26 fxe5. Therefore the queen has to retreat to a more passive square.

24 rxeS 25 aS 26 b3

dxeS �f8 ife6(D)

ANAND - KAsPAROV, PCA WORW CH., NEW YORK (9) 1995 187

27 :IdS! �xdS?? An inexplicable mistake. Black

should just wait and make a useful move on kingside, eg. 27 ... h7 then if White plays 28 c5, Black can reply 28 .. 1i'c6 or 28 .. . .1:tec8. If White retreats the bishop from b6, then Black can safely take the exchange since White needs the b6-bishop to support the advance of the pawns.White remains with the better position but has to find a way to break through.

28 exdS White is clearly better and, more

over, his position is very easy to play: he just has to push his pawns.

28 'lkg6 (D)

29 cS e4 30 .te2 :.es

After 30 ... .i.e7 31 d6 .tf6 32 d7 1: f8 Black threatens ... .i.e5 with some counterplay, but 33 .tc7! kills any potential black activity.

31 W'd7! .l:tgS (D)

3 1 ...W'g3 32 'i'xb7 ngs 33 l:gl is hopeless for Black.

After the text-move I just had to calculate a little to be sure of victory.

32 l:gl! e3 33 d6 llg3 34 'iixb7

The simplest, although 34 �h2 is also good.

34 'iie6 (D)

35 �h2! At first I intended to continue 35

.l:tfl, but then I saw Kasparov's trap:


35 ... l:tb8! 36 ifxb8 l:txh3+ 37 �gl 'ife5 and, amazingly, Black forces a draw. I then saw that 35 l:tdl l:tb8 36 'ifd5 would lead to a win, but the

text-move is even easier, because 35 ... .1:Z.e8 may be answered by the simple 36 d7.

1-0

This gave me my first win in a world championship match and my first win against Kasparov (in a tournament game) since Reggio Emilia 199112. I was ecstatic. Needless to say, the rest of the week was a cold shower.

I think one of my main problems was that I had no idea how tense you

could become in such a match. Although I had gained a lot of match experience during the cycles, I had no idea what it was like to play a world championship match; it is genuinely different to lesser matches. Looking back at the l i th and 13th games, I didn't need four seconds to tell me what I did wrong in these games - I just blundered.

Kasparov's play was far from exemplary during the match, but he did'nt make any real blunders. This shows that he was able to keep his nerves under much better control, which could, of course, be a function of his much greater world championship experience. If there is a next time, I belie....: I would be much better prepared to cope with the pressure of the match

I had noticed that a number of players had been badly affected by match defeats. Andrei Sokolov's loss to Karpov sent his career into a tailspin.Likewise Hjartarson against Karpov. After my defeat by Kasparov, I gave some thought as to how to get my career back on track. The memory of the chess world can be very short-lived -you can become a nobody within a year if you don't back your reputation up by good results. I was still strongly motivated to stay at the top, but I understood that waiting for the next Candidates wouldn't be enough - I would have to keep making good results in the inter· vening period.

After the match, my first event was the tournament at Wijk aan Zee, which was Heaven compared to the New York match. You could just go to Wijk aan Zee, everybody understands chess and is enthusiastic about chess; you can just play chess and need not be distracted by changes of venues and prizefunds and 10 I other things. I felt happy in this event, despite the bitterly cold weather, and I played reasonably well. One of my targets was to regain my appetite for chess and to this end I decided to vary my openings. In some games I played I d4, and in others I adopted very sharp lines -all to keep my interest and motivation alive. The following game shows one of these opening experiments.

Game 35

V. Anand - B. Gelfand Wijk aan lee 1 996

Sici l ian, Grand Prix Attack

1 e4 cS Gelfand is a very straightforward

player who doesn't vary his openings much - with Black against 1 e4 he likes to play the Najdorf. I had noticed that he makes no effort to avoid the Grand Prix Attack and had shown some vulnerability against this line. For example, in his Candidates match against Short at Brussels 1991, he lost a game against the Grand Prix Attack and for the rest of the match abandoned the Sicilian. Later, however, Gelfand beat Sax when the Hungarian Grandmaster tried to repeat Short's success. In view of these games, I knew that I couldn't really catch Gelfand by surprise with the Grand Prix Attack, but I did have one new idea to try out...

2 ll:lc3 d6 3 f4 g6 4 ll:lf3 .i.g7 5 .i.c4 liJc6 6 d3 e6 7 0-0 liJge7 8 'i'el! h6

Not 8 ... d5? 9 exd5 exd5 10 liJxd5!, while after 8...0-0 9 f5 !? exf5 10 "lh4! White has an automatic kingside attack.

9 .i.b3 a6 (D) After 9 ... liJd4 (9 ... 0-0 10 "iVh4 is

slightly better for White) 10 liJxd4 cxd4 l l liJe2 0-0 1 2 �hl f5 l3 liJg1 �h8 14 lL!f3 .i.d7 15 .i.d2 l:l.c8 1 6 "iVg3 fxe4 1 7 dxe4 d5 18 exd5 exd5 White had the advantage in the game Topalov-Van Wely, Wijk aan Zee 1996.

10 eS!? ... and this is it! After 10 a4 l:l.b8 1 1 'itg3 lL!d4?!

( l l ...b5 !? may be better) 12 lL!xd4 cxd4 l3 lL!e2 b5 14 axb5 axb5 1 5 'itf2! "iVb6 1 6 f5 exf5 17 exf5 gxf5, which was played in Anand-Gelfand, Reggio Emilia 199112, White could have gained the advantage


by continuing 18 tllf4, but clearly Gelfand was not going to repeat the whole variation.

The idea of this type of pawn sacrifice, which occurs relatively often in the Closed Sicilian and Grand Prix Attack, is simply to fight for the dark squares.

10 ••. ltlrs If Black plays IO ... d5, then 1 1

'ilff2 b6 12 �d2 tlla5 (12...0-0?! 13 tlle2 �b7 14 a4 is good for White; a5 is a threat as c5 is weak) 13 tlle2 tllxb3 14 axb3 favours White, who will play b4 or d4 and exploit the absence of Black's dark-squared bishop from the queenside.

If 10 ... dxe5 1 1 fxe5 tllxe5 ( I I . . .g5 12 lDe-4 g4 13 tlld6+ �f8 14 tllg5 ! with a winning attack for White) 12 tllxe5 'i!fd4+ 13 �hl (not 13 �e3 'i!fxe5 14 'l'f2? 'i!rxe3 ! 15 'i!fxe3 �d4) 13 ... 'i!fxe5, then 14 tlle4 (14 "iff2 i s also possible) 14 ... 0-0 (l4 ... f5 15 �f4 1Wxb2 16 tl\d6+ and 14 ... 'i!fc7 15 �f4 e5 16 �xf7+! are also very good for White) 15 �xh6 f5 16 M4 (even stronger than 16 �xg7 �xg7 17 tllg5) 16 ... 'i!rxb2 17 l1bl 'i!fd4 18 tt:\g5 and White wins.

The move Gelfand played is the best.

11 �hl! 11 lDe-4 would be premature, as

Black could take on e5. 11 ••• tl\rd4 (D)

After 1 1 ...0-0 White could play 12 lDe-4, since if Black takes the

pawn on e5 then White has g4 foJ. lowed by �xh6 at the end Also I l ...tllcd4 12 �d2 (since the pressure on e5 has been lifted, White doesn't have to play lbe4 immedi· ately) and I I . .. d5 12 l2le2 (or 12 'i!ff2) would give White a slight advantage.

12 tlle4 I spent a long time thinking about

12 tt:\xd4 cxd4 13 lbe4 dxe5 14 'I' g3 ( l 4 fxe5 is met by 14 ... �xe5 andnot 14 ... tllxe5? 15 'irg3! g5 16 .if4!, winning) with the point that 14...0-0 allows 15 f5 ! exf5 16 �xh6! (16 'ifxg6Wh8 17 .hfl l2le7!) 16 . ..la7! 17 �g5! fxe4 18 'l'h4 with a clear advantage for White, e.g. 18 ... W5 19 .lbf5 'ifh6 20 l:tf6!. However, I

just couldn't find anything against 14 ... 'ire7 ! 15 fxe5 he5 16 .if4 �xf4 17 l:txf4 f5 ! (17 ... 0-0 l8 lU6! is too dangerous) 18 'i!rxg6+'i>d8 19 tllg3 (19 tllc5 l:te8!) 19 .. ."l'g5!. AI· though Black needs to play a whole

ANAND - GElFAND, WIJK AAN ZEE 1996 191

sning of 'only' moves, I didn' t doubt that Boris would find the correct path. In the end I settled for the more modest text-move, but this has the defect that Black can gain time against White's rook.

12 ... <'i)xf3 12 . .. dxe5 13 <'i)xe5 ! is promising

for White. 13 !W3 (D)

Not 13 �xd6+? i¥xd6 and Black wins.

B

13 ••• dxeS 14 fxeS <'i)xeS

After 14 ... i.xe5 White can choose between two favourable lines: 15 lf2 and 15 <'i)xc5 i.xh2 16 i.xe6 !.

IS J:[fi gS! Boris immediately found this

forced move. 15 ... 0-0 loses to 16 ixh6! and now:

I) 16. ..�xd3 17 'i'e3 <'i)xb2 18 ixg7 1Pxg7 19 'i'xc5 f5 20 'i'e5+ �h7 2 1 1i:lg3 'i'f6 22 'i'xf6 l:l.xf6 23

l:abl trapping the knight.

2) 16 ... i.xh6 17 <'i)f6+ �g7 18 11Vxe5 'l'd4 1911Vg3! with a very dangerous attack.

After other 15th moves, White just plays i.f 4 and 'I' g3, with a very strong attack.

16 11Vg3 16 <'i)xc5 0-0 17 'l'e4 'l'e7 regains

the pawn, but White's attack has gone.

16 ••• 0-0 (D) Once again the correct move. In

stead, for example, 16 .. .f5 (16 ... b6 1 7 i.f4! 'i'c7 18 l:l.ael and 16 ... l:l.g8 17 <'i)xc5 also favour White) J7<'i)xc5 (17 i.f4 'ifc7!) 17 ... 'ilfe7 18 d4<'i)c6 19 c3 is good for White.

17 i.xgS!? I didn't think much about this

piece sacrifice, since it seemed to be the natural follow-up to White's earIier plan. I just checked that Black had no obvious defence, and then played it. While this may appear reckless, it fitted in with my ambition to play


interesting chess during the tournament

With best play Black can survive, so it is probable that White should look for an improvement earlier. Still, Black has to walk a tightrope for several moves, no easy task in such a complex position.

17 lL!xc5 is not dangerous for Black and he can equalize comfortably by 1 7 ... b6 1 8 � ..i.b7.

17 ••• hxgS 18 �xgS (D)

Threatening 19 'i!fh4.

18 ••• lDg6 Best. After 1 8 ... 'i'd4 ( l8 ... b5 19

l:tael lDg6 20 lDxf7! is good for White) 1 9 'i'h3! l:le8 20 l:lael l:le7 21 .l:.e4 White's attack is dangerous, for example 2I . ..'i'xb2 (21 .. .1Wd6 22 l:le3, intending l:lg3, also poses problems for Black) 22 d4! c4 23 ..i.a4 (so that Black doesn't get a pawn on b3, though even 23 dxe5 cxb3 24 1Vh7+ �f8 25 �f4! looks

good) 23 ... 'i'xa2 24 dxe5 'l'xa4 2S 'i'h 7+ �f8 26 .l:.ef4 and White wins.

19 hell (D) Not 19 'if g4? 'ifd4 20 'irh5 'lh4

and the attack collapses.

B

I had seen up to here when I sacri· ficed the piece and thought that Black would have a tough job defending the position, but in fact he can hold on.

19 ... 'fle7 A critical moment. lbe alterna

tives are: I ) 1 9 ... c4? 20 hc4b5 2l ib3 is

pointless as 2I. . . ..i.b7? loses to 22 l:lxe6!.

2) 19 . . . .ih6? 20 ttlxf7 .l:.xn 21 'i'xg6+ wins.

3) 19 ... Lb2 (Black can just about hang on after this move) and now:

3a) 20 lDxe6 .i.xe6 21 ixe6 �g7 ! defends.

3b) 20 ..i.xe6!? .ie5 (20 .. .fxe6 21 lL!xe6 :Xfl + 22 l:txfl .ixe6 23 'i'xg6+ and White may have no

ANAND - GELFAND, WIJK AAN ZEE 1996 193

more than perpetual check) 21 l:l.xe5 fxe6 22 l:l.eel �g7 and Black is slight! y better.

3c) 20 ll'lxf7 ! l:l.xf7 2 1 'i'xg6+ :g7 22 .i.xe6+ .i.xe6 23 'i'xe6+ 'i>h8 24l:l.e3 l:l.h7 (24 ... 'i'g8 loses after 25 l:l.h3+ l:h7 26 l:l.xh7+ 'i'xh7 27 l:l.f3) 25 l:l.f7 (D) leading to a final branch:

3cl) 25 ... l:l.h4 26 l:l.h3 ... g5 27 l:l.xh4+ 'i'xh4 28 :C3 wins.

3c2) 25 ... .i.g7 26 l:l.xg7! l:l.xg7 27 l:l.h3+ l:h7 28 'i'e5+ �g8 29 l:l.g3+ �8 30 'i'f4+ l:l.f7 3 1 'i'h6+ �e8 (31 ...'i>e7 32 l:l.e3+ �d7 33 'i'e6+) 32 :r. g8+ �7 33 'i' e3+ �d7 34 l:l.xd8+ with a winning ending.

3c3) 25 ... l:l.h5! 26 l:l.xb7 (26 l:l.h3 'i'g5 27 l:l.xh5+ 'i'xh5 28 l:l.f3 l:l.e8 wins �or Black) 26 ... .i.g7 is unclear. White has three pawns for the piece but Black has enough pieces in play

to defend his king. 4) 19 ... .i.f6!? (D) (this may also

enable Black to draw) and now:

w

4a) 20 ll'lxf7? �xf7 is unsound. 4b) 20 ll'le4? .i.h4 solves all

Black's problems. 4c) 20 l:l.f5?! �g7 repulses White,

but not 20 ... exf5? 2 1 ll'lxf7 l:l.xf7 22 'i'xg6+ �h8 23 i.xf7 Vcs 24 l:l.e8 'i'xe8 25 'i'h6#.

4d) 20 l:xe6!? .he6 (20 ... .i.xg5? 2 1 l:l.xg6+ wins) 21 ll'lxe6 'i'e7! (not 2 l ...fxe6? 22 'i'xg6+, nor 21 ...'i'b8?! 22 fuf8 'i'xg3 23 hxg3 with excellent winning chances for White) and now:

4dl ) 22 �f4?! �h7 (22...�g7? 23 <'Llh5+ wins for White) and Black has some advantage after 23 �d5 .i.h4 or 23 'tFh3+ .i.h4.

4d2) 22 lL!xf8 �xf8 23 'i'f3 �g7 24 .i.d5 lbe5 25 'i'e4 is unclear. White will gain three pawns for the knight, but Black's king is not in danger.

4d3) 22 �7 (with the threat 23 'i'xg6+) 22...�g7 23 �5 'i'e2! (23 ... 'i'e5 24 �xf6 'i'xg3 25 �h5+ wins while 23 .. ..i.h4 24 lL!xe7 .i.xg3 25 �f5+ �h7 26 �xg3 is a very


favourable endgame for White) 24 l:l.e 1 .i.h4 25 l:l.xe2 .i.xg3 26 hxg3 and in view of White's broken kingside pawns, Black has no problems.

4e) 20 li)xe6 fxe6 and now: 4el) 2l l:l.xe6 �g7 ! (certainly not

2l . . . Le6 22 'ilfxg6+ �h8 23 'l'h6+ llig8 24 .i.xe6+ winning for White) 22 l:l.d6 'file? 23 h4 'ife5 and Black wins.

4e2) 21 'fllxg6+ .i.g7 22 .i.xe6+ .i.xe6 23 'fllxe6+ 'i>h8 and White has a perpet11al check although possibly not more.

20 l:l.fS!! (D)

B

Defending the knight on g5 so that White can play 'irh3.

lO ••• .i.£6 The main alternative is 20 ... .i.h6

(20 ... exf57 loses to 21 l:l.xe7 li:Jxe7 22 'ifh4 l:l.d8 23 'ifh5) and now:

1) 21 li)xe67! .i.xe6 22Le6 (22 l:l.xe6 fxe6 23 'ifxg6+ 'fig7 defends) 22 ... fxe6 23 'ifxg6+ 'ifg7 24 'ifxe6+ �h7 is unclear.

2) 21 li)xf77! and now: 2a) 2I...�g7 22 �5! and Black

is helpless: 2al) 22 ... exf5 23 'fixg6+ 11ih8 24

'irxh6+ mates. 2a2) 22 ... l:l.xf57! 23 'l'xg6+ wins

after 23 ... �h8 24 'fixf5 or 23 ... � 24 11'xf5+! exf5 25 li)g6+.

2a3) 22 ... l:l.f6 23 J:fl! (threaten ing 24 l:l.efl ) wins.

2b) 2I...l:l.xf7 22 '1i'xg6+ �g7 (or 22 .. .l:l.g7 23 'ilfxh6) 23 lhe6 .he6 24 .i.xe6 .l:l.af8 25 l:l.g5! (threatening mate in one) 25 ... �h8 (25 ... '1'f6 26 'fllxf6- isn't that a nice pair of pin!) 26 l:l.h5+ .i.h6 27 lhh6+ llh7 28 'ittg1 'llg7 29 l:l.xh7+ 'l'xh7 30 l'g5 should be winning for White.

2c) 2 L �h7! (Gelfand's suggestion) is unclear after 22 �xh6 �xh6 or 22 li)g5+ Lg5 23 lhg5 'ifn 2A h4 b6.

3) 21 h4 (D) and Black seems have no adequate defence against the threat of h5:

ANAND - GElFAND, WIJK AAN ZEE 1996 195

3a) 2l .. .'t>g7 22h5 i.xg5 23 �g5 liltS (23 ... 1!6'f6 24 hxg6 .l:th8+ 25 �gl wins a pawn with a good posilion)24 l:txg6+ 'MS (24 ... fxg6? 25

'ifxg6+ wins) 25 .l:tg5 and again White has a clear pawn more.

3b) 21. ..-t>hS 22 h5 exf5 (the line 22 .... bg5 23 .:lxg5 is very good for White) 23 l:txe7 fi)xe7 24 fi)xf7+ lbf7 25 i.xf7 i.g7 26 'ifg5 and White wins.

3c) 2L.i.xg5 22 .l:txg5 ot>g7 23 h5 is line 3a.

3d) 2I ... c4!?22 i.xc4 b5 23 i.b3 (D) (23 i.d5 l:ta7 24 h5 1!6'c7! is less clear-cut) and now:

3d3) 23 ... 11Vb4 24 fi)xf7 ! i.g7 25 .l:te4 'ife7 26 fi)d6 l:txf5 27 fi)xf5 and again White wins.

21 ltlxe6 (D)

21 f:xe6?? Up to here Black has found the

correct defence time after time, but now he goes down without a fight. 2l.. . .l:te8! was the right defence:

1) 22 l:r.efl was my original intention. During the post-mortem we agreed that after 22 ... i.xe6 23 i.xe6 'ifxe6 24 l:r.xf6 White has enough compensation. Later Gelfand sent me an e-mail pointing out that 22 ... i.h4! is very good for Black.

3dl) 23 .. .'it'h8 24h5 exf5 25 l:txe7 2) 22 'iff3 i.h4 (22. ..fi)h4 23 fi)xe7 26 fi)xf7+ :Xf7 27 i.xf7 and 'i'g4+ fi)g6 24 'iff3 is a likely draw, White wins. but not 22 ... i.e5? 23 .l:txf7 ! 'ifh4 24

3d2) 23 ... i.b7 24 h5 'i'b4 (White l:txe5 i.xe6 25 .l:txe6 .l:txe6 26 'iffl also wins after 24 ... exf5 25 .l:ttie8 27 l:r.f6 and White wins) 23

fi)xe726 lbxf7+ 'it>h7 27 fi)xh6) 25 .l:txf7 'ifxf7 24 'ifxf7+ �xf7 25 hxg6 exf5 26 i.xf7+ l:txf7 27 gxf7+ ltlc7+ i.e6 26 .l:txe6 l:lxe6 27 i.xe6+ �f8 (or 27...Wg7 28 l:te8) 28 .l:te6 is �e7 28 fi)xa8 �xe6 29 fi)b6 and the winning for White. ending slightly favours Black.


3) 22 l:ffl .be6 (22 ... fxe6 23 'l'xg6+ Ji..g7 24 l:tf6! 'l'xf6 25 'l'xe8+ 'ft'f8 26 ii..xe6+ ii..xe6 27 'li'xe6+ 'l'f7 28 'l'b6 is unclear) 23 ii..xe6 fxe6 24 'l'xg6+ i.g7 25 l:tf3 l:tf8 with advantage for Black.

4) 22 l:te4! i.xe6 23 i.xe6 fxe6 24 'l'xg6+ Ji..g7 and after 25 li.f3 or 25 l:th5 'l'f6 26 'l'xf6 hf6 27 l:txc5 White still should be able to make a draw owing to the reduced material. However, whatever winning chances there are lie with Black.

22 lbe6! (D) A deadly blow. Not 22 i.xe6+?

ii..xe6 23 l:txe6 'l'g7! and Black is better, nor 22 'ilhg6+? 'i'g7 (however, 22 . . . Ji..g7? 23 l:txe6 wins for White) 23 'i'h5 leading only to a draw.

22 ... �g7 Or 22 ... he6 23 'ffxg6+ 'i'g7 (if

23 .. . Ji..g7, then 24 ii..xe6+) 24 he6+ l:tf7 (24 ... �h8 25 l:th5+ mates) 25

Ji..xf7+ �f8 26 'fl'xf6 and White wins.

23 lbe7+ i.xe7 24 lbf8 Ji..xfB

White's large excess of pawns would be enough to win in any case, but in fact Black doesn't even get a chance to bring his queenside pieces into play.

25 h4! 1-0 In view of25...�h726h5 �7 27

'l'f3 i.f5 28 'i'xb7.

The Wijk aan Zee tournament went well in general, although the lililed brilliancy against Sokolov was an unnecessary loss. At fust I thought that my loss to Topalov was also unnecessary, but it turned out later that he had seen much more than I had. In the last round I beat Tiviakov in a game lasting 107 moves, gaining revenge for a loss I had suffered against him in 1989. 1n fact i have only lost once to him at a normal time-limit, but he has written about it so often that most people think I have a huge negative score against him! My score of 8/13 was sufficient for second place behind Ivanchuk.

In April I participated in the annual Amber tournament in Monaco, which consists of a mixture of blindfold and rapid games. At the start I played a number of difficult opponents, but in these early rounds managed l 'h-'h against Karpov and Lautier and 2-0 against Nikolic. However, Kramnik had

5TH AMBER RAPID, MONTE CARLO 1996 197

raced into the lead and I was never able to catch up with him. Half-way through I had a bad patch with six draws and two losses from eight games (I was even a bit lucky to achieve this meagre total). Finally I came to the end of the bad patch beat Xie J un 2-0 and then met Judit in the following round. In the blindfold game she had the advantage but then played some strange move and after hard a fight I won. Then we met in the rapid game.

Game 36

V. Anand - J. Polgar 5th Amber Rapid, Monte Carlo 1 996

Pirc Defence

1 e4 My win in the earlier blindfold

game had been tiring - after getting a position where she could barely move a piece, I allowed myself to be swindled and had to win the game several times. Surely the rapid game (with my eyes open! ) would be more relaxing?

1 g6 I didn't expect . . . g6 and was quite

surprised that she played it. 2 d4 �g7 3 lbc3 d6 4 �e3 c6 S .. d2 bS 6 f4

I couldn't really remember what to play here, but in a rapid game you shouldn't worry too much about minor details. By now the exertions of the first game had faded away, at the prospect of a good hackfest! Somehow, I couldn't bring myself to be solid, especially after three wins in a row.

6 ... ltlr6 7 �d3

Not 7 e5? b4!. 7 eS 8 lbr3 exd4

9 �xd4 0.0 10 0-0

At first I was quite happy with my position as all White's pieces are developed to reasonable squares. However, after her next two moves !realized that my opening had not been very successful.

10 11 lbe2 12 �h1 13 �g1 14 ibg3 15 b3 16 l:l.ae1

b4 lbbd7 cS i.b7 .. c7 l:l.ae8 l:l.e7 (D)

Now 17 l:l.e2 l:l.fe8 18 l:l.fel leads to a horribly passive position, and 1 8 . . . lbg4 intending 19 ... i.c3 would

ANAND - J. POLGAR. MONTE CARW RAPID 1996 199

be very unpleasant. Therefore, I decihl kl throw caution to the winds and swing my queen over.

11 rs :rea 18 'l'r4 &s (DJ

18...h6!? is also possible: after 1 9 fxg6 fxg6 20 l:e2! (direct attempts like 20 i.c4+ don't work) the position is quite messy.

Understandably, Judit tries something else.

19 'i!t'h4 'ifd8 20 i.e3 �xf3 21 gd3 'i!Vb6 22 i.f4

To meet . . . d5 with e5. Nevertheless,Biack must play ... d5 as otherwise White would have time for

l:te2-g2. 22 ... d5 23 eS c4 24 fxg6 (D)

Not 24 bxc4? dxc4 25 fxg6 hxg6! ( byincluding the exchange on c4, White has forced Black to recapture

with ... hxg6, since 25 . . . fxg6? loses to 26 i.e3 !, but he has outfoxed himself, since that mighty b7-bishop is now staring at f3!) 26 �f5 gxf5 27 i.xf5 �d7! (now this works since Black can meet l:gl with ... i.xf3+) and I suspect that Black is already winning.

24 exf6? is also bad as Black can reply 24 ... l:!xel 25 fxg6 (25 fxg7 cxd3) 25 .. .fxg6!.

24 hxg6 (D) 24 .. .fxg6! is better: 1 ) 25 bxc4 �7 is clearly good

for Black. 2) 25 exf6 i.xf6 26 J:be7 i.xe7 !

(26 . . . l:be7? 27 i.g5) is slightly better for Black.

3) 25 �f5 !. We thought in the post-mortem that Black was better after 24 ... fxg6, but later I found this move, which, while not as promising as in the game, still offers White good chances. Besides, White doesn't have much of a choice at this point. I am


not sure how to assess the position aside from the usual cop-out of unclear! One possible continuation is 25 ... gxf5 26 ..i.xf5! h6 (26 ... ..i.c8 27 exf6!) 27 llgl with messy complications.

25 lOrS!

Now this is extremely strong and I think White is already winning.

25 ••• gxfS Or 25 ... cxd3 26 lllxg7 ! and wins.

26 ..i.xiS d4 26 ... llld7 27 'iWh7+ �f8 28 :Z.gl is

decisive. 27 :Z.e2 d3

27 ... llld5 28 11'h7+ wf8 29 :Z.g2 f6 30 .i.h6 and White mates in a few moves.

28 :Z.g2 :Z.xeS 29 :Z.xg7+

Better than the unclear continuation 29 'iWh6 lllg4 30 ..,h7+ �f8 31 :Xg4 "ilff6 32 .i.xe5 (32 :Xg7 "ilfxg7! 33 ..i.xe5 :Z.xe5 is good for Black) 32 .. .llxe5.

29 ••• Wfll Or 29 ... Wxg7 30 .lth6+! �h8 31

..i.e3+ Wg8 32 'iWg5+ � 33 .ixb6 picking up the queen.

30 'iWh6 l:Z.e2 (D)

w

31 bxc4!! I am very proud of this move.

which was the main reason why I selected this game. Basically, I quickly rejected 31 :Z.g6+ followed by 32 llxf6 because of 32 ... hf3+. Judit had gone much further in this line, but I couldn't be bothered - I W3lted something cleaner. Black's pieces are perfectly placed; indeed the only piece which can be better placed is the e8-rook (aside from the black king, of course!). I noticed l:Z.g2t followed by :Xe2 and I also saw bxc4 in connection with the move cS dis· turbing the black queen. Suddenly I realized that the d3-pawn could not move ! Bingo!

After 31 l:.g6+ We7 32 :xf6 Black may play:

ANAND - J. POLGAR, MONTE CARLO RAPID 1996 201

I) 32 ... .i.xf3+ (I had only seen this)33 l:xf3 l:el + 34 �g2 l:tgl+ 35 �h3 'i'xf6. I stopped here - I was convinced I was winning and didn't want to waste my time making this work.

2) 32 ... l:tg8! (full marks to Judit for noticing this move!) and now I am indebted to Fritz4 for the reply 33 .id6+ ! 'i'xd6 34 .i.xd3! (the only way - White needed to jettison his f4-bishop in order to cover f3 and to lure the enemy queen to d6, from where it no longer eyes the glsquare) 34 ... 'i'xf6 35 'i'xf6+ �6 36 .be2 c3! and Black is certainly not worse here.

White can also try 3 1 cxd3, with the same idea as in the game, but then Black can limp on with 3 1 .. .�e7.

31 ••• d2? After 3l ...'�e7 32 .i.xd3! l:tf2

White has: l ) 33 l:txf2 'i'xf2 34 l:txf7+ �xf7

35 'i'g6+ (35 .i.g6+ �g8!) 35 ... �e7 (35 ... �e6 36 .i.f5+ �e7 37 1Wg7+ �d8 38 .i.c7#) 36 'i'g7+ �e6.

2) 33 l:txf7+! �xf7 34 'ilfg6+ �e7 (34 ... �e6 35 :Xf2 is winning for White) 35 1Wg7+ and now either 35 ... �d8 36 l:txf2 l:te1 + (36 ... l:tg8 37 .i.c7+! 'i'xc7 38 'i'xf6+ is decisive) 37 �g2, or 35 ... �e6 36 l:te1+ lL!e4 (36 ... .i.e4 37 .i.xe4 is winning) 37 .i.xe4, winning for White in both cases.

32 l:tgl+ She had missed this one. Now I

win the house. 1-0

Topalov was easily the most successful tournament player in 1996, finishing either first or joint first in Amsterdarn, Novgorod, Leon, Madrid, Vienna and Dos Herm as an .I had already lost two games against him in 1996, so I was looking forward to stopping the sequence.

Game 37

V. Anand - V. Topalov Dortmund 1 996

Sici l ian, Scheveningen

1 e4 2 m 3 lbc3

c5 e6 a6

A slightly unusual move-order, but both of us seemed to have decided that the game was going to be a .i.e2 Scheveningen regardless of the move-order! The comment in Game 25 about this being an inflexible move-order for White only applies if White is going to play the aggressive f4 and 'iff3 system. You can play the solid .i.e2 line against almost anything.

4 d4 S �xd4 6 .i.e2 7 0-0 8 .i.e3 9 f4

10 a4 11 �h1

cxd4 �c6 d6 �6 .i.e7 0-0 'ifc7 l:te8

A position on which I had done a lot of work for the Kasparov match. Although I wasn't so successful in the match itself, the analysis paid off over the succeeding months. At the time this game was played, I was still far ahead of other grandmasters in my understanding of this line. Later on they caught up with me, but

not before I had notched up several wins.

12 .i.f3 llJaS (D)

13 g4! Kasparov played 1 3 .igl against

Topalov a month earlier at Dos Hermanas, but I decided to be less subtle. That game conlinued 13 ... 1f8 14 'ife 1 l:tb8 1 5 h3 �7 16 .t!alilc6 17 l:td l �xd4 18 :Xd4 b5 19 axbS axb5 20 e5 with an edge for White.

13 ••• lLld7 13 ... �4 14 .i.cl e5 l5 l2f5 exf4

16 g5 is good for White. 14 .i.g2

This move reflects one a my discoveries: that when Black plays . . . �7 voluntarily, it may not be

ANAND - TOPALOV, Do/U'MUND 1996 203

necessary to play g5. Of course, you may want to play g5 in the end for

attacking purposes, but White can time it much better. One point behind leaving thepawn on g4 is that an early g5 can be met by ... g6 followed by ... e5. If the pawn is still on g4 it is much easier for White to meet this manoeuvre by the piece sacrifice lili5, opening the g-file after ... gxf5 gxf5.

14 ... .i.f8 Black has a range of possible

plans ; foc example he could continue 14 ... b6, but again White plays 'i'el, :dl and any other useful moves he can find before pushing the g-pawn.

1S .. el b6 After 15 . ./oc4, White plays 16

.icl intending b3 and .i.b2. 16 l:tdl .ib7 (D)

17 'i'h4 White's plan is to move the e3-

bishop out of the way (possibly to gl) and then to play J:l.d3-h3. After

Black defends the h7-square by ... g6, .. .i.g7 and ... ltlf8 White again makes use of the position of the pawn on g4 by playing f5, since then ... exf5 can be met by gxf5.

17 ••• li:lc6 Mter l7 ... li:lc4 18.i.cl g6 19 lld3

White proceeds with his kingside attack.

18 ltlde2! White must make time for this

move as 18 llf3 lets Black free his position by 18 ... li:lxd4 19 .i.xd4 e5 20 llh3 h6.

18 lC!b4 19 Jtdl (D)

19 1f l9 ... d5, then 20 e5 f6 2 1 exf6

lC!xf6 22 .i.d4 gives White a positional advantage. The text-move aims to provoke g5. This looks odd unless you have read the earlier discussion about the merits of withholding g5!

20 gS f6 (D)


After 20 ... g6 21 l:l.f3 llc8 22 l:h3 h5 23 ll'lg3 a deadly sacrifice on h5 is looming, while 20 ... h6!? 21 'i1Vf2 (21 ll'ld4!? hxg5 22 fxg5 ll'le5 23 l:df2 is also possible) 2l . ..hxg5 22 fxg5 ll'le5 23 g6! fxg6 24 .ixb6 favours White.

If Black continues quietly by 20 ... :c8, White ptays 21 .igl followed by l:f3-h3.

21 ll'ld4! Returning to exert pressure on the

new weakness at e6. 21 --- fxgS

After 2 1...ll'lc6 22 gxf6! Black runs into problems:

I ) 22 ... 1Vxf6 23 1Vxf6 ll'lxf6 24 e5 ll'lg4 25 .igl liJa5 26 ll'lb3 .ixg2+ 27 l:xg2 ll'lh6 28 ll'le4 with a clear endgame advantage.

2) 22 ... ll'lxf6 23 e5 (23 ll'lxc6 .ixc6 24e5 .ixg2+ 25 llxg2 ll'ld7 is only equal) 23 ... ll'ld5 24 1Vxd8 ll'lxd8 25 ll'lxd5 .ixd5 (25 . . . exd5 26 e6! ll'lxe6 27 ll'lxe6 l:xe6 28 .ixb6 is

very pleasant for White) 26 i.xd5 exd5 and now both 27 lbf3 !? and 27 e6 ll'lxe6 28 ll'lxe6 lhe6 29 i.xb6 give White some advantage.

22 fxgS li)cfi 23 l:dn! :cs

After 23 ... ll'lxd4 24 .ixd4 lbes Black has finally occupied the e5-square, but it's a bit late for this to matter: White continues 2S .ixeS dxe5 26 l:[f7 with an excellent posi-tion.

24 ll'lce2! In order to make sure that White

can maintain a knight on d4. 24 ••• M

After 24 ... ll'lde5? White can choose between the quiet 2S b3 illll the sharp 25 ll'lxe6 l:xe6 26 l:xf8+ 1Vxf8 27 l:xf8+ l:hf8 28 lbf4 :eeB 29 ll'ld5, with some advantage for White in either case.

25 ll'lxc6 .ixc6 26 lM4 (D)

26 .id7

ANAND - TOPALOV, DORTMUND 1996 205

I dat't think either of us looked at 26 ... .ixa4 for more than one second In such a position you just know that Black cannot afford to go

35 b4! 'Wa3 (35 .. .'l'xc2 36 lt:lxf8 wins) 36 .ixh6! and the long-awaitlld sacrifice on h6 finishes Black.

29 lt:lxeS .ixg2+ pawn-grabbing. White would con- 30 :Z.xg2 (D) tinue 27 :0 (threatening 28 l:l.h3; 27 b4 e5 is less clear) and now:

I) 27 ... e5 28 ::th3 h6 29 'Wh5! and White's attack is very strong.

2) 27 ... :c7 28 b4! and now: 2a) 28...ltlb7 29 !Dxe6! :Z.xe6 30

rfJ and White is winning. 2b) 28 ... !Dd7 29 .ih3 !De5 30

g6!! ltlxg6 (30 ... "ii'xh4 3 1 .ixe6+) 31 'lxd8 l:.xd8 32 !Dxe6 and wins.

2c) 28 ... e5 29 bxc5 exd4 30 cxb6 is the critical line. White has the advantage but the game is far from over.

27 eS! The point of Black's previous

move is to set up latent threats along the c8-h3 diagonal; for example 27 l:lf3 e5 28 ltlr5 g6 29 ::th3 h5! refutes the attack The text-move effectively counters this plan.

27 dxeS After 27 ... d5 28 ::tf3, followed by

:Z.h3, White has an immense attack. 28 lill'3 .ic6

After 28 .. .'i'c7 (28 ... .id6 29 l:l.d2 'Wc7 30 ::tfdl and wins) White continues his attack by 29 g6 h6 30 �g5. It looks slow, but White will gradually use his greater firepower on the kingside: 30 ... .ie7 (30 ... .ic6 31 �f7) 31 l:l.f7! .ic6 32 .ixc6 '6'xc6+ 33 �gl .irs 34 ltlh7 'Wxa4

B

The threat is 3I lilg4 followed by 32 ltlf6+.

30 :Z.c7 White's preponderance on the

kingside is simply too great: l ) 30 . . . .ie7 3 l ltlg4! and now: la) 3 l ...�h8 32 ltlf6 .ixf6 (if

32 ... gxf6, then 33 g6) 33 gxf6 gxf6 (33 ... g6 34 .ig5) 34 :Z.xf6 and the white attack is decisive.

lb) 3 l .. .'Wd5 32 lbf6+ .ixf6 33 gxf6 :Z.c7 34 a5! and White is winning, for example 34 ... bxa5 35 f7+ l:l.xf7 36 :Z.xf7 �xf7 37 'l'f2+ picks up a piece.

2) 30 ... 'Wd5 31 g6 h6 32 .ixh6! gxh6 (32 ... 'Wxe5 33 .icl mates) 33 lbg4 lbd7 34 g7 'Wxg2+ 35 �xg2 :Z.xc2+ 36 �hl .ixg7 37 lbxh6+ wins.


31 ll:lg4 �h8 White also wins after 3 l . . .e5 32

ltJ£6+ gxf6 33 gxf6+ �h8 34 'l'g3 i.d6 35 .ih6 or 31 . ..'i'd5 32 li)f6+ gxf6 33 gxf6+ .ig7 34 i.h6.

32 'i'h3! Threatening 33 g6 followed by a

sacrifice on h6. 32 •.. 'i'dS

The continuation 32 . . . g6 33 ..tf4 i.d6 34 'i'c3+ J:lg7 35 l:tdl ll)e4 36 'Wc6 is decisive.

33 g6 h6 (D)

34 i.xh6! White has more than one good

continuation, but this is the most convincing. After 34 ll:lxh6 gxh6 35 i.xh6! (not 35 g7+? i.xg7 36 .ixh6 �g8, but 35 Lf8+ J:lxf8 36 g7+ �g8 37 gxf8W'++ �xf8 38 i.xh6+

�e8 39 W'g4 is also very good fiX" White) 35 ... i.xh6 36 11'xh6+ 'if.>g8 37 g7 'l'xg2+ 38 �xg2 lhg7+ 39 �h l ll:lxa4 White should win, but the text-move is instantly deadly. 34 ... gxh6

35 g7+! 35 ll:lxh6 (35 D.xf8+? lhfB 36 g7+

�xg737 'ifxh6+ �fl) 35 ... .ixh636 W'xh6+ �g8 transposes to the last note. The order of moves in the game rules out the possibility of ... 'ill'xg2+.

35 ... hg7 Or 35 .. .l:l.xg7 36 l:lxf8+ :Xf8 37

W'xh6+ �g8 38 ll:lf6+ Wfl 39 'flxg7#.

36 li)xb6 Threatening 37 ll:lf7++ Wg8 38

W'h8#. 36 ... W'xgl+ 37 'ifxg2 .ixh6

Or 37 .. .'�h7 38 'ifh3 .ixh6 39 l:tf6 �g8 40 J:lxh6 with further rna· terial gains to follow.

38 'ifg6 1-0 38 ... J:lg8 39 'i'xh6+ l:th7 40 'l'r6+

l:thg7 41 W'h4+ l:th7 42 'fld4+ :hg7 43 l:tn wins at least another piece.

This was almost a model game, but it is not easy to appreciate unless you are familiar with all the intricacies of the Scheveningen.

Towards the end of the year I competed in a very strong double-round event in Las Palmas. The other players were Kasparov, Karpov, Topalov, Kramnik and Ivanchuk. I started with two draws, but in the third round my tournament came alive with the following game.

Game 38

V. Anand - V. lvanchuk Las Palmas 1 996

Ruy Lopez

This game was played on my birthday. The tournament had begun slowly, with only one decisive result in the first two rounds and there had been a lot of whining amongst the public and press about it. This seems a bit unjustified, given the fighting spirit that prevails these days, but perhaps chess fans are right to worry lhat chess will revert back to the 17-move draws prevalent in the 1980s.

1 e4 e5 If anyone doesn't know already,

Ivanchuk plays everything. In fact I expected this, but of course couldn't be sure!

2 lDf3 lDc6 3 .i.bS a6 4 h4 lbr6 5 0.0 .i.cS

Ivanchuk had already played this move twice during I 996. 5 ... b5 6 ib3 .i.c5 was all the rage in I 995, but by this time it had been superseded by the text-move.

6 lDxeS 6 c3 b5 7 d4 bxa4 8 dxc5 lbxe4 9

�xe5 1llxe5 10 'l'd5 .i.b7 was played in Short-Ivanchuk, Novgorod 1996, which ended in a draw. Somehow I didn't feel like repeating this line.

6 llJxeS 7 d4 lbxe4 8 l:tel .i.e7 9 l:txe4 lbg6

10 c4 0-0 11 � d6

If 1 1 ...c6, intending 12 ... b5, then 12 d5 is slightly better for White.

12 llJds .i.b4 For some reason, this had escaped

my attention; the point is that 1 3 g3 is met by 13 ... c6 (this actually happened in Smirin-lzeta, Las Palmas 1993). After the text-move I thought for a long time, because if Black drives the knight back with ... c6 then he has solved all his opening problems. Then I saw the exchange sacrifice, but I consumed a lot of time before taking the plunge. I didn't want to concede equality too easily on my birthday, but the possibility of just being an exchange down was sobering. Finally I decided that Black wouldn't be able to break White's bind.

13 'iniS! An innovation.

13 .•• c6 (D) 1 3 . . . .i.e6!? is an alternative since

14 lbf4? .i.xc4 15 lbxg6 fails to


15 ... �xf2+!. Therefore White is better advised to play 14 i.c2.

14 .l:.xb4! Consistent. Otherwise Black is

just better. 14 ... 'il'xb4

Not 14 ... lDxh4? 15 �g5 with the lines:

1) 1 5 ... 'il'a5 1 6 lDe7+ �h8 17 �c2 h6 (17 ... lDg6 18 �xg6 fxg6 19 lDxg6+ and 17 . . . lDf5 18 lDxf5 �xf5 19 �f6! are also hopeless for Black) 18 'il'xh6+! gxh6 19 �f6#.

2) 1 5 .. .f6 16 �xh4 ! cxd5 17 'il'xd5+ �h8 18 �g3 with excellent compensation for the exchange.

15 'iixh4 lDxh4 16 lDb6 ltb8

Trying to keep the material. Black can bail out by playing 16 ... �f5, but the two bishops give White a slight edge.

17 �f4 lDCS Not 17...ltd8? 18 �g5.

18 dS

Establishing the bind. 18 :e1? fud4 19 �xd6 �e6 20 i.xb8 lhb8 is even slightly better for Black.

18 ••• W (D) After 18 ... cxd5 1 9 cxd5 lld8 20

ltel linB 21 h3 lDe7 22 g4! Black still has to find a way to untangle.

19 � White would like to play 19 g4,

but this is impossible because of the reply 19 ... lte4. The text-move, however, threatens 20 g4 because White can meet I 9 ... lte4 by 20 ltel !. Other methods of preparing g4 are less effective:

1 ) 19 h3 h5! ( 19 .. . :e5 20 :dl and 1 9 ... lte2 20 g4 :Xb2 21 .ib3 lDd4 22 �xd6 lDxb3 23 axb3 l:txb3 24 c5! favour White) 20 �fl :e4! and White no longer has the move :tel .

2) 19 f3? (this stops 19 ... :e4 and threatens g4, but there is another problem) 19 ... lte2! 20 g4 ltJd4 21 �I ltxb2 22 �xd6 lDxf3 23 .ixb8

ANAND - [VANCHUK, LAS PACMAS 1996 209

ixg4 and there is no defence to mate!

It is hardly necessary to mention that the greedy 19 dxc6 bxc6 20 ixc6 releases the bind and gives Black the advantage after 20 ... D.e2 ! .

19 ... h6? Up to here Black has defended

well, but after this move he gets in serious trouble. The idea of relieving the back rank in order to prevent lle 1 in response to ... lle4 is correct, but he has chosen the wrong pawn move. The alternatives are:

I) 19 ... :e4? 20 lle l ! is good for White.

2) 19 .. .f6 20 h3 lle5 2 1 r!di c5 22 ixe5 fxe5 23 g4 lbd4 (23...l0e7 24 :d3 !, followed by llb3, and White will win at least a pawn) 24 f4! is very good for White, since Black's queenside pieces are still immob ilized. If Black continues 24 ... exf4, then 25 J:tei followed by :le8 wins a piece.

3) 19 ... h5!? and now: Ja) 20 dxc6 (White can always

bail out this way) 20 ... bxc6 21 .i.xc6 :Xb6! (better than 21 ... J:te7 22 �d5 or 21...l:td8 22 c5!) 22 .i.xe8 J:txb2 and Black should be fine.

3b) 20 l:tdi J:te4! is very unpleasant for White.

3c) 20 J:te1 llxel + 2 1 �xel f6 is a risky line for White.

10 h3 Now Black is in difficulties.

10 J:te4

20 ... l:td8 21 g4 �4 22 c5 is good for White.

21 .i.h2 Black's problem is that ... h6 has

not prevented White's g4, as 19 ... h5 would have done.

21 cxdS (D)

22 g4 White's threat is not so much 23

gxf5 as 23 .i.c2. Black is obliged to surrender material in the hope of obtaining three pawns for a piece.

22 lhc4 23 ltlxc4 dxc4 24 llel

This intermezzo does the trick -White succeeds in exchanging rooks. After 24 lldl .i.e6! (but not 24 ... b5 25 .i.c2 �h4 26 .i.xd6 llb7 27 .i.g3 threatening both mate and the h4-knight) 25 gxf5 .i.xf5 Black avoids the rook-swap.

24 ••• .i.e6 Or 24 ... ltld4 25 lle8+ �h7 26

.i.xd6 lla8 27 .i.e5 b5 (27...�c6 28

2/0 VISHY ANAND: MY BEST GAMES OF CHESS

�c2+ f5 29 f4! is hopeless; the threat is 30 llxcS, and if Black exchanges on e5 the passed pawn will decide) 28 �d1 lLlc6 29 �c2+ f5 30 l:.xcS llxc8 3 1 �xf5+ g6 32 �xc8 lLlxe5

36 �f3 37 �b2 (DJ

33 ha6 b4 34 �e2 and the minor- B

piece ending is won for White. 25 gxf5 �xr5 26 �xd6 �xh3+ 27 �g1 lldB 28 lle8+ llxeB 29 �xe8 �e6

Black has three pawns for the bishop, but the pawns are far back and Black has only one passed pawn, so White should win, although care is required.

30 a4! g5 31 aS

Now there is only the kingside to worry about.

31 <j;g7 32 �a4 �g6 33 �d1

After 33 �c2+ �f5 34 �1 �e4 Black's bishop takes up its optimum square.

33 �d5 34 �c2+ �6 35 JJ..c7

Stopping 35 ... h5 because of the reply 36 �d8+.

35 ••. �e6 36 �h7

Now White prevents .. .f5. Black is gradually running out of active moves, when it will be time for the white king to advance.

37 ••• � 38 �c2!

Not 38 �g3? �e4. It is still too early to allow the exchange of bish-ops.

38 .i.e4 39 �d1 �4 40 �e2 �d3 41 �b6+

The two bishops form an effective team.

41 �d5 42 �d1 fS 43 �g3 �eS 44 �c5 �6 45 �b5 f4+

Or 45 ... �g7 46 � �h7 47 �f7, followed by �d5, and one of Black's queenside pawns falls.

46 �b2 1-0 The pawns are blockaded and

White will soon win one by eilher �f3 or �f8.

Game39

V. Anand - A. Karpov Las Palmas 1 996

Queen's Gambit Accepted

In round 6 I slid back to 50% after a horrible loss to Kramnik in which I failed to put up any resistance. I spent the evening disgusted with my play and decided to adopt an uncompromising style the next day against Karpov. The result was my best game of the tournament.

1 �3 There was no way I could face a

boring Caro-Kann and trying to deal with an improvement on move 45 leading to a difficult ending, etc. I felt that it would be better to go down in style than to do something like that. Now, how does a move like I llli3 allow me to get interesting positions? Well, to be honest, I llli"3 can lead to positions even more boring than after 1 e4 c6, but at least they would be unfamiliar boring positions!

1 dS :z d4 e6 3 c4 dxc4 4 e4

I played this without hesitation as I wanted to liven things up immediately.

4 ••• bS s a4 c6

6 axbS 7 b3

cxbS

He had already started thinking a lot and I knew that he wasn't familiar with this opening- not that there is a great deal of theory on it.

7 .b7 8 bxc4 .ixe4 9 cxbS �6

10 .ie2 I 0 .id3 has b.en played more fre

quently. 10 11 0-0

12 lbc3

h7 0-0 (D)

The game Lutz-P.Schlosser, Berlin 1989 continued 12 lbbd2 .ib7 1 3 lbc4 a6 1 4 b6 lbc6 15 .id2 lL!dS


with an obscure position - it isn't clear whether the b6-pawn will be weak or strong. Developing the hiknight to c3 appears more natural as it exerts some influence over the important d5-square.

12 ••. i.b7 13 lDeS a6

After 13 ... i.b4 14 i.b2 i.xc3 (or 14 ... a6 15 i.f3) 15 i.xc3 a6 (15 ... 'i!t"d5 16 tt:lf3 is also a little better for White) 16 i.f3! (better than 16 i.a5 'i!t"d5 ! 17 i.f3 'i!t"xb5 and White cannot profit from the b-fi.le line-up) White has an edge.

14 ..U3 (D)

14 Karpov finds the safest solution,

leaving himself with only a slight disadvantage. Black can also try 14 ... i.xf3 15 'Wxf3 'i!t"xd4 16 'l'xa8 'i'xc3 17 i.f4, when again White has a slight edge.

15 tt:\xdS exdS 16 l:tbl!

After 16 'Wb3 axb5 17 l:txa8 .b.a8 18 'Wxb5 White has an edge,but facing Karpov's defensive skills I preferred to aim for a large advantage!

16 ... 'i'b6 17 i.e2!!

I decided that the bishop had nothing more to do on f3 and the best plan was to relocate it to d3.

17 ... axbS There is no choice; both 17 ... aS

and 17 ... f6 18 i.e3! a5 19 00 00 20 i.f3 would leave White with a very strong passed b-pawn.

18 lbbS 1kc7 19 i.r4 i.d6 20 i.d3 h6 (D)

After 20 ... i.c6 21 J:1b3, the white pieces are ominously aimed at the black kingside.

21 i.xh7+! Here, I spent a few seconds look·

ing at 21 .:xd5, which leaves White with a clear extra pawn, but as I mentioned earlier I couldn't face

ANAND - KARPOV, LAS PALMAS 1996 213

long, technical game. In many lines Blackcan exchange on e5, leaving a position with 4 vs 3 on one side. Depending on which pieces are left, this might or might not be a win, but the game would certainly continue for a long time. Then I saw ..txh7+ and didn't waste any more time on lhd5. I spent some time analysing the sacrifice, and didn't see a defence foc Black. By now I was too excited to analyse and decided that I would simply play it. Karpov had hardly any time left and I was sure he wouldn't find a defence.

Perhaps this decision was somewhat reckless but I wasn't punished for it- indeed, I was rewarded with a nice win.

21 �7 22 'l'hS+ �8 23 l'l.b3 (D)

2J ..txeS? After this error there is no saving

Blade's position. There were two

alternatives that would have offered Black more defensive chances, although White retains a very dangerous attack in every line. Certainly it would be a monumental task to defend this position over the board, especially taking into account Black's time shortage.

The alternatives are: I) 23 ... ..tc8 24 l'l.g3 and now: Ia) 24 ... l'l.a3 (this move was sug-

gested by a New in Chess reader, Maarten de Zeeuw) (D) with a further branch:

lal) 25 l'l.xg7+ �xg7 26 ..th6+ �6 27 l'l.el l'l.g8 28 'ifh4+ (28 �g6 l'l.xg6 29 'irh4+ �5 30 'irh5+ is perpetual check, while 28 f4 is unclear, but could also be a draw) 28 ... �e6 29 �g4+ �d7 30 �6+ �c6 31 ::.C I+ �b7 32 l'l.xc7+ ..txc7 looks unclear.

la2) 25 f3 'ile7 and White has various attacking ideas:

la21) 26 ..th6 ..txe5 27 dxe5 g6 28 ..txf8 'ira?+ 29 �h l l'l.al 30

214 VTSHY ANAND: MY BEST GAMES OF CHESS

l:bg6+ fxg6 3 1 'it'xg6+ leads to a draw.

1a22) 26 J.g5 f6 (26 ... 'it'e6 27 .l:th3) 27 lLlg6 'l'e8 and White has no effective way to proceed.

1a23) 26 llc1 ! 'iWf6 27 J.h6 (27 J.g5 J.xe5 28 hf6 hf6 29 'iWxd5 J:td3 is a likely draw) 27 ... J.xe5 (if Black allows White to take on g7 then White should have the advantage) 28 dxe5 "i1Vh6+ 29 Wh1 g6 30 J.xf8 .l:tc3 3 1 'it'h6 .l:txc 1 + 3 2 'it' xc 1 and White wins.

1b) 24 ... 'it'e7 (D) and now:

1b1) 25 J.g5 and now Black should play 25 ... f6 26 lLlg6 'lieS, with an unclear position, rather than 25 ... 'it'e6, when 26 .l:th3 'iWxh3 27 gxh3 f6 28 lLlg6 fxg5 29 'iWh8+ Wf7 30 lLlxf8 J.xf8 3 1 f4 g4 32 'iWh5+ is very promising for White.

1b2) 25 J.h6! i.xe5 26 dxe5 g6 27 e6! i.xe6 (27 ... 'iWxe6 28 i.xf8 wins material) 28 'it'e5 f6 29 .l:txg6+ and White wins.

2) 23 .. 16! 24 .l:th3 fxe5 (24 ... ..ixe5 25 dxe5 transposes to the game) 25 dxe5 'lfc4! (25 ... .1:txf4 26 e6 � 27 'iWhS+ We7 28 1Wxg7+ �xe6 29 l:te1+ .l:te4 30 D.h6+ furces mate) and now:

2a) 26 1Wh7+ � 27 e6t �6 (27 ... Wxe6 28 l:le1+! wins focWhite and 27...We8 28 1Wg6+ �d8 29 i.g5+ 'ili'c8 30 .l:tc I is clearly better for White) with another fork:

2a1) 28 .J:[h6+? gxh6 29 •xh6f �5 (29 . . . We7 30 i.g5+ mates) 30 g4+ We4 3 1 l:te1+ 1i'e2! (3l...�d3 32 'it' g6+ Wd4 33 i.xd6 probably favours White, but is extremely messy; I didn't bother to analyse this line deeply as the strength of 3L'I'e2! made it irrelevant) 32 :Xe2+ i.xe2 33 i.xd6 .l:tal+ 34 Wg2 i.fl-+- 35 Wg3 l:tf3+ 36 Wh4 .l:th3+ and Blade wins.

2a2) 28 i.g5+ 'ili'xe6 29 :CI+ Wd7 (not 29 ... i.e5 30 :Xe5+ �d6 31 .l:tel lLlc6 32 1Wg6+ �7 JJ •xg7+ Wb6 34l:tb1 + i.b5 35 l:thbJ winning for White) 30 .. xg7+ �c6 3 1 .l:tc3 lLld7 32 .l:txc4+ hc4 with an unclear position. White has a lot of pawns on the kingside, but since Black's king is now safe he can acti· vate his pieces.

2b) 26 .l:tel ! (cutting off the en· emy Icing's escape route) 26 .. . i'xf4 27 'iWh7+ Wf7 28 exd6 and now:

2bl ) 28 ... .1:te8 29 _.h5+ g6 30 l:te7+! ! .l:txe7 (30 ... Wffi 31 i'fJ! is an unexpected win) 31 'i'h7+ Wf6

ANAND - KARPOV, LAS PALMAS 1996 215

(31...Tf8 32 dxe7+ We8 33 'i'xg6+ Wxe7 34 llh7+ mates )32 'i'xe7+ Wf5 33 'i'f8+ We5 34 lte3+ with a decisive advantage for White.

2b2) 28...ll:lc6 29 ltf3 'i'xf3 30 gxf3 .1c4 31 Wh l. White has the advantage because of his d-pawn and persisting attack, but this was Black's best chance.

24 :b3 f6 25 dxeS '¥iie7

25 ... '1'c4 26 ltel 'ifxf4 27 '¥iih7+ �n28e6+We8 29 'illg6+ is also no help for Bli!.!<k.

26 Wh7+ �7 1:1 llg3 (D)

1:1 ·- �e8 27 ... 1:tg8 28 'if g6+ �f8 29 exf6 is devastating.

28 lbg7 From now on, the game pretty

much plays itself - there are several ways to win on every move ! One alternative is 28 exf6! gxf6 (the lines 28 ... llxf6 29 llxg7 'l'e6 30 'l'h5+ llf7 31 llxf7 'i'xf7 32 lle I+ and 28 ... 'i'xf6 29 llel+ are also terminal) 29 lle3 'ilfxe3 30 fxe3 i.xfl 3 1 i.d6 and White will have a decisive material advantage.

28 'ille6 29 exf6 ll:lc6 30 lla1 �d8 31 h4

To clear the back rank- it's always nice to have time for such details.

31 OM i.b7 3l . . .li:ld4 loses to 32 i.c7+ �c8

33 i.a5. 32 llcl i.a6 33 lla1

Again, there are other routes to victory, e.g. 33 i.c7+ �c8 34 i.b6.

33 i.b7 34 lld1 i.a6 35 11'bl! lbf6 36 i.gS �c8

1-0 Karpov lost on time while in the

act of playing 36 ... 1t>c8. The reply 37 'l'b6 wins on the spot.

Winning this game gave me a 6-3 score against Karpov in decisive games. My other games ended in draws and my score of+ 1 was sufficient for out

right second place behind Kasparov. Finally, here is an effort of which I am very proud, from the Credit Suisse

chess festival in Biel (July-August 1997).

Game 40

V. Anand - J . Lautier Bie/ 1 99 7

Scandinavian Defence

At the opening ceremony the previous day, the chess players had to play a match against representatives from the Swiss Skiing Federation. Each participant from the A and B tournaments had to play two moves and if they didn't mate the skiers within 24 moves, then the skiers won. The chess players chose the Scandinavian. To my great surprise, this is exactly what happened in my first round game!

1 e4 dS I don't recall Joel ever having

played this before, but he hadn't competed at all since Monaco in April, so I assumed this was an opening he had prepared during the intervening three months.

2 exdS 3 lOc3 4 d4 5 ltlf3 6 .i.c4

'lll'xdS 'ill' aS ltlf6 c6

Nowadays 6ltle5 is more popular, but during my preparations for the World Championship, I noticed that the lines with 6 .i.c4 were very dangerous for Black to navigate. Knowing that Joel didn't have a great deal of experience with the Scandinavian,

I decided to test him in this critical variation.

6 ••• tis 7 ltJe5

The variations with 7 .i.d2 leave White with a slight edge, but I remembered that 7 ltle5 and 8 g4 gave Black more problems.

7 e6 8 g4 .i.g6 9 h4 ltlbd7! (D)

Joel chooses the best line. After 9 ... .i.b4 10 .i.d2 ltle4 1 1 f3! White gained some advantage in CaqJOra •

Curt Hansen, Palma de Mallorca 1989 and it was after this game that people started to look at 9 . ..li:lbd7 more seriously.

ANAND - LAUTIER, BIEL 1997 217

10 �xd7 i2Jxd7 ll hS h4 12 m.3 .i.g2

A nice finesse - if White plays :g3, then Black will gain a tempo with a later ... .i.d6. However, if the game continuation is correct, then Black will have to abandon his finesse and play 12 ... i.d5. Then 13 id3 i.d6 14 .i.d2 ii'c7 1 5 ll'lxd5 cxd5 16 'ili'e2 i.f4? 17 0-0-0 0-0-0 18 i.xb7 won a pawn for White in Ochoa de Echagiien-Denker, New York Open 1989, but of course this was not forced.

13 :le3! I bad wanted to play 13 :lg3,

which gives White a slight edge despite the fact that Black can play ... id6, but a fresh look at a position during a game can often tum up better moves than those found during home preparation!

13 ... ll'lb6 After 13 ... b5 14 .i.d3 b4 1 5 ll'le4

Black will have to play ... i.xe4 sooner or later, when White will be better due to his two bishops. Understandably, Joel didn't want to resign himself to an inferior position with few prospects of counterplay.

14 i.d3! This move, which I found at the

board, was the reason I decided to go foc 13 :e3 instead of 13 :lg3.

14 .ib37! is inferior after 14 ... c5!, when Black has good counterplay.

14 ll'ld5 (DJ

The obvious reply, attacking c3 and e3.

15 f3! This remarkably calm move is the

point behind White's play. He is willing to jettison a few pawns and/or the exchange in order to snare the bishop on g2. When the bishop is finally trapped, Black will probably end up with a rook and two pawns for two minor pieces. Owing to the lack of open files, the minor pieces will be very much more at home in the resulting position than the rooks and this, coupled with White's lead in development, will almost inevitably give him a clear advantage.

In C.Bauer-Prie, French Championship 1996, the continuation 15 .:g3 ll:lxc3 1 6 bxc3 i.d5 1 7 i.d2 11fa4 1 8 1i'e2 b5 19 h6 0-0-0 was fine for Black.

15 ... .i.b4 (D) After 15 ... ll:lxc3 16 bxc3 11fxc3+

17 .i.d2 11fxd4 18 �f2 .i.xf3 19


�xf3 or 15 . ..fi)xe3 16 i.xe3 'irb6 17 �£2 i.h3 1 8 ltbl a position of the type mentioned above is reached; White holds the advantage in both cases.

16 �f2! i.xc3 The above comment also applies

to the line 16 ... lllxc3 17 bxc3 i.xc3 1 8 ltbl i.xd4 19 �xg2 i.xe3 20 i.xe3.

17 bxc3 "i'xc3 18 J:l.b1 "i'xd4

White wins after 1 8 ... i.xf3 19 'i'xf3 'i'xd4 20 J:l.xb7 0-0 21 We4.

19 J:l.xb7 J:l.d8 (D) The other critical variation runs

19 ... i.h3 20 J:l.xf7! (I stopped here, but Joel saw two moves further!) 20 ... c5 (Black simply protects his queen and threatens 2 1... Wxf7; if instead 20 ... lllxe3, then 21 i.xe3 'i'd6 22 l:l.f4 is very good for White) 21 J:l.f5 ! ! lllxe3 22 i.xe3 Wb2 23 J:l.xc5 0-0 24 Wg3 ! winning, as Black will be lucky to get a single extra pawn.

If 1 9 ... lllf4, then 20 �g3 'i'd6 21 i.a3! lllxh5+ (2l ..."iixa3 22 i.e4! also wins) 22 Wxg2 'irg3+ 23 �fl is winning for White.

After the text-move, I saw the possibility of i.g6 but then I realized that it didn't work immediately because Black could run with his king, e.g. 20 i.g6 'irxd 1 21 J:l.xe6+ �f8 22 i.a3+ (or 22 J:l.xf7+ �g8) 22...1Lle7 23 i.xe7+ �g8, and the attack fails.

Then I saw the possibility of ioserting h6 at the start of the combination, when a later ... �g8 could be met by J:l.g7+!. Suddenly, all that remained was to check the details ...

20 h6!! gxh6? This gave me a chance fcr a really

beautiful finish. Black could still fight on with 20 ... lllxe3, txt Joel hadn't seen the idea behind h6!. The analysis runs:

1) 20 ... g6 2 1 i.xg6! 1fxdl 22 J:l.xe6+ �f8 23 J:l.xf7+ �g8 24 J:tg 7+ �f8 25 i.a3+ followed by mate.

ANAND - LAU11ER, BIEL 1997 219

2) 20 ... liJxe3 (absolutely the only move that doesn't lose by force) 2I .he3 'iWe5 22 hxg7 llg8 and now 23 lei ! threatening 'iW a3 and 'ili>xg2 is strong (not 23 ..ih6 'iWh2 !).

21 .ig6!! (D)

21 ••• liJe7 There is no way out: I) 2 I ...'i'xe3+ 22 ..ixe3 fxg6 23

J.c5 wins.

2) 2 1 ...1Wf6 22 ..ixf7+ ilxf7 23 l:lxf7 ll'lxe3 24 ilxd8+! (24 'iWe2 ll'ld I+ 25 'ili>xg2 'ili>xf7 is less clear, although White remains much better after 26 1!Ve4) 24 ... 'ili>xd8 25 ..ixe3 ..ih3 26 D.xa7 lle8 27 l:l.xh7 and Black loses several pawns, followed by his bishop!

3) The key variation is 2 I ...ilxdi 22 D.xe6+ 'ili>f8 23 ..ixh6+ �g8 24 ..ixf7#.

22 11Vxd4 llxd4 23 lld3!

There is no nlllld to bother with 23 D.xe6 lld7 when the text-move wins effortlessly.

23 lld8 24 lhd8+ �d8 25 ..id3! 1-0

Because after 25 ... -th I 26 ..ib2 l:l.e8 27 ..if6 Black will soon be in zugzwang and have to surrender at least a piece.

Biel l997 is a tournament that I remember very fondly. Biel was and still is a superbly organized tournament, with very comfortable conditions for the players and pleasant walks by the lake. I felt very happy there and this was reflected in my chess. The first game against Lautier (Game 40) was a game I could be proud of and I also won a nice game against Gelfand. After a further win against Pelletier, I faced Lautier for the second time.

Game 41

J . Lautier - V. Anand Bie/ 1 99 7

Sicil ian, Najdorf

1 e4 A minor surprise - Joel does play

both first moves, but at the time 1 d4 was his main weapon. Earlier in 1997, at Ubeda, he played I e4 four times in six games and I suppose that he was in the process of enlarging his opening repertoire.

1 ... cS It is often tempting to look for

something offbeat against people who vary from their 'normal' routine, but this can backfire unless you know what you are doing! Caught by surprise, I decided to stick to something I knew.

2 'Llf3 d6 3 d4 cxd4 4 'Llxd4 lt::lf6 S 'Llc3 a6 6 .ic4

The next surprise. I had no idea what line Joel would choose, but since he had a lot of experience playing on the black side of 6 .te2 and 6 .te3, I thought he might go for one of those.

6 ... e6 7 .ib3 bS

One of the main lines is 7 ... 'Llbd7, but I had something else in mind.

8 0-0 b4 Once again, 8 ... .te7 9 'i'f3 would

lead to a large body of theory. I had studied the slightly offbeat 8 ... b4 before the last VSB tournament in Amsterdam 1996 and found t to be surprisingly interesting. White is obliged to place his knight offside on a4, but the defect is that Black gets a weak pawn on b4. This was my first opportunity to try the idea out.

9 lt::la4 .td7 9 ... lt::lxe4?! 10 f4 transposes tothe

famous Fischer-Tal game from the Candidates in 1959. 10 f4 lLlc6 (D)

11 .teJ

LA UTIER - ANAND, BIEL 1997 221

I I f5 is the sharpest tty. After ll . .. e5 ( l i ...lLixd4? 12 'iWxd4 e5 13 l'xb4 d5 14 'Wei ! favours White) 12 �6 (asso often in the Najdorf, there is no turning back; White has a badly placed knight on a4 and a weak pawn on e4, so he must throw caution to the winds) 12 .. .fxe6 13 fxe6 �c8 14 1g5 Black can play:

1) 14 ... h5 15 �xf6 gxf6 16 �d5 1b7 17 c3 ( 17 a3 is a possible alternative) and now Kalegin and Dvoirys just stop, saying that White has compensation, but is that true? White certainly has a lot of trumps - pawn on e6, the open f-file and Black's exposed king, but Black has an extra piece and when White plays cxb4 to rescue thestranded knight, Black's knight gains access to the d4-square.

2) 14...�e7 15 �xf6 bf6 16 J:xf6 'i'xf6 17 'Wxd6 lLld4 18 lLlb6 (D) and now:

2a) 18. ..�xe6 19 lLlxa8 (the alternative 19 he6 J:d8 20 �d7 + �f7

21 'iWd5+ is interesting) 19 ... lLle2+ 20 �h 1 lLlg3+ led to a draw in Kalegin-Dvoirys, USSR 1988. Interested readers can find the relevant game and analysis in lnformator45 (game 282).

2b) 18 ... lLle2+ 19 �hl lLlg3+ 20 hxg3 'Wh6+ 21 �gl 'We3+ 22 �fl J:f8+ 23 'Wxf8+ �xf8 24 e7+ �e8 25 �a4+ �d7 26 �xd7+ �f7 27 lLlxa8 �xe7 28 .:Z.dl was a line given by Kalegin and Dvoirys; they continued 28 ... '1Vxe4 and assessed the resulting position as clearly better for White. However, after the improvement 28 ... 11t'xg3! (VujadinovicNovak, e-mail 1999) the removal of the g3-pawn makes it very hard for White to avoid the coming checks.

11 ••• .:Z.b8 Not l l ....i.e7? 1 2 f5 lLlxd4 (12 ... e5

13 lLle6 fxe6 14 fxe6 .i.c8 15 �b6 traps the qu.en) 13 1Wxd4 e5 14 1Wxb4 and White stands very well.

12 c3 �e7 13 e5?!

13 cxb4 .:Z.xb4 14 lLlc3 0-0 leads to equality, so White tries for more. However, in my preparations I had already come to the conclusion that the text-move is dubious because sacrificing the exchange gives Black a big advantage. Fritz is not convinced because it finds a very convoluted way for White to keep the extra exchange, but I took one look and decided that Black must be better. The main problem for White, besides his


poorly placed a4-knight, is his kingside. Without his e- and f-pawns, his king is quite vulnerable, and every one of Black's minor pieces is well positioned to join the attack. White's Sozio bishop on b3 is totally misplaced in this situation, since it cannot defend the kingside.

Despite all these factors, I was surprised at how quickly White's position went downhill.

w

13 dxeS 14 fxeS lL!xeS 15 .i.f4 lL!g6! 16 i.xb8 1Vxb8 17 cxb4 i.xb4 (D)

18 l:lcl After 18 lLlf3, Black gains the ad

vantage by 18 ... i.b5! 1 9 l:lf2 lL!g4. 18 ••• 0-0 19 �1 (D)

Since this move loses material, one might expect there to be something better. However, it may well be that White is practically lost already.

The alternatives for White are not very attractive:

1 ) 19 lL!c5? i.xc5 20 l:xc5 'lb6 and Black wins material due to the pin; for example, 21 'ii'c2 lLle4 22 l:lc4 lbd6 23 l:lc5 e5!.

2) 1 9 lLlf3 and now: 2a) 19 ... l:ld8?! 20 'ii'c2 l:c8 21

'ii'f2! (21 lL!c3 loses material after either 2I . ..lL!g4 or 21 ... i.bS) and the position is not so clear.

2b) 19 ... i.d6! 20 lLlc3 .ic6! (the immediate 20 ... lL!g4 also looks good) and White is in serious trouble:

2bl) 21 'ii'e2 lL!g4! 22 h3 (22 g3 i.c5+ 23 �hi lLl6e5 24 lLle4 ibS wins for Black) 22 ... i.xf3 23 gxf3 (23 lhf3 i.c5+ is also winning for Black) 23 ... lLJf4 with a decisive attack.

2b2) After 21 lbe2 i.xf3 22M i.xh2+ 23 Whl i.e5 Black stands very well. He has two pawns for the exchange, and while his own king is completely safe, White's is seriously exposed.

LAUTIER - ANAND, 8/EL 1997 223

19 ... l:ld8 Over the board, I wasn't totally

sure that this natural move won material,but in fact White must already return the exchange due to the threats along the d-file.

20 li)cS (D)

.txc5

the white king, which is very insecure thanks to its scanty pawn-cover.

23 ••• .IteR 24 �c6 .i.xc6 25 J:xc6 a5

The rest of the game is fairly sim-ple - White's king is too exposed.

26 'ilf2 �g7 27 b3 J:td7 28 J:tc5 'i' d8 29 ife3 J:td3 (D)

White is gradually driven back and Black moves in to occupy the centre.

30 'irel 'i'd4

ldidn 't want to allow unnecessary complications. Still, 20 ... .tb5 was also good; for example, 21 �xe6 (White should prefer 21 J:txf6 gxf6 22 'ill'gl, which will transpose to the game as I don't really see a better move than 22 ... .txc5) 21.. .fxe6 22 �xe6 :xd l 23 l:tfxd l .lta5! and Black wins.

21 :xeS 22 :xt'6 23 'i'gl

Black can give up the a-pawn, since the game will be decided by a direct attack against the white king.

31 :Xa5 lle3 .tb5 gxf6

White frees himself from the dfile,pin but at a heav)Price. Instead of being material up, he is now a clear pawn down.Moreover, Black has good attacking chances against

32 'li'cl 11i'e4 Threatening 33 ... 1lxh3+.

33 �h2 11i'f4+ 34 'it>g1 'ilfg3

0-1 After 35 <Jolfl {i)f4, White has to

surrender his queen (for a start).


After this second excellent win against Lautier, I was euphoric. Evidently this sensation went rather to my head because the very next day I lost to Mid lov. However, l went on to win the tournament in any case, half a point ahead of Karpov.

The following game was played under entirely different circumstances lost to Lautier in the first round at Belgrade, and then had the unpleasent prospect of facing Kramnik with Black the following day. It seems as if the tournament might not get off to a good start. ..

Game 42

V. Kramnik - V. Anand Belgrade 1 997

Semi-Slav

1 m lDr6 2 c:4 e6 3 &3 dS 4 d4 c6

I decided to play the Semi-Slav, although this was quite a risky choice as at the time it was one of Kramnik's favourite openings. Playing one of your opponent's preferred openings is an accepted strategy, and it can be quite successful. For example, Kiril Georgiev, who was one of the world's leading experts on the Dragon, has nevertheless lost a number of games on the white side. It is not easy to reoriantate yourself to the other side of board and you can even end up in the situation of 'knowing too much ', and be unable to decide which line to play. Having said that, it is a doubleedge sttategy and in this game I was in big trouble right in the opening.

5 ..tgS h6 6 ..th4 dxc4 7 e4 gS

5 8 ..tg3 b 9 �

10 eS 11 a4?!

..tb7 lObS (D)

White should probably have tried the immediate 1 1 lllxg5. Apparently

w

Krarnnik wasn't convinced by this and decided to 'improve' it with a preliminary a4, trying to provoke Black into replying l l ...a6. However, it turns out that Black has several good alternatives to pushing his a-pawn, so the critical line is 1 1 lllxg5 lllxg3 12 lllxf7 �xf7 1 3 fxg3, and now:

1) 1 3...�g8 14 0-0 and now: Ia) 14 ... llla6?. In my original

notes, I stopped here with the comment that ... lllc7 would defend the weakness on e6. In fact, this is irrelevant and White wins by 15 llle4 ..tg7 16 ..th5, when I don't see any defence. A sample line is 16 ... J:lh7 17 llld6 'i1Fb6 1 8 ..tg6 ..txe5 19 ..txh7+


�xh7 20 'ilfc2+ 'ili>g8 21 'ilfg6+ .i.g7 22 l:!.ad l ! with decisive threats.

lb) 14 ... l:!.h7 (this is often a useful move) 15 .i.g4 �h8 (15 ... 'iWe8 16 .!Lle4 and 15 . . . l:!.e7 16 .i.xe6+! are good for White) 16 .i.xe6 and I suspect that White's compensation is more than enough.

lc) 14 ... �d7 (just as in the game) 15 .i.g4 (15 .i.xc4 bxc4 16 'iWg4+ .i.g7 17 'i!he6+ �h7 is a perpetual check if White wants it, but he cannot achieve more; e.g., 18 l:!.f7 l:!.f8 ! 19 l:!.xd7 .i.c8 20 l:!.xd8 .i.xe6 and the two bishops are potentially dangerous) 15 ... 'ilfe7 and now:

le i ) 16 �e4 (D) is less of a problem for Black than it is with a4 and ... a6 included. He can try:

Je l l ) 16 ... c5 17 l:!.f6 (the best idea) 17 ... .i.d5 18 l:!.g6+ �h7 with a final fork in the path:

l c l l l ) 19 'ilfc2 cxd4 20 l:!.xe6! .i.xe6 21 �g5+ �g7 22 �xe6+ �g8 23 1i'g6+ .i.g7 24 �xg7 �f8 25

'iWe4! 'ilxg7 26 'ilfd5+ �h7 27 l:!.fl (27 .i.f5+ lbg6 28 .i.xg6+ 'ilfxg6 29 1Vb7+ is also a draw) 27 ... 'ilfxg4 28 l:!.fl+ lii>g6 29 l:lf6+ with a draw.

lc1 12) 19 1Vbl ! �xg6 20 �6+ �g7 2 1 .i.h5 'ilfxf6 22 exf6+ � looks unclear to me.

l c l 2) 16 ... l:!.h7! 17 ed6 :b8 18 a4 and now instead of transposing to the game with 1 8 ... a6, Black can play 1 8 ... c5! 19 axb5 cxd4 20 I'xd4 .i.d5 21 l:!.xa7 'ilfd8. White has some compensation for the piece, but Black has defended the e6-square and certainly has chances in the battle to come.

lc2) 1 6 'ilfc2! and in this line I don't see any significant difference from the situation with a4 and ... a6 (see the note to White's 17th move).

2) 13... �e8 14 (}.() �a6 aims to run with the black king 10 the queenside. My hunch is that White has compensation, but it would take us too far afield to analyse this position exhaustively.

It is clear that the early sacrifice on g5, while dangerous, entails a degree of risk for White as well. Now let's get back to the game.

11 •.• a6? White's plan works! To be honest

�xg5xf7 had crossed my mind, but I dido 't fully believe it. Meanwhile I made this 'automatic' move. Black should have played l l ...b4 12 .!Lle4, when he has a choice of good alter· natives:

KRAMNIK - ANAND, BELGRADE 1997 227

1) 12 ... cS is a solid continuation: Ia) After 13 �xeS .ixcS 14 dxcS

'fxd1+ 1S J:xd1 �d7 Black has a very comfonable ending, V.PopovDreev, Russian Team Championship, Kazan 199S.

lb) 13 �d2 and then: 1bl) 13 ... "ii'xd4?! 14 .ixhS .ixe4

15 /Llxe4 W'xe4+ 16 q;,n and now 16 ... 1'b7 fails to 17 'irf3 ! �c6 18 ixf7+, so Black would have to find 16 .. . g4 17 .ixg4 11Vb7 18 'i'f3 �c6 19 ih5 0-0-0 20 .ixf7 �d4 with an unclear position.

!b2) 13...�g3 is simpler: lb21) 14 hxg3 "ii'xd4 1S .if3

li:kl7! kills the line for White; e.g., 16 liJd6+ .ixd6 17 .ixb7 �xeS !.

lb22) 14 �f6+ 11e7 1S hxg3 'fxd4 16 �xc4 .ig7 and White might have some compensation, but one certainly cannot say more than that.

2) 12 ... g4 and now: 2a) 13 �d2 "ii'xd4 14 .ixc4

li.lxg3 1S hxg3 �7 16 "ii'xg4 "ii'xb2 17 l:.b1 'l'xeS 18 l:thS �f6! is fine for Black.

2b) 13 �4 cS 14 �xeS .ixcS 1S dxc5 'l'xd1+ 16 J:xd1 lt:lxg3 17

hxg3 h5 (17 ... �d7 is also possible) 18 ixc4 �e7! leaves White with a sad-looking knight on h4.

2c) 13 .th4 "ii'd5 14 �d2 (14 ott"6+ �f6 l S .ixf6 J:g8 is clearly better for Black) 14 ... c3 also looks reasonable for Black.

Note that l l...�xg3 is less effective, because 12 hxg3 .ib4 13 �fl.

intending �e4, gives White quite good compensation.

12 �xgS! Now this is very strong.

12 �g3 13 �'Kf7 � 14 fxg3 1;g8

After 14 ... 1Je8 lS 0-0 Black sorely misses the defence ... �a6-c7, and 1S ... 1Jd7 1 6 .ig4 gives White a strong attack.

15 0..0 16 .tg4

17 ll)e4

�d7 11e7 (D)

1 7 "ii'c2 is a dangerous alternative. The critical line runs 17 ... l:th7 1 8 W'g6+ �h8 19 .ixe6 liVgS 20 'i'e4 l:te7!, and now:

l ) 2 1 axbS l:txe6 (not 2 l ...axbS? 22 J:xa8 .ixa8 23 .ixd7) 22 J:f7 lLlf6 (22 ... .ig7? 23 h4 1ih5 24 l:xd7 axbS 2S :n .ta6 26 J:fS 1IV g6 27 h5 "ii'xg3 28 �e2 wins for White) 23 1if3 (23 1if4 cS is slightly better for Black) and now:


:xh6 cxd4 27 e6 1V.xh6 (27...�e7 loses to 28 .l:tff6 dxc3 29 .l:.fg6) 28 i'xh6 .l:.h7 29 1Vg5+ .l:.g7 30 1Ve5 dxc3 31 bxc3 �e7 32 axb5 axb5 33 'i'xb5 .l:.d8 34 h4 favours White.

2b) 21...1Vd2 (D) and now:

2bl) 22 .1:.f2 'l'd3 2Hfg4 .1:.g7 24 lh4 c5 25 d5 �e5 is fine for Black.

2b2) 22 .l:.d1 1Vxb2 and now both 23 ixd7 1Vxc3 and 23 .l:.dn 1Vxc3 24 ixd7 �g7 favour Black.

2b3) 22 axb5 is interesting, but Black can defend:

2b31) 22 ... 1Vxb2?! 23 :an with another branch:

2b311) 23 ... 1Vxc3 24 �xd7 .l:.xd7 25 e6 'i'xd4+ (25 ... .1:.dd8 26 .:.n .tg7 27 .l:.xg7 'l'xd4+ 28 'l'xd4 .:.Xd4 29 llff7! and White wins) 26 1Vxd4+ Ld4 27 .l:.xf8+ .l:.xf8 28 .l:.xf8+ �g7 29 e7 wins for White.

2b312) 23 ... .1:.xe6 24 .:.n �g7 25 Ld7 'i'xc3 26 'i' g4 .l:.g8 (26 ... 1Ve3+ 27 �h1 .l:.g_8 28 'i'xe6 �c8 29 bxc6 c3 30 'i'c4 �xd7 31 cxd7 and White

is clearly better; e.g., 3 1 ..."1' g5 3 2 1Fxc3 1Vg4 33 'l'c6 .l:.d8 34 e6 '1We2 35 .l:.a1 a5 36 1Vd5 a4 37 h3 winning) 27 '1Wxe6 �c8 28 bxc6 �xd7 29 1Vxd7 1Ve3+ 30 �hi c3 3 1 c7 '1We2 32 .l:.g1 'l'c4 33 h3 c2 34 .l:.c1 1Vc3 35 'l'd8 'l'c4 36 e6 'l'xe6 37 d5 and White is winning.

2b32) 22 .. ..1:.xe6! 23 .l:.d I 'l'xb2 (23 ... cxb5 is also possible, with unclear play after 24 1Vxb7 1Ve3+ 25 'it>hl .l:.d8 26 'l'c7 .l:.ee8 27 .:.n ltlxe5 28 .l:.h7+ �g8 29 .:.n ltld7 30 .l:.xd7 .l:.xd7 3 1 'l'xd7 'l'f2 32 'l'd5+ �h8 33 'iVo) 24 .:.n ltlf6! (24...�g7 25 .l:.xd7 .l:.f8 26 ltle2 �c8 27 .l:.c7 'l'b3 28 .l:.al is only equal) 25 1Vf5 cxb5 and Black is better.

2b4) 22 .:.an is the critical line: 2b41) 22...�g7 23 .1:.5f2 1Vd3 24

1V g4! (24 .l:.f7 .l:.xf7 25 .:.Xf7 ltlf8 is less dangerous) and now:

2b41 1) 24 ... .1:.d8 25 1Vh4 .l:.ee8 26 �f7 .l:.f8 27 ltle2, heading for g6, leaves White better.

2b412) 24 ... .:.Xe6 25 1Vxe6 1Vxd4 26 1Ve7 ltlc5 (26...�xe5 27 .l:.d I and 26 ... �f8 27 'iVn ltlc5 28 ltle2 1Ve3 29ltlf4 �g7 30 �hi are winning for White) 27 e6 .l:.f8 28 ltle2 (again the ltle2-f4-g6 manoeuvre looks good) 28 ... 1Ve3 29 ltlf4 .l:.g8 30 �hi with a clear advantage for White.

2b42) 22 ... .1:.xe6 23 .l:.f7 �g7 (if 23 ... 1Vd3, then 24 .l:.xd7 is promising for White) 24 1Vg4 (24 .1:.xd7 .l:.xe5 is fine for Black) 24 ... .1:.g8 25 1Vxe6 (25 .1:.7f2? �xeS 26 1Vxg8+ �xg8


I a) 23 ... 1Vg4 24 'Wf2 (24 "l'xg4 �xg4 25 l:txb7 is unclear, while 24 exf6 'l'xd4+ 25 �h1 axb5 26 l:tf1 �g8 27 l:txb7 l:txf6 definitely favours Black) 24 ... cxb5 (24 ... c5 25 d5 .bd5 26 �xd5 �xd5 27 l:txf8+ l:txf8 28 1Vxf8+ �h7 29 .1:!.fl 1Vg8 30 'l'xc5 'l'd8 3 1 bxa6 l:txa6 32 'l'xc4 is good for White) 25 l:txb7 b4 26 exf6 bxc3 27 bxc3 with an edge for White.

1b) 23 ... ..ie7 24 .l:!.xe7 l:txe7 25 bxc6 �g4 26 cxb7 l:txb7 27 �h1 l:tab8 28 h3 �xeS 29 dxe5 'i'xe5 30 l:txa6 l:tb6 3 1 l:txb6 .l:!.xb6 should be a draw.

2) 2 1 l:tf5! (D) and now:

2a) 2 1 ...1Vg7 leads to a further branch:

2a 1) 22 .l:!.afl �c5 23 dxc5 l:txe6 24 1Ve3 ..ie7 25 �e4 l:tf8! and now Black can defend:

2a1 1) 26 �d6 .:txf5 27 l:txf5 ..ixd6 28 cxd6 ..ic8 29 'i'f3 and Black is holding on after 29 . . . �g8

30 'i'xc6 ..id7 or 29 ... l:te8 30 :t"6 'i'a7+ 31 �fl 'l'g7 32 'i'xc6..id7 33 'i'O bxa4 34 "iff4 c3!.

2a12) 26 'i'd4 l1d8! 27ll'ld6 i.xd6 28 cxd6 l:texd6 29 l:tf8+ .l:!.xf8 30

l:txf8+ 'i'xf8 3 1 exd6+ �g8 and White has to be satisfied with a draw.

2a2) 22 ..ixd7! l:txd7 23 l:tafl and now:

2a21) 23 ... c5 24 d5 b4 25 e6 :C7 (25 ... l:tdd8 26 �e2 is also good fer White) 26 �e2 l:td8 27 1Vxc4 with a clear plus for White.

2a22) 23 ... ..ib4 forces White Ill play accurately:

2a221 ) 24 'i'e3? .l:!.f8! 25 � l:txf5 26 l:txf5 ..ie7! (26 ... c5 27 "lf4 cxd4 28 e6 is satisfactory foc White) and White is slruggling foc compensation.

2a222) 24 l:th5 ! creates greater problems:

2a222 1) 24 ... ..ie7 loses to 25 lll7 1Vxf7 26 l:txh6+ �g8 27 e6 'i'g7 28 l:tg6.

2a2222) 24.. .�g8 25 :C6 c5 26 1Vf5 l:tf7 27 1Ve6 .i.c8 28 'i'e8+ :18 (28 ... 1Vf8 29 l:tg6+ l:tg7 30 lhg7+ �xg7 3 1 'i'c6 and White wins) 29 'i'c6 l:txf6 30 exf6 .i.b7 31 'l'e6+ 1Vf7 32 'i'g4+ �h7 33 0.e4 is again winning for White.

2a2223) 24 ... l:tf8 25 11f6 l:txf6 26 exf6 1Wf8 27 g4 and White has a clear advantage.

2a2224) 24 ... c5 25 'i'h4 �g8 (or 25 ... l:txd4 26 l:txh6+ �g8 27 'lh5 l:td2 28 l:th8+ 1Wxh8 29 'i'f7t) 26


27 l:l.xd2 .i.g7 gives Black a clear advantage) 25 ... 'ilfxd4+ 26 Whl ltJc5 27 'ife7 ! 'ife3! (Black's best try, the point being to have ... 'ilfel + available and prevent ltJe2-f4; the alternatives 27 ... .i.c8 28 l:l.xg7 l:l.xg7 29 l:l.f8+ Wh7 30 'ilfe8, 27 ... b4 28 l:l.xg7 l:l.xg7 29 l:l.f8+ �h7 30 'ife8 and 27 ... 'ilfxe5 28 l:l.xg7 'ilfxg7 29 'ilfxc5 offer Black little hope) 28 lM 1 'ifd4 (after 28 ... 'lWxe5 29 .l:txg7 'Wxg7 30 'ifxc5 White is better due to Black's exposed king) and now:

2b421) 29 e6 .i.c8 30 l:l.xg7 (30 l:l.7f4 liJe4 3 1 1fh4 1fd3 32 e7 liJg5 is also unclear) 30 ... l:l.xg7 3 1 1ff8+ �h7 32 1fxc8 1fe4! and Black has counterplay.

2b422) 29 h3 ! 'lWxe5 30 l:l.xg7 'ilxg7 3 1 'lWxc5 'lWxg3 32 liJe3 and although White has lost the g3-pawn, his king is still safer and Black cannot equalize; for example, 32 ... 'i'g7 (or 32 ... .i.c8 33 l:l.f6) 33 'i'd6 .i.c8 (33 ... c5 34 ltJg4 .i.e4 35 l:l.f6 also favours White) 34 l:l.f6 �h7 35 'i'f4! and White is clearly better.

Thus 17 'i'c2 would have been good for White, but there is no reason to criticize the text-move.

17 l:l.h7 The immediate attempt to break

out by 17 ... c5 (D) is also inadequate: 1) 18 liJd6 was discussed in the

post-mortem, but in fact Black can defend here:

Ia) 18 ... cxd4 19 l:l.f7 ! (19 ltJxb7 ltJxe5 20 'i'xd4 .i.g7 21 'iVe4 ltJxg4

w

22 'i'xg4 'i'xb7 23 'i'xe6+ Wh7 24 l:l.f7 'i'c8 is equal) 19 ... 'i'xf7 20 /fut7 �xf7 2 1 .i.f3! .i.d5 22 hd5 exd5 23 'i'f3+ � (23...�e7 24 'i'xd5) 24 lUI and White's attack is too strong.

lb) 1 8 ... .i.d5 19 .i.f3! and now: I bl ) 19 ... cxd4 20 .i.xd5 exd5 21

'i'g4+ .i.g7 22 l:l.f7 'i'xe5 23 'l'xd7 .i.f6 (23 ... d3 24 l:l.e7 'iVd4+ 25 �hi d2 26 l:l.dl and White wins) 2A :XC6 'i'xf6 25 l:l.fl l:l.h7! (25 ... '1'g6 26 ltJc8! is decisive) 26 1fxh7+ �xh7 27 l:l.xf6 d3 28 �f2 should win foc White.

lb2) 19 ... .i.g7 20 lLif5 'i'e8 21 .i.h5 'i'd8 22 ltJxg7 �xg7 23 :t7+ �g8 24 h4 l:l.h7 25 'iVg4+ 'iPhB 26 l:l.xh7+ �xh7 27 'i'g6+ Wh8 28 'iVxh6+ �g8 29 .i.f7+! Wxf7 30 l:l.fl+ �g8 31 'i'g6+ Wh8 32 .llf7 is again a win for White.

lb3) 19 ... l:l.h7 ! 20 .i.xd5 (after 20 dxc5 .i.xf3 21 'i'xf3 l:lb8 Black defends) 20 ... exd5 21 'l'g4+ .llg7! (not 2 l ...�h8 22 'ifg6 l:lg7 23 li:Jf7+ �g8 24 ltJxh6+ �h8 25 'l'c6 and

KRAMNIK - ANAND, BEWRADE 1997 231

White is better) 22 'l'h3 :h7 and White seems to have nothing better than to repeat moves.

2) 18 :C6! .i.d5 (the only move as White can't be allowed to take on e6) 19 l:l.g6+ �h7 (19 ... Wf7 1oses to 20 .ih5, while 19 ... .i.g7 20 lt:lf6+ is very good for White) 20 'llt'bl (20 lc2 cxd4!) 20 ... q..xg6 21 lLlf6+ �g7 22 .ih5 'I' xf6 23 exf6+ q..xf6 24 lg6+ q..e7 25 axb5 ! when White should stand better.

18 lLld6 l:tb8 (D)

Vladimir now sank into thought for a long time, apparently trying to decide which of several promising continuations to go for. As it turned out, he spent a long time on many lines. and didn't find anything convincing. Then he saw a move that discourages Black's 'only' resource ( ... c5) and decided to go for it.

19 b4? What on earth is this move? I

hadn't even considered it. After the

surprise faded, I realized that unless Black takes drastic action he is going to be squashed. Incidentally, it is difficult to imagine that this pawn move is going to be the cause of White's defeat!

White should have continued 19 axb5! cxb5 20 lLlxb7 J:l.xb7 21 J:l.xa6 l:tb6 22 l:txb6 lt:lxb6 23 J:l.f6, winning the e6-pawn, after which it's curtains for Black. Ward-Grabliauskas, Copenhagen 1998 tested this assessment and White duly won.

19 .•• hS! If White is allowed to play .i.h5,

Black would hardly have a move left.

20 .i.h3 20 W h4 21 g4 .i.h6 and 20

..i.xh5 11fg5, followed by ... .i.xd6, are fine for Black.

20 ... .i.h6! 21 ..ti>h1 .i.gS

Taking some vital squares from White's fl-rook. He can't play :C4 or l:tf6 any more, so eventually Black will be able to play ... lLlf8. What a relief it was to get one bishop out! Now, if only the other one could escape ...

22 11fc2 Here I saw that he wanted to 'tri

angulate' his queen to e4. I then saw a variation which led to White playing 26 lLlxg5. I didn't see a defence to this and was about to play 22 ... lt:lf8 when a crazy idea occurred to me -was it actually possible to allow his


knight to take one of my pieces on g5 and then ignore it? I spent some time checking the idea and convinced myself that it worked.

After 22 .lb3 c5! 23 dxc5 (the alternatives 23 li:lxb7 cxb4 and 23 bxc5 �d5 are also good for Black) 23.. .�d5 24 axb5 axb5 25 �f5 :g7 ! 26 :a7 :t8 the extra piece should tell eventually.

22 ••. :g7 We both played the next few

moves quickly. 22 ... li:lf8 is also possible, but I wanted to keep f8 free for a rook swap with ... :f8.

23 "ire2 �a8! Not 23 ... :h7? 24 1re4!, when

White is better. 24 "irxbS lU8 25 lDe4

Or 25 axb5 cxb5! and Black consolidates with ... �d5.

25 cSI 26 li:lxgS (D)

26 �dS!!

After 26 ... :xg5? 27 �xe6+ Wg7 28 1Fh4 I couldn't find a way out during the game and afterWards I confirmed that Wbite is winning; e.g., 28 ... cxd4 29 �xd7 :Xfl+ 30 :xn :xe5 3 1 1rxd4.

After the text-move, the bishop finally gets out Unbelievably, Black is already better. Vladimir was shott of time and now missed his last chance.

27 ffi? 27l:xfl!+ ?! li:lxf8 28 Le6+ �

is also excellent for Black, so White should have tried 27 �xe6+! .be6 28 Jhfl!+ fuf8 29li»4 l:l.h7 30 .dl (30 'i1Ve2 cxd4 is similar) 30 ... l:l.xh2+! 31 �xh2 "irh7+ 32 Wgl 'l"xi4. Black has a clear advantage but White is not quite dead.

27 ... cxb4 Capturing the pawn which ad

vanced so audaciously on move 19. White is already lost - the bishop on h3 is dead and the passed pawns on the queenside will decide the issue.

28 axbS axbS 29 �h4 'l'gS

In his hurry, Vladimir had overlooked this and now it's over.

30 Jbf8+ 0..t8 31 'l'e8 l:l.f7

Stopping :n. 32 ffi "irg6!

Not 32 ... '1'e3? 33 ,i.g4!. 33 'l'xbS b3 34 lU1 "ird3 35 ��· 1We3+


36 �bl (D)

36 .•. c3 I considered playing 36 ... l:.b7??,

but smelt a rat and decided to stick with 36 ... c3. In fact 36 .. 11b7?? loses to37 ..txe6+!.

37 ..i.xe6!? I thanked my intuition for choos

ing 36 ... c3 over 36 ... l:.b7 and continued ...

37 38 d5 39 gxf3

..i.xe6 J:.xf3

An unusual situation with seven passed pawns on the board. 39 1lxf3 c2 40 11'fl 11'xf3 41 gxf3 b2 is also winning for Black.

39 •••

40 1Wc4 Either pawn move wins, but why

go for something pretty on move 40?

41 11Vg4+ �b7 42 e6 lQg6

0-1

This game had the same effect on me as the win against Lautier at Biel (Game 41 ), but this time the euphoria affected me positively. I finished joint fust with Ivanchuk at Belgrade, and went on to the FIDE World Championship at Groningen in a really good mood.

There hadn't been a FIDE World Championship for a while, and I regarded this tournament as an important event I was highly motivated for it and I spent the whole of October training. During this time I worked intensively on the Semi-Slav, so my choice of opening in the above game wasn't based entirely on ignorance. After the Belgrade event (in which the previous game was played) there was onJy a two-week gap before the World Championship, not enough for any additional preparation, and I spent this time resting.

Earlier in the year, at Linares, I had lost to Nikolic in a game in which I equalized with Black straight out of the opening. I could have forced a draw immediately,but I became too ambitious and unjustifiably tried to stir up complications. The resulting ending proved worse for me and Nikolicplayed win. to convert his advanta ge into a win. lbis time I deciced to be morecaful.

Game 43

P. N ikolic - V. Anand FIDE World Championship, Groningen 1997

Semi-Slav

1 d4 2 c:4 3 �c3 4 �3 5 i.g5

d5 c6 �6 e6 h6

The so-called Moscow Variation, which makes a change from the enormous complexities of the Botvinnik System (5 ... dxc4).

6 i.xf6 White's traditional reply, gaining

time at the cost of conceding the two bishops. These days there is more interest in the pawn sacrifice 6 .th4 (see Game 42, Kramnik-Anand).

6 'ifxf6 7 e3 �d7 8 i.d3 dxc4 9 .bc4 g6

10 0-0 i.g7 11 b4 0-0 12 :tel 'it'e7 13 'it'b3 (D)

Up to this point, it had all been played many times before. White controls more space and has a lead in development, but Black's position is solid and in the long term the two bishops might pose a danger to White.

13 �b6

B

A slightly unusual move. 13 ... b6 has been played most often, with pretty good results, while 13 ... :ds is another option.

14 i.d3 14 �e4!? is an interesting idea.

After 14 . . . �xc4 (14 ... li)d5 15 b5) 15 'it'xc4 Black has two bishops against two knights, but he cannot easily free himself; e.g., 15 ... b6 16 "Wxc6 i.b7 fails to 17 'it'd6!.

14 l:l.d8 15 � � 16 a3

After 1 6 b5 i.d7 Black is able 10 complete his development.

16 i.d7 17 lL!c5

NJKOUt - ANAND, FIDE WORLD CH., GRONINGEN 1997 235

White usually concentrates on restraining Black and only later goes for his own plan (for example, e4-e5). A typical continuation is 17 1:lc2 ie8 18 llb1 , fortifying the queensidepawns. After 1 8 ... J:Idb8 19 lt:\ed2 Oonov-Andreev, St Petersburg 1998), for example, Black has a solid but passive position.

17 ..ie8 For the moment Black must be

content with quiet play. 18 ..ib1?!

This slip allows Black to take over the initiative on the queenside. 1 8 JUdi was better, intending ..ifl.

18 ..• b6 19 lt:\d3 aS! (D)

1t becomes clear that the bishop is not well placed on b1 . Not only is it blocked in by the d3-knight, which has to stay in place to defend b4, but it also obstructs the first rank, and so prevents White from meeting ... ..if8 by :bl .

20 :C4 White finds another way to de

fend his b4-pawn, but it is clear that the rook is clumsily placed on c4.

20 ••• axb4 21 axb4 1:ldb8!

Black makes use of the opposition of queen and rook to threaten 22 ... c5.

22 e4 22 llfc1 is met by 22 ... c5!, and

now: I) 23 dxc5 ..ia4 and Black wins

the exchange. 2) 23 bxc5 bxc5 24 'ird1 cxd4 25

lt:\xd4 (Black has the two bishops and White's pieces are awkwardly placed, but Black needs to be very accurate - all the pawns are on one side and if White can regroup then it's just a draw) 25 ... e5! and Black keeps the initiative:

2a) 26 lt:\c6 ..ixc6 27 J:lxc6 e4 28 lt:\el (28 lt:\f4 lt:\xf4 29 exf4 e3 30 fxe3 'irxe3+ 31 �hi 'ifxf4 gives Black a safe extra pawn) 28 ... J:Id8 29 'ife2 (29 'ifg4 lt:\c3! 30J:I6xc3 ..ixc3 31 'ifxe4 'iff6 is good for Black) 29 ... 'ifa3 is awkward for White.

2b) 26 lt:\e2 lidS (26 ... ..ib5 27 J:lc5 'ire6 28 e4 is OK for White) 27 e4 lt:\f4 28 lt:\exf4 exf4 and the two bishops give Black an edge.

3) 23 e4! ..ia4 (23 ... lt:\c7?24 bxc5 bxc5 25 'ird I favours White) 24 'ifb2 .ib5 25 exd5 .ixc4 26 J:lxc4 exd5 27 J:lcl 'ird6 28 lt:\de5 c4 29 ..ic2 b5. Black will play ... iH8 and then double on the a-file, so I would say that


it's slightly better for Black, but not more.

22 ••• 0Jc7 The knight is heading to b5, to

step up the pressure against d4. 23 Mel 0Jb5 24 0JdeS ltd8 25 0Jxc6?

White decides to liquidate, but this decision was mistaken since in the ensuing position Black's bishop is far more active than White's. 25 'ifb2! was correct, keeping an eye on b4 and d4. Then the position would be roughly equal.

25 ••• .i.xc6 26 l:.xc6 l:.aJ! (D)

Black makes use of a tactical point to activate his rook with gain of tempo. The alternative 26 ... l0xd4 27 l0xd4 .i.xd4 gives Black a smaller advantage.

27 'ifc4 27 'ifb2? loses immediately to

27 ... l:.xf3 28 gxf3 0Jxd4, while 27

'ifdJ 0Jxd4 28 l0Jtd4 :Xd4 COSIS White the b4-pawn.

27 ••• liJxd4 28 0Jxd4 l:txd4

White is in trouble. His bishop is badly placed on bl, where it serves no other function than to look at the e4-pawn. By contrast, Black's bishop is active and will take part in the attack against White's two main weak spots: b4 and f2.

29 :ell+ �b7 30 'ifc6 (D)

After 30'ifc7 'ilxc7 (stronger than 30 ... 'ilf6 3 1 'ifxb6 or 30 .. .:117 31 'ilxb6 .i.d4 32 'ifb8 11'f6 33 ll8c2) 31 l:.8xc7 .:.Xb4 32 e5 �g8 Black simply plays his bishop to c5, with a large advantage.

30 ••• :d2! Black is not interested in captur·

ing the b4-pawn, which would give White time to mount a counterattack against f7, but goes for bigger game: the f2-pawn and White's king.

NIKOUC - ANAND, FIDE WORLD CH., GRONINGEN 1997 237

31 l:tc7 "i'd8 White's weak back rank costs him

a tempo. 32 g3

There is no good way to counter the back-rank threat. The text-move alows Black's other rook to join the attack, but 32 l:tfl loses to 32 . .l:Xf2! when both 33 'iii>xf2 ld4+ 34 \ii>e l l:te3+ and 33 l:txf2 ldl+ 34 l:tfl �d4+ 35 �hl 1Wxfl# lead to mate. 32 h3 l:tdl + 33 �h2 1e5+ is also hopeless for White.

32 .•• l:tf3 Defending f7 and attacking f2.

33 l:tf1 (D) 33 l:tdxf2

Black has a choice of wins; for example, 33 ... 1Wd4 34 ... c4 l:tfxf2! leads to a quick mate.

B

34 l:txf2

Now Black forces mate in seven. 34 1Vdl+ 35 �gl l:txf2+ 36 �2 �d4+

0-1 In view of 37 'iL'g2 1We2+ 38 �h3

'iffl + 39 �g4 h5+ 40 �h4 �f6#.

It was an indication of my good form at the time that despite a rather drawish position, I was able to concentrate on exploiting my advantages and gained a surprisingly easy win as a reward.

The FIDE World Championship was my first knockout event for quite a while. Sometimes the nervous tension of this type of event can affect the players, and indeed a number of top seeds fell by the wayside at a relatively early stage. Fortunately, I managed to avoid this fate - perhaps I had learned some lessons from Tilburg 1992, where I had been eliminated by Tiviakov early on. In a knockout event, it is important to be aware that the odds are against any particular player winning, so you should not have expectations that are too high. You just have to play each match as it comes, and do your best to maintain your concentration. At Groningen I managed to achieve this, at least in the first half. Some players seem better suited to the tensions of the knockout format than others, and I would single out Khalifman and Adams as being in this category.

11le following game was my best achievement at Groningen, and from it You can see that I was still in peak form.

Game 44

V. Anand - A. Shirov FIDE World Championship, Groningen 1997

Ruy Lopez, M0ller

1 e4 eS 2 lff3 lbc6 3 .i.b5 a6 4 .i.a4 lff6 5 0-0 b5 6 .i.b3 .i.c5 7 a4 .:tb8 8 c3 d6 9 d4 .i.b6

10 lba3 0-0 1 1 axb5 axb5 12 lbxb5 .i.g4

For the alternative 12 ... exd4, see Game 49 (Anand-011).

13 .b3 Later, it became clear that Black's

previous move is an inaccuracy that can best be exploited by 13 l:z.el !.

13 exd4 14 cxd4 (DJ 14 'l'e8

Here, too, theory has advanced since this game was played. The improvement 14 ... lbxe4! was played in Nijboer-Piket, Wijk aan Zee 1998. This is based on the tactical point 15 .i.d5 'l'e8 16 'l'c2 (Nijboer-Piket continued 16 h3 .i.f5 17 l:z.el lbb4 18 .i.c4 c6 19 lba3 and although White went on to win the game, here Black is more comfortable) 16 ... .i.xf3 17

.i.xc6 (17 gxf3 lbb4 18 'l'xe4 ._xb5 favours Black) 17 ... '1'e6and now 18 gxf3 fails to 1 8 . . . 'l'g6+.

15 h3! 15 lbc3 lbxe4 16 h3 .i.xf3 17

'l'xf3 lbxd4 18 .i.xd4 li:ld2 19 'lf4 lbxb3 and White has nothing better than to force a draw by 20 .i.xg7.

15 ••• .i.d7 There is nothing better. 15 ... ih5

16 .i.a4 "ifxe4 17 lbc3 .i.xf3 18 cfue4 .i.xd I 19 lbxf6+ gxf6 20 bc6 favours White owing to Black's weak pawns and the inactive b6-bishop.lf instead 15 ... .i.xf3?! 16 'lWxf3 lZlxe4. then 17 .i.d5 lbe 7 18 lbxc 7 .i.xc 7 19 .i.xe4 l:Z.xb2 20 .i.g5 and Black's position looks quite unpleasant.

ANAND - SHIROV, FIDE WORLD CH., GRONINGEN 1997 239

16 llk3 Black must act now or remain a

pawn down. 16 .••

lbxe4 17 :e1 lbxc3

!7 ... i.f5 loses material to 1 8 i.a4, while after 17 ... lba5 1 8 i.c2 lbxc3 19 bxc3 'i'c8 White plays 20 c4! in any case, since Black cannot take on c4.

18 bxc3 (D)

Black has regained the sacrificed pawn, but he cannot equalize. White controls more space and Black has problems with the b6-bishop, which is exposed to attack by c4-c5. Whilst the poorly placed bishop might appear relatively insignificant, it turns out to be a crucial factor in the game.

18 ... 'i'c8 19 c4! i.t'S

Activating the bishop and avoiding the tactical point mentioned in the next note.

20 :e2

Thanks to Black's ... i.f5, White cannot hunt the bishop down by 20 cS i.a5 21 d5 since after 2l . ..i.xel 22 dxc6 there is no en prise bishop on d7. After the text-move, however, this is a genuine threat.

20 .•• llJaS Black takes drastic action to res

cue the b6-bishop, but the result is two misplaced minor pieces. h would probably have been better to play 20 ... :d8 21 i.a4 lbb4 (2l ... lba5 22 :c I is slightly better for White) 22 i.g5 (22 .:a3, followed by doubling on thee-file, is a saferoute to a slight advantage) 22 ... f6 23 i.d2 c5 and now:

I ) 24 :a3 is an interesting idea which leads to some nice variations, but it falls short:

Ia) 24 ... i.d3 25 .:e7 hc4 (not 25 .. .'�ffi? 26 hb4 and now 26 ... cxb4 27 'i'xd3 �xe7 28 'i'xh7 wins for White, while 26 ... hc4 27 :ae3 cxb4 28 'i'bl ! gives White a dangerous attack) 26 i.xb4 cxb4 27 b3 'i'a6 (27 ... :b7 loses to 28 :e8+ �7 29 lbe5+ fxeS 30 'i'f3+) 28 liJh4 is good for White.

1 b) 24 ... d5! looks fine for Black. 2) 24 i.xb4 cxb4 25 :b2 is a bet

ter idea, with a slight advantage for White.

21 h2 cS 22 dS! (D)

Now both the b6Cbishop and the aS-knight are poorly placed. The bishop might eventually emerge via


d8 and f6 although, as the sequel shows, achieving this is no easy task. However, the plight of the knight is more or less permanent. Even if it manages to grovel back to b7 and d8, it would still have little future. King's Indian players will be only too familiar with the problems posed by an offside knight on a5!

22 ... i.d8 23 i.d2

Preventing ... i.f6 and thus giving White time to switch his bishop to c3.

23 ••• 11'a6 24 'li'a4

White cannot play the immediate 24 i.c3 due to 24 ... lt:lxc4 25 i.bJ (25 i.bl 'i'xal !) 25 ... 'i'b5, but the queen move pins the dB-bishop down for another move.

24 ... :as Not 24 ... i.d3 25 l:l.e3 i.xc4? 26

i.xc4 'ifxc4 27 i.xa5 l:l.b l+ 28 i.el and White wins a piece.

25 .i.cJ

White has definitely prevented i.f6 and has a clear advantages.

25 - li'lb7 The knight drops back, not so much

in the hope of achieving any activity itself, but more to let the d8-bisbop out via a5.

26 'li'dl! White keeps queens on, as the lack

of enemy pieces on the kingside wiD give White good attacking chances there.

26 - .taS?! The bishop emerges, but it has no

effective role on the a5-el diagonal. 26 ... 11b6 was a better chance; although White retains a clear advantage, at least the dB-bishop controls g5 and h4.

27 i.b2 .i.b4? (D) Black should hunker down wilb

27 ... l:l.ae8, although after 28 .ib3 followed by i.a4 his position is unenviable.

lS �4!

ANAND - SHIROV, FIDE WORLD CH., GRONINGEN 1997 241

Now White's advantage reaches decisive proportions.

28 �g6 The alternatives also lose: 28 ... J..d7

29 l:le7! wins material, as 29 ... .:.ad8 fails to 30 �bl 'lWb6 31 �xh7+ �xh7 32 "lb5+ �g8 33 .!i::lg6 and mate next move, while 28 . . . J..c8 29 'ird3! gives White a ferocious kingside attack

29 C4 (D)

To save the bishop, Black will have to weaken his kingside.

29 ... 'ira4 After 29 .. .f6 30 .!i::lxg6 hxg6 3 1

'l'd3 'i>f7 32 �bl White wins at once.

30 'irxa4?!

There is nothing seriously wrong with this move, because White stili liquidates to a winning ending. However, he could have decided matters by a direct attack: 30 �b3 ! 'ird7 31 :as l:lxa8 32 f5 ! ..txf5 33 i.a4! :xa4 34 .!i::lxf5 l:la8 35 .!i::lxg7, with

.!i::lh5 to come, leaves Black with no defence.

30 ••• l:ba4 31 rs :CaS

3 L.i.h5 32 g4 wins for White. 32 :e7!

Not 32 fxg6? hxg6 33 .!i::lf3 (after 33 J::l.e7 J::l.xa2 White no longer has a back-rank mate -this is why he must delay capturing on g6), which prevents 33 ... .:.Xa2 owing to 34 .:.Xa2 J::l.xa2 35 J::l.e8+ �h7 36 .!i::lg5+ fol-lowed by mate, but after 33 ... f6! White is suddenly unable to save the a2-bishop.

32 ... .i.b5 (D) 1be only chance, as after 32 ... .!i::las

33 fxg6 White wins a piece in perfect safety.

33 g4! Just as before, White must not

grab material too soon. 33 :xb7? f6 leaves White with the familiar problem regarding his a2-bishop.

33 f6


Black's reply is again forced, since otherwise he drops a piece.

34 gxh5 w Now, however, this liquidation

leads to a forced win.

w

34 :Xa2 35 bl bl (D)

36 h6! White's b2-bishop is doomed, but

before it finally dies, it helps seal Black's fate.

36 .•. :Xbl 36 ... gxh6 37 .i..xf6 ll:ld8 38 ll:lg6!

leads to a quick mate. 37 :Xg7+ �h8 38 :Xb7

Not only is Black a pawn down, but his king is also in terrible trouble.

38 .i..cl 39 l1d7 �g8 40 l1d8+

It is not the d6-pawn that White wants, but the one on h7.

40 'M1 (D)

With the time-control passed, I took time to get up from the board and relax, just as the old Soviet trainers used to advise. This turned out to be time well spent, because it enabled to me to find an accurate move and avoid a potential pitfall in the position.

41 l:l.h8! 41 ll:lg6? is flashy, but would al

low Black unnecessary chances after 4I .....Ild4+ 42�hl (or42 'it>fl llf2+, followed by ... llxf5) 42 ... llbl+ 43 �g2 llb2+ 44 �f3 l1f2+ 45 �g4 llg2+ 46 �4 J:f2+ 47 'it>e4 hxg6 48 h7 gxf5+ 49 �d3 Iif3+ 50 �d2 .:r.xh3, when matters are far from clear.

41 ••• .i..d4+ 42 �

Black's situation is hopeless. As soon as the checks run out, White wins by l:Z.xh7+ followed by li:lg6 (or l:Z.g7+ first, if Black meets l1xh7+ by ... �g8).

1-0

ANAND - SHIROV, FIDE WORW CH., GRONINGEN 1997 243

Sometimes you get topsy-turvy games which go first one way and then another, but here White's play creates a seamless impression from beginning to end - one could almost imagine Capablanca playing a game like this. Had I found 30 .i.b3, then I think it would have been a virtually flawless game. However, even this inaccuracy is not a serious defect because White was also winning after the game continuation, although it required a second bout of accurate play to wrap up the full point.

After the Shirov match, tiredness started to set in. Against Gelfand, I was still playing well, but against Adams you could see that exhaustion was affecting both players. Then we come 1D the match against Karpov in the final. The following game was played when I was 3-2 down, with one game still to play. I can't say that it is a particularly good game, but it shows something of my character that I was still able to pull off a win. At one time I probably couldn't have achieved this, but over the years I have become tougher, especially in critical situations. The difference between a good performance and an extra-special one is often not to be found in the technical aspects of the game, but in sporting characteristics such as will-power and resilience under pressure.

Game 45

V. Anand - A. Karpov FIDE World Championship Final (6),

Lausanne 1 998 Trom powsky

I needed to win this game to stay in the match. What should I do? I recalled my previous match against Karpov at Brussels 1991. Trailing by one point after four games, I played calm chess and still got winning positions in both the fifth and sixth games. I also derived some hope from the 24th game between Kasparov and Karpov in Seville 1987 when Karpov was in the same situation. Kasparov avoided a theoretical battle and got no advantage. Karpov, however, was too eager to draw and soon got himself into hot water. With these thoughts in mind, I decided to play

1 d4 li)f6 2 �gS

the Trompowsky! I couldn't really bring myself to play something like Kasparov's Reti, so this seemed a good compromise. On the one hand, it offers interesting positions. Since the theory consists mostly of Hodgson games, there are no long forced lines leading to draws (yet!), but lots of fascinating and creative chess. Perfect, I thought, for this critical game.

2 e6 3 e4 h6 4 .bf6 'l'xf6 s li)c3 d6 6 'ilt'dl gS

Black wants to prevent White from expanding with f4. Since Black is going to castle queenside, the weakening of the kingside pawn-structure is not especially serious.

7 �c4!? My team was looking for some

thing off the beaten track, but with some venom. Eventually �supov came up with this idea. White is g� ing to castle kingside, as opposed to the normal queenside castling in this variation. The g 1-knight should go to e2 (on f3, it would be harassed by ... g4) and so the idea arose to develop the bishop first. The knight also has ideas like li)ge2-g3-h5.

7 ll)c6 8 li)ge2 � 9 :d1 �d7

10 0-0 0-0-0 (D} 10 ... 0-0 1 1 'l'e3 ( 1 1 f4 gxf4 12

:xf4 'l'g5 is slightly awkward for White) offers White chances based on f4 or li)g3-h5 .

ANAND - KARPOV, FIDE WORLD CH., LAUSANNE 1998 245

11 lllbS! Black will ignore any queenside

thrust like b4-b5 by simply playing his knight to a5, so White must first provoke weaknesses and only then prepare b4. For example, the immediate I I b4 g4 12 b5 llla5 13 ..i.d3 h5 (13...�b8 is also possible) 14 liJd5 exd5 15 •xa5 �b8 ! (safer than 15 ... dxe4 1 6 ..i.xe4) 16 exd5 ( 16 lllc3 'lxd4 17 ll\xd5 .. c5 gets White nowhere) 16 ... h4 gives Black enough compensation for the pawn.

11 ... a6 12 liJa3

Intending to set a queenside attack in motion by c3 and b4.

12 g4 Subsequently, I found some games

where Black went for .. .f5 immediately. Indeed, 12 ..... g6 13 f3 f5 appears more effective than the Karpov chose.

13 r4 13 b4?! .. g5! is fine for Black.

13 gxr3

14 l:!.xf3 _.e7 Black's two bishops and possible

g-file play balance the weak f-pawn. 15 cJ h5 16 .l:ldfl lldf8 17 b4 �1 (D)

This is not a bad move, but it perhaps shows an over-developed sense of danger.

18 lllc2 19 _.e1 20 ..i.d3

..i.h6 �b8 ..i.c6!

Black has played this phase well; his two bishops and the open g-file give him counterplay.

21 lllr4 J:fg8 After 21 ... e5 White can bravely

take the pawn with 22 lllxh5!. 22 d5

Now 22 lllxh5 f5 ! gives Black too pbmch play.

22 ••• ..i.e8 (D) After 22 ... ..i.xf4 23 dxc6! (23 .lbf4

exd5 24 l:!.xf7 _.g5 is excellent for Black) 23 ... ..i.xh2+ (23 ... e5 24 cxb7


breaks open the defences to Black's king) 24 �h2 ll:\xc6 25 llh3 (not 25 l:lxn 'if g5! 26 'ife2 lL\e5 T7 ll7f3 ll:lxf3+ 28 llxf3 1We5+ 29 �hl h4, which is slightly better for Black) 25 ... ll:\e5 26 i.e2 "flg5 27 ll:le3 h4 (after 27 ... ll:\g4+ 28 i.xg4 hxg4 29 llxh8 llxh8+ 30 �gl White keeps a slight edge) 28 "flf2 White can slowly consolidate.

23 'iff2 If White grabs the pawn by 23

ll:\xh5 Black can choose between: l ) 23 ... 'irg5 gives Black enough

compensation in the case of 24 ll:\g3 i.g7, but 24 lLif4 is better, since the knight is well-placed on f4. Then 24 ... e5 doesn't work due to 25 .l:l.g3, so White should have the advantage.

2) 23 .. .f5 !? 24 lL\f4 and now: 2a) 24 ... e5 25 ll:\e6 f4 26 c4!

(White is perfectly willing to give up the exchange on f3; his knight on e6 would be secure and he could play on the queenside) 26 ... 'ifh7 (26 ... i.h5

27 b5 and 26 ... .td7 27 b5 axb5 28 cxb5 ll:lxb5 29 'iVb4 are also promising for White) 27 b5 and White's attack develops quite fast.

2b) 24 ... .txf4! 25 :txf4 fxe4! (25 ... 1Wh7 forces 26 h4, but I can't see a follow-up for Black) 26 .be4 (after 26 11fxe4 i.g6 Black wins a piece) 26 ... exd5 27 i.f5 (27 .bd5 'irh7 28 h3 'Mhc2 29 hg8 Jhg8 30 lllf2 'irg6 31 .l:l.f8 lL!c8 and Black consolidates) 27 .. .'1/g7 28 .l:l.g4 i'h6 looks fine for Black.

23 ... .tgr Or 23 ... h4 24 ll:\d4 bf4! (after

24 ... i.d7, 25 ll:\h5 favours White as his knight is heading for f6) 25lW4 and White has a slight advantage after 25 ... ll:\c8 26 c4 oc 25 ... l:g7 26 'iff3 llhg8 27 llf2.

24 ll:\d4 (D)

24 .td7 The first slight slip. Black should

not have missed the oppoltunity to swap the a7-knight by 24 ... 1i:\b5.

ANAND - KARPOV, FIDE WORW CH., LAUSANNE 1998 247

Then 25 dxe6 li:lxd4 26 cxd4 fxe6 is OK for Black.

25 dxe6 .bd4 26 cxd4 fxe6 1:1 eS .i.c6 28 lilg6 (D)

28 l:tg3 is met by 28 ... h4.

28 'l'd8?? As soon as he made this move, I

saw a grimace on his face. It turned out that he hadn't seen 29 ... lilc8 in the line below and so went for the alternative, but saw 30 lilf7 too late.

He should have played 28 ... l:l.xg6 29 .bg6 (29 l:l.f8+? lilc8 ! 30 .i.xg6 l:l.xf8 31 'l'xf8 1l'g5 wins for Black) 29 . ... bf3 30 11Fxf3 dxe5 3 1 dxe5

11'xb4 (31...lilc6 32 11Ff6 gives White an edge) 32 11Ff6, with an unclear position.

29 lilxh8 Suddenly I was back in the match.

I tried my best to calm down, as I realized that 3-3 was now a matter of technique.

29 ••• .i.xf'3 30 lilf7 'l'h4

After some thought he came up with this move, which is the best practical chance, but really White only has to play carefully to rake in the point. 30 ... .i.xg2 3I lilxd8 .i.e4+ 32 11Fg3 l:l.xg3+ 33 hxg3 .i.xd3 34 ltf8 ! wins for White.

31 11Fxf'3 Not 3 1 11Fxh4?? ltxg2+ 32 �hl

ltf2+ and Black draws. 31 ••• 11Fxd4+ 32 �h1 d5

H Black could save his h-pawn, matters could still get complicated due to my stranded f7-knight.

33 .l:l.dl! (D)

8

The b-pawn is not important; what is crucial is to capture the h5-pawn.

33 ••• 'l'xb4 34 .l:l.b1 'l'a4 35 'l'xhS lilc6 36 'l'e2 �a7 37 'l'rl+ b6


Intending ... �b4 with some slight counterplay.

38 :ci! �b7 w 39 h3!

White has made all the necessary precautionary moves and can now go for the e6-pawn.

39 .••

40 'Wf6 41 lbcl8+!

J:l.c8 lbcl4 (D)

I saw that 41 �6+ won as well (after 41 ...cxd6 42 'ire7+ �bS 43 :xeS+ �xeS 44 exd6 'i'c6 45 'i'f8+ �d7 46 'i'g7+ �xd6 47 'i'xd4), but Karpov has swindled far too many points from me over the years, so I looked for something less complil:ated.

41 �b8

42 �xe6 1-0 Here he resigned, since 42 ... 'i'al

43 :dl is hopeless for Black. I had drawn the match! The game itself was not of the highest quality, but I was extremely proud to have pulled it off in a must-win situation.

After this game I botched the tie-break, but that is history now. I regard Groningen as a great success and in a way consider that I 'won' the tournament, since the conditions for the final were so unequal that I can hardly count it as part of the event.

When Topalov came on to the scene (he started playing at the highest level round about 1 993), many players, myself included, found it difficult to cope with his style. His unadulterated aggression, backed up by absolutely superb preparation, proved a handful for almost everybody. The very top players meet each other across the chessboard quite frequently, so that the style of each player gradually becomes familiar and you know more or less what 1D expect. Topalov was a 'new kid on the block', with a new style which was f� a time very successful. However, the impact of his excellent preparation was obvious and it stimulated the leading players to raise the level of their owo preparation in order to combat him. I was pleased with the following game because Topalov had been a very difficult opponent for me. By winning this I came back to a level score with him. Although the excitement all takes place within the space of a few moves, it is nevertheless an attractive game.

Game 46

V. Anand - V. Topalov Wijk aan lee 1 998

Ruy Lopez, M11111er

1 e4 eS 2 �3 ltlc6 3 .tbS a6 4 .ta4 lLlf6 5 0-0 bS 6 .tb3 .tcS

At the time of this game, Topalov and Shirov had already been playing the MBI.Ier Variation for two years and had been responsible for its growth in popularity. As with all topical lines, the theory developed rapidly. I beat Shirov in Groningen 1997, but in the bird round at Wijk aan Zee 1998, Piket had shown an improvement over Shirov's play. This game was played in round five, when the ball was back in White's court. I was attracted to a plan that Topalov used against Shirov at Madrid 1997, and decided to give it a shot.

7 a4 l:l.b8 8 c3 d6 9 d4 .tb6

10 axbS The main line is 10 ltla3, but for a

shon time in 1998 the continuation 10 axb5 axb5 I I h3 held centte stage.

10 axbS 11 b3 0-0 12 l:l.e1 (D)

12 l:l.e8!? Topalov prefers to leave the bishop

on c8 and pressurize e4. The alternative is 12 ... .tb7 13 ltla3 exd4 14 cxd4 ltla5 15 .tc2 b4 1 6 ltlbl c5 17 .tg5 h6 1 8 ..i.h4, and now:

I ) 1 8 ... g5? 1 9 ltlxg5 hxg5 20 ..i.xg5 b3 and then:

Ia) 21 ..i.d3 c4 was played in Topalov-Shirov, Madrid 1997. I feel that White should be better even here, but this is a moot point since line ' lb' looks so strong.

lb) 21 .txb3 ! is very good. It is surprising that White should go after a measly pawn, but in fact it's the third rank he wants. After 2 l ...cxd4 22 l:l.a3 ! White was clearly better in


the game Hraeek-Shirov, Bundesliga 1997/8.

2) 1 8 ... cxd4! is an improvement. B I faced this in the last round at Wijk aan Zee (against Shirov) and didn't really get any advantage. However, it's time to leave theory and go back to the game!

13 llJa3 b4! The next few moves all appear

forced. 14 llk4

Better than 14 dxe5 bxa3! (the alternative 14 ... dxe5 15 lDc4 is slightly better for White) 15 exf6 axb2 and now White has to allow easy equality after 16 .ixb2 1Wxf6, because 1 6 .ixf7+ �xf7 17 Wd5+ .ie6 1 8 tt:lg5+ �g8 1 9 f7+ ( 1 9 tt:lxe6 Wxf6) fails to 19 ... �h8! ! 20 tt:lxe6 bxal"ii' 21 tt:lxd8 l:texd8 and Black has turned the tables.

14 bxc3 15 bxc3 exd4 16 tt:lxb6

White must eliminate this bishop because 16 cxd4? may be answered by 16 ... tt:lxd4!.

16 •••

17 cxd4 l:txb6 tt:lxe4

Or else White has the two bishops combined with an imposing centre. Not 17 ... l:txe4? 1 8 J.xf7+ ! , when White is clearly better.

18 .ixf7+ �xf7 19 l:txe4 l:txe4 20 tt:lgS+ �g8 21 tt:lxe4 (D)

Black has simplified the position and even forced bishops of opposite colours. Nevertheless, he still has to be careful. His pieces is misplaced - the rook on b6 and even the knight on c6 are vulnerable in some lines. Black's king is slightly exposed, not only because it has just two defensive pawns, but because of vague threats of l:ta8, etc.

21 ... ids 22 tt:lgS! tt:le7

Black's first slip. 22 ... h6! is a more accurate defence,although after 23 Wf3 Black must continue precisely:

1) 23 . . . tt:le7? 24 l:ta8 l:tb8 (or 24 ... .1c8 25 1if7+ �h8 26 Wxe7!) 25 Wb3+! and White wins.

2) 23...Wc8 24 Wd5+ �h8 2S tt:lf7+ �h7 26 tt:lxh6! tt:le7 (Black is lost after 26 ... gxh6 27 1Wf7+ �h8 28 l:ta3) 27 11t"a2! (more accurate than 27 Wf7 11t"e6) 27 ... .ie6 28 'lc2+ tt:lg6 29 .l:la7! and White has a clear advantage.

ANAND - TOPALOV, WIJK AAN ZEE 1998 251

3) 23 ... 1ff6! 24 'l'd5+ � (not 24 ... �h8? 25 W'xc6!) and Black appears to be surviving.

23 g4! i.g6 23 ... i.d7 fails to 24 'Wf3 !.

24 &6 (D)

24 1fc8? After the game, we concluded that

22.. .i0e7 was a big mistake, but in fact it is only 24 . . . 1fc8? that proves fatai.At home I found the defence 24 ... 'i'b8, trying to stop l:ta 7, and have been unable to find anything really convincing against it. Maybe White is still better, but he doesn' t seem able to prove a serious advantage. The analysisruns 25 d5 (25 lf3 d5) 25 ... ..tf7! (the difference is that Black has ... l:tb1 in many lines; 25 ... c5 is bad in view of26 i.g5) and now:

I) 26 li:Jxg7 �xg7 27 'Wd4+ � 28 ih6+ (28 'Wh8+ lLlg8 29 i.h6+ �e8 30 l:tel+ �d7 favours Black) 28 �e8 29 l:tel l:tb1 ! 30 i.e 1 l:txc1

(not 30 ... �d7? 31 1We4! llJxd5 32 1Wxh7 and White is better) 31 :Xc1 � (3I. ..i.xd5 32 D.e1 'iltb3 33 1Wh8+ �d7 34 1fxh7 is unclear) 32 1rh8+ i.g8 33 1Wf6+ i.f7 with a draw by perpetual check.

2) 26 i.g5 llJg6 and Black can defend after 27 lLldB ll:le5, 27 i.e3 llb5 or 27 1Vd4 llJe5.

3) 26 1i'd4 he6 27 dxe6 D.bl 28 llxbl 'ii'xbl 29 'ii'f4 h6 (29 ... 'ii'g6 30 'ii'a4 gives White an edge) 30 1ff7+ �h7 3 1 'ii'xe7 1fxcl+ 32 �g2 'ii'c6+ 33 �h2 1Vd5! 34 1Vd7 (34 1Vf7 c5 and the pin stops White from advancing his passed pawn) 34 ... 'ii'e5+ and it is doubtful if White has any more than a draw.

2S dS! I don't see a defence for Black af-

ter this. 2s ••. ..tn

Or 25 ... c5 26 lla7! llb7 27 llxb7 1fxb7 28 'ii'f3! 1Vc8 (28 ... i.f7 29 1Vxf7+! and 28 . .. �h8 29 i.h6! are also hopeless) 29 11Vc3 and White wins.

26 ll:lxg7! �xg7 There isn't really much else, as

26 ... i.xd5 and 26 ... ll:lxd5 both fail to 27 ll:lf5.

27 11Vd4+ � 27 ... �g8 loses to 28 i.h6, but

27 ... �g6 is the most tricky line. White wins after 28 11'e4+ �f6 (28 ... ll:lf5 29 lla3! is decisive) 29 i.h6! (this was my intention during the game; at the time I thought that 29 g5+


�g7 ?n 'ill'xe7 'ill'f5 3 1 ll:a8! 'iii'g6 might not be so easy, but later Fritz showed that this wins too after 32 ..,f8! 'ill'xh3 33 lta3! - still, this isn't the easiest line around) and now:

1) 29 ... .i.xd5 30 'i'xh7 lDf5 3 1 ll:el fuh6 3 2 f4 lDf7 3 3 g5+ lDxg5 34 ..,e7+ is a nice finish.

2) 29 ... .i.g6 30 'ill'd4+ 'J;f7 31 ..,g7+ �e8 32 .i.g5 'i'd8 33 lta8 wins for White.

3) 29 ... lDf5 30 gxf5 'i'xf5 (or 30 ... 'ill'g8+ 31 �h2 .i.xd5 32 ..,h4+ 'iii>xf5 33 'ill'h5+ �6 34 ltgl and White wins) 31 'ill'h4+ 'J;g6 (3l ...�e5 32 ltd l ! .i.xd5 33 ire7+ is no improvement for Black) 32 �h2 is winning for White.

28 .i.h6+ �e8

29 llel 1-0 The position is hopeless f<r Black

for example, 29 ... 'J;d7 30 :Xe7+! �e7 3 1 'i1Fe4+ .i.e6 (31...�6 32 g5#; 3I ...'J;d7 32 'i1Ff5+) 32 1Wxh7+ .i.f7 (32...�e8 33 'i1Fh8+ �e7 34 .i.g5+ 'J;f7 35 irh7+ � 36 'l'e7+ 'J;g8 37 .i.h6 'i'd7 38 11ff8+ �h7 39 dxe6 wins for White) and now either check wins:

1) 33 .i.g5+ 'J;e8 (33 ... � 34 irh8+ .i.g8 35 'ill'h6+ �e8 36 'l'g6+ .i.f7 37 'ii'e4+ transposes) 34 'l'e4+ �f8 35 .i.h6+ �g8 36 'i'd4 wins.

2) 33 'i1Fe4+ .i.e6 34 .i.g5+ � 35 dxe6 'i'e8 36 .i.h6+ �g8 (or 36 ... 'J;e7 37 'i'h7+ �d8 38 'l"f7) 37 'i1Ff5 and Black has absolutely no defence.

I was extremely pleased to win a game in direct attacking style, and as are· suit I shared the lead with Kramnik. The tournament ended with the situation unchanged, the two of us sharing first prize.

During January, I tried to withdraw from the forthcoming Linares tourna· ment on the grounds of tiredness, but the organizer persuaded me that it would be a severe blow to the tournament, and I participated after all My play in the first part of Linares was very wobbly; I was lucky to win my first two games and I felt that I was swimming against the tide. Fortunately, I had a double rest day before the last three rounds, and I was able to go out with some friends and distance myself from chess for a time. Thanks to this, I felt much fresher for the final sprint. The first of the three final games was the fol· lowing game against Ivanchuk. 'Chucky' had started the tournament badly. but then he won two consecutive games against Shirov and Svidler so at the time of this game he was back on a respectable score.

Game 47

V. lvanchuk - V. Anand Linares 1 998

Sici l ian, Richter-Rauzer

1 e4 c5 2 lllf3 d6 3 d4 c:xd4 4 lllxd4 .U6 S lllc3 lL!c6

At the start of the event I played three Caro-Kanns, but switched to the Oassical Sicilian after a tough game against Topalov.

6 �gS e6 7 'l'd2 �e7 8 0-0-0 0-0 9 f3

This came as a surprise to me since I couldn't remember any games where Ivanchuk had played it. Still, Khalifman had played something similar against me at Groningen 1997, 1ll I wasn't caught entirely unawares.

9 lL!xd4 10 'l'xd4 a6 11 h4 (D)

An interesting move, since White rules out any tricks based on the undefended g5-bishop. However,it seems that White's plan of h5 and

then g4-g5 is a bit slow. Note that 1 1 .ixf6?! hf6 12 'l'xd6 'l'a5 gives Black good play for the pawn.

11 bS

12 Wb1 'irc7 13 hS

Now, however, 13 �xf6 could have been considered, although both recaptures are playable:

1) 13 ... ..i.xf6 14 'irxd6 l:a7! (after 14 ... 'ira5? 15 e5 l:d8 16 'irc6! White wins material) 15 'irxc7 l:xc7 16 lLle2 �b7 and I think that Black has enough compensation.

2) 13 ... gxf6!? and Black can continue with ... 111' c5, foUowed by .. .'�h8 and ... l:g8.

13 ••. h6 14 �h4

Here 14 ..i.xf6 gxf6, followed by ... 'ir c5, is fine for Black as the move h5 doesn't help White at all.

254 VtSHY ANAND: MY BEST GAMES OF CHESS

14 ••• �b7! 15 li:le2

Continuing with the attack by 15 g4 runs into 15 .. . b4 16 li:le2 ( 16 �xf6 �xf6 17 Wxb4 J:Lab8 ! gives Black a very strong attack) 16 ... l:Z.ac8! 17l:lcl ( 17 1Vd2 0.xe4! 18 fxe4 he4 19 .i.xe7 Wxe7 is very good for Black) 1 7 ... e5 18 1Vxb4 d5 with a massive initiative for Black.

15 ••• :Sc8 16 Wd2 Md8 (D)

Not 16 ... d5? 17 e5! 1Vxe5 1 8 �g3 'l'f5 19 �4 1Wh7 20 �d3 1Vh8, which is amusing, but only for White! After the text-move Black has equalized.

17 .l:l.e1?! Too slow. White is still unable to

push forward on the kingside since 17 g4?! is met by 17 ... lllxe4!, but l7 llld4! was possible, when 17 ... d5 18 e5 'ii"xe5 19 g4 �c5 20 c3 b4! leads to a complicated position in which White is still in the game.

17 .•• e5! Preventing White's e5advance

once and for all. 18 .ixf6

The alternatives are: 1) 18 li.'lc3 d5 19 exd5 (19 L£6

dxe4 is also good for Black) 19 ... b4 ( 19 ... li.'lxd5 20 lllxd5 .i.xd5 21 i.d3 gives Black only a slight advantage) 20 llle4 �xd5 (20 ... �xd5 2l lZlxf6+ ..txf6 22 �d3 .i.xh4 23 l:xh4 is again just a little better for Black) 21 'ii"f2 a5 and Black has a clear advantage.

2) 18 lllg3 d5 19 .i.d3 dxe4 20 fxe4 lt:lh7 21 �xe7 'ii"xe7is promising for Black, since after ... lbg5 he will have the only active playin the position.

18 19 li:lc3

.lhf(j .igS! (D)

This is the problem with White's plan. Black doesn'thave to play •. b4 or ... d5 straight away,but can first force White to misplace his queen.

IVANCHUK - ANAND, LINARES 1998 255

20 'i'dl?

'i'aS Now Black's attack becomes irre

sistible. 21 lLJdS

There is no defence: 21 116'd3 b4 2210d5 b3! wins for Black, while 2 1 id3 :xc3 22 bxc3 .ic8!, followed by ... ie6, gives Black a decisive attack.

21 22 exdS

23 �xc2

i.xd5 llxc2 (D)

'i'xa2 Simple - White can't do anything

to stop 24 ... llc8+. 24 f4 25 �d2 26 �e2

27 �f3

.l:c8+ .hr4+ 'i'xb2+ .l:cl

There are probably other ways to win, but this move just forces resig-nation.

0-1

In the next round I beat Topalov, which boosted me into a clear lead. A draw in the last round against Kramnik preserved this lead and took me to outright first prize, ahead of Shirov, Kasparov and Kramnik.

1998 was my most successful year up to that time. I won the chess Oscar, and took first place in five major tournaments. My only failure was in Dortmund,where i scored just under 50%. Sometimes playing chess is an uphill struggle as in Linares, but the rest of the year I felt quite fresh. The tournaments were nicely spaced out, and I had time between for preparation. Dur-ing October, before the Tilburg tournament, I went to a Center Pare in Holland for a training camp. For those who are not familiar with Center


Pares, they are a chain of holiday villages in forest settings with sport facilities and water attractions. While I was there I worked with my regular second, Ubilava, and for a time with Peter Leko. At the time I was quite keen on going away for a training camp, although since then I have tended to work more at home.

At the Frankfurt rapid tournament in June, the final was between myself and Krarnnik. Although I won the match in a blitz game, I had not been able to break through his Petroff Defence. Thckling the Petroff was one of the tasks we set ourselves while at the Center Pare. The resulting work proved its value just one month later.

Game 48

V. Anand - V. Kramnik Tilburg 1 998 Petroff Defence

In our seven previous encounters where I had the white pieces, I had managed only draws (when Vladimir had the white pieces, we exchanged two wins apiece). I had no idea, of course, that this one was going to be any different, but at least this time I had a great idea in the Petroff, the opening Vladimir had made his principal weapon following his match against Shirov.

1 e4 eS I was quite happy to see this ...

2 m l'i:lr6 ... and even happier now. I played

the next few moves quite quickly. 3 lilll:eS d6 4 m ltlxe4 S d4 dS 6 W ltlc6 7 0-0 i.e7 8 l:te1

In genera� the lines with c4 work better when it is played at this .Point, so this was the first hmt that I didn't iJtm;l Ill play c4 today.

8 ·- i.g4 !I c3 rs

Black gives up several squares on thee-file, but gets a big knight on e4 and chances for an attack.

10 1Vb3 0-0 11 tLlbd2 ltlaS

l l ... �h8 is the other critical line, and Shirov played itagainstlvanchuk in Dortmund 1998. However, I expected 1 1 ... � as this line has a good reputation, theoretically speaking.

12 1Va4 lilc6 13 .i.bS lilxd2

Subs�uently, Kramnik improved with !3 ... i.h4 14 g3 i.f6 and ifWhite now takes the pawn with I 5 i.xc6 bxc6 16 1Vxc6, then after 16 ... l:te8 Black has compensation. Theory marches on. Back to the game!

14 ll:lxd2 'ild6 (D)

Thus far we had followed my game against Anur Yusupov from


Linares 1993. In that game I had the feeling that White must be better as a result of having got the dreaded B

e4-knight out ofthe way, but I soon discovered that the black attack packed sufficient punch even with-out this knight. Luckily, I was able to maintain the balance and make a draw.

15 h3!! My second, Ubilava, suggested

interposing this before continuing .!tlb3-c5. After some time, I realized that Black's attack loses a lot of its force as a result of this interpolation. There are two main reasons for this; the first is that the bishop no longer controls e6, so White can play J:.e6, and the second is revealed later in the game. Black still has dangerous threats, but analysis indicated that White is able to parry all of them. So, five years after the Yusupov game, I managed to play the right move!

The earlier game had gone 15 �b3 .ih4 16 l:.fl (D) and now Black has an interesting choice:

1) 16 . . . 'iWg6!? (a strong move which forces White to react quickly - he must be prepared to give up the exchange) 17 i.f4 .ih3 and now:

1a) 18 i.g3? f4! 19 gxh3 (19 i.xc6 loses to 19 . . . bxc6 20 gxh3 fxg3 21 fxg3 i.xg3 22 hxg3 'ifxg3+ 23 �h1 .Z:tfe8 24 'ifxc6 'ifxh3+ 25 �g1 'ifg3+ 26 �h1 ltad8) 19 .. .fxg3 20 fxg3 .ixg3 2 1 hxg3 'ifxg3+ 22 �h1 'ifxh3+ 23 �g1 ltfe8 24 ltf2

'i!fg3+ 25 ltg2 'ife3+ 26 \!i>hl L4! 27 i.fl l:.f8 and Black wins.

lb) 18 g3 gives White some compensation for the exchange.

2) 16 ... f4 (the game continuation) 17 f3 i.f5 1 8 llx:5 1fg6 19"1'dl ih3 20 'ifd2 J:ae8 21 lDd3 a6 22 .ixc6 bxc6 23 l2\e5 lbe5 24 dxe5 .ie6 (it's quite difficult for White 10 use his extra exchange and Black can still develop a kingside attack, so I decided to force a draw) 25 'l'd4 ib3 26 'l'd2 i.e6 27 'iWd4 1h-1h Anand· Yusupov, Linares 1993.

IS ••• .ihS After 15 ... i.h4 16 hxg4 fxg4 17

ltfl ! ( 17 lte2 is less accurate in view of 17 ... J:ae8) I don't think Black has enough compensation, sinceWhite easily copes with the immediate sacrifices: 17 ... i.xf2+ 18 ltxf2 :xt2 19 �xf2 llf8+ 20 �e2 "l'g3 21 �d 1 ! and the two extra pieces are enough. The only trap that White must avoid is 21 i.xc6?? llf2+! 22 �dl 'l'd3 23 i.xd5+ �f8 24 'i'b4-t

ANAND - KRAMNIK, TILBURG 1998 259

�e8 25 i.fl+ �d7 26 'iWa4+ b5 27 lxb5+ 'i'xb5, when Black wins.

16 lilb3 i.b4 17 lt:Jcs r D >

17 .ilf2+? This move loses at once - in my

game against Yusupov, the same continuation would have led to the capture of the pawn on h2 and a decisive attack for Black, but having it on h3 makes all the difference. The main alternatives are:

I) 17 ... l:l.ae8? 1 8 i.g5! i.xg5 19 �xb7 wins for White.

2) 17 ... lild8 defends b7 and e6, but i<> very passive. After 1 8 i.fl !, intending1i'd7, Black cannot equalize:

2a) 18 ... b6 19 lild7! (19 lt:Jd3 lt:Je6 20 lb4! is also good) 19 ... b5 (after 19 ... :e8 20 l:l.xe8+ i.xe8 21 lilf6+

i.xf6 22 'i'xe8+ 1Wf8 23 1i'd7 White wins a pawn) 20 i.xb5 c6 21 li:lxf8 cxb5 22 'i'xb5 �xf8 23 l:l.e5 with a large advantage for White.

2b) 18 ... c6 19 lild3 lile6 20 b3! is good for White; 2 1 i.a3 is lhreatened and Black has nothing to show for the weak squares on the e-file.

3) 17 ... f4 and now: 3a) 18 lilxb7? 1Wg6 19 �h1 ( 19

�1 i.xf2! 20 i.xc6 i.xe1 and 19 i.xc6 i.f3 win for Black) 1 9 ... i.xf2 20 lift f3 ! 2 1 g4 i.g3 gives Black a very strong attack.

3b) 1 8 l:l.e6! "l'd8 (D) and again White must be accurate:

w

3bl ) 19 li:lxb7 i.xf2+ 20 �xf2 (20 � h 1? 1Wh4 wins for Black after 21 i.xc6 i.f3! or 21 i.fl f3 22 'l'xc6 fxg2+ 23 i.xg2 i.f3) 20 ... 1i'h4+ 2 1 �gl f3 2 2 1Wc2 (22 i.xc6? 'l'g3 23 11fc2 f2+ 24 �fl 1Wxg2+! 25 �xg2 fi1W+ is lost for White) 22 .. .f2+ 23 �h2 (23 �fl ? 1Wg3) 23 ... i.g4 24 l:l.e3 (24 l:l.e4 dxe4 25 1Wxe4 fl'l' 26 i.xfl l:l.xfl 27 i.g5 1txg5 28 l:l.xfl is unclear) 24 ... fl1i 25 i.xfl l:l.xfl 26 l:l.g3 l:l.h1+ 27 �xhl 1txg3 28 hxg4 11fel+ is a draw.


3b2) 19 f3 !? ixf3 (19 .. .'6'g5 20 g4!) 20 gxf3 '11rg5+ 21 �hi 16'g3 22 16'c2 '11rxh3+ 23 1fh2 1fxf3+ 24 1i' g2 is slightly better for White.

3b3) 19 tt:ld3 f3!? 20 .i.xc6 (20 g3? 16'c8! 21 lt:\c5 ..txg3 22 fxg3 tt:ld8 favours Black) 20 ... bxc6 21 g3 16'c8 22 16'xc6 ..tf7 23 J:!.e3 may be good for White.

3b4) 19 .i.fl is the most solid. After 19 ... '11rc8 (19 ... .i.xf2+ 20 <t>xf2 'ii'h4+ 21 �gl f3 22 ..te3 is clearly better for White) 20 .i.d2 White has a safe, if slight, advantage.

18 �xf2 'i!Vb2 19 ixc6 bxc6

Or 19 ... f4 20 ..txd5+ �h8 21 tt:ld3 ! f3 22 lt:\r4 :xf4 (22 .. .fxg2 23 .i.xg2 l:Xf 4+ 24 .i.xf 4 :f8 25 :e4 is also winning for White) 23 .i.xf4 16'xg2+ 24 �e3 and White has too much extra material.

20 'i'xc6 f4 Black has nothing else.

21 'i'xdS+ �h8 2I....tf7 22 'i'f3 (22 'i'g5 is also

possible) 22 ... ..th5 23 'i'xh5 transposes except for an insignificant difference in the position of Black's king.

22 'i'xhS f3 (D) 22 ... 'i'g3+ 23 �fl f3 24 gxf3

.l:l.xf3+ 25 �e2 defends quite simply - the knight on c5 and the queen on h5 control all the key squares.

23 'i'xf3! In my preparation I had analysed

23 'i'xf3, but during the game I

noticed that White has two alterna tives:

1) 23 .i.f4 is not so good; Black has some chances after 23 ... fxg2 24 �e3 .l:l.xf4 25 �d3 .l:taf8! since 26 tt:le6? fails to 26 ... gl'i'! 27 lfufll (ZT .l:l.xgl :f3+ 28 �c4 16'e2+ 29 �b3 'i'xe6+ is also winning for Black) 27 ... 'i'gg3+ and Black wins.

2) 23 .i.h6 fxg2+ 24 �e3 "lg3+ (24 ... g l 16'+ 25 .l:l.xgl 'i'f2+ 26 �d3 :B+ 27 'i'xf3 ! 'i'xf3+ 28 � is similar to the game) 25 �2 and now:

2a) 25 ... :f2+ 26 :e2 gl'l' 27 .l:l.xgl 'l'xgl 28 .txg7+ (28 ie3 .l:l.xe2+ 29 'l'xe2 also wins for White) 28...�xg7 29 'l'e5+ picks up the rook on a8.

2b) 25 ... gl'l' seemed more trobleblesome. When I reached 11m point in my over-the-board analysis, I de cided that following my preparation with 23 'i1Vxf3 was muchmore sensi· ble. However, I later saw a win for White by 26 .txg7+! (not 26 1xgl•

ANAND - KRAMNIK, TILBURG 1998 261

:.12+ 27 'ilfe2 l:txe2+ 28 �xe2 l:l.e8+ and Black wins) 26 .. .'1i<g8 27 l:l.xgl (27 'i"d5+ 'ii<xg7 28 "ife5+ 'ii<h6 29 le6+ �h5 30 "ife2+ is also good) 27 ... l:.f2+ 28 "ife2 :Xe2+ 29 'ii<xe2 :ea+ 30 j.e5 with two extra pieces.

23 ... l:I.D3+ 24 'ii<xf3 l:l.f8+ 25 �e2 'i'xg2+

25 .. . l:te8+ is slightly better, but still loses after 26 .i.e3 "ifxg2+ 27 "'d3 'i'xb2 28 .i.cl ! 'Wb5+ 29 c4 lc6 30 l:txe8+ "ifxe8 3 1 .i.e3.

26 'ii<dJ 'i'xh3+ 27 �c2!

27 j.e3 was my original intention, but at the board I changed my mind because 27 ... l:l.f2! complicates matters a bit.

27 ••• 'i'g2+ 27 ... 'i'f5+ 28 ll:\d3 is also winning

for White.

28 .i.d2 'i'g6+ 29 :cl.e4!

Even better than 29 ll:\d3 UB 30 l:.e3.

19 ••. hS 29 ... l:l.e8 30 l:.gl ! is decisive.

30 l:.ae1 l:l.e8 31 'IL>cl lbe4 32 lbxe4!

This finishes it - after ll:\g5, the black pawns are firmly blockaded.

32 h4 33 lbg5 'i'hS 34 l:.eJ �g8 35 c4

Black can't capture on g5 due to 36 l:l.e8+ and can't move his queen from g6 or hS due to the mate on e8. He can therefore do nothing to prevent d5, c5, etc., creating a passed pawn.

1-0

The next game is from the Final of the European Clubs Cup, which was held in Belgrade at the start of 1999. At the time I was playing for the Agrouniversa!Club from Belgrade. We found ourselves facing a Polish team and the result was a fierce struggle. On top board, Shirov (playing for the Polish club) beat Kramnik, and the following game was played on board 2.

Once in a while you get a game where you totally lose control; the tactics start to swirl, and you can no longer hope to calculate everything - you just have to play on feeling. I was pleased to discover when analysing the game afterwards that it was, within reasonable bounds, accurate. Actually, I'm happy to win any game, but it is more pleasant to win a game where Fritz does'nt blow all sort in holes in it afterwards.

Game 49

V. Anand - L. 011 European Clubs Cup, Belgrade 1 999

Ruy Lopez, M!llller

1 e4 Lembit had a broad opening rep

ertoire, but with a certain fondness for topical lines. Therefore I anticipated either the Sicilian or the M0ller Variation, especially because Piket drew easily against Almasi with 1 2 ... exd4 (see the game) a month earlier in Groningen.

1 e5 2 ti,)f3 lbc6 3 �bS a6 4 h4 lDI'6 S 0-0 bS 6 �b3 �cS

I had guessed correctly! 7 a4 .:l.b8 8 c3 d6 9 d4 �b6

10 lba3 0-0 11 axbS axbS 12 lbxbS

All well-known stuff. After Svidler-Shirov, Linares 1998, this line was thought to be under a cloud from Black's point of view. However, players subsequently discovered that if Black played 12 ... exd4 before ... �g4, then some unpleasant variations could be avoided. I don't know if it was Glenn Flear or Kiril

Georgiev who played this idea first, but in any case it becameclearthat it was White's tum again!

12 ••• exd4! 12 ... �g4 13 l:l.el (for 13 .ie3,see

Game 44, Anand-Shirov) 13 ... exd4 ( 1 3 ... �xf3 14 gxO lLlh5 was at first thought to be fine for Black, but 15 �hi ! '6'f6 16 .:l.gl lLlf4 17 i.e3 proved good for White in SvidlerShirov, Linares 1998) may be met by 14 lLlbxd4! .

1 3 cxd4 Black has forced White to capture

with the c-pawn, as 13 ltlbxd4?! lbxd4 14 lLlxd4 hd4 15 cxd4 ltlxe4 only leads to equality.

13 ••• i.g4 14 .:tel

At that time, this was considered the main line, though subsequently new ideas such as 14 .:l.a4 have appeared. Morozevich-Ki.Georgiev, Bundesliga 1998/9 continued 14 �c2 d5 15 e5 l0e4 16 i.e3 and now Kiril played 16 ... f6, but 16...ltlxe5! is even easier for Black.

14 - dS (D) IS exdS

The alternative 15 e5 lLle4 16 '6'd3 �xf3 17 gxf3 lbxe5! 18 'i'e2 (18

ANAND - OU, EUROPEAN CWBS CUP, BELGRADE 1999 263

ldl 'ftb4! 19 fxe4 dxe4 20 �g2 <'tld3 is good for Black, while 18 dxe5 is met by 18 .... .bf2+ 19 �fl lt:lc5) !8 ... i'f6 (18 ... 'int4? 19 fxe4 lt:lg4 20 .if4 defends) 1 9 fxe4 (19 .i.dl ltld3 20 fxe4 dxe4 is also OK for Black) 19 ... 1ilf3+ 20 �g2 lt:lxe l+ 21 1Wxel c6 22 li:lc3 .i.xd4 is fine for Black.

15 lt:lxdS 16 h3 ..thS 17 g4

A few months later, in the game Shirov-Piket, Amber blindfold, Monaco 1999, White tried 17 .i.c4 1Wd7 18 .tg5 h6 19 .i.h4 J:fe8 20 J:txe8+ lhe8 21 .i.g3 (21 lt:lc3 .i.xf3 22 l'xf3 li:ldb4 23 d5 lt:le5 is also fine for Black)2 1 ... lt:lf6 22 J:ta3 lt:lxd4 23 <tbxd4 ..txf3 24 J:txf3 .i.xd4 with a roughly equal position.

17 ••• i.g6 18 lt:lc3 lt:ldb4!

The only way. 19 i.gS

At home, I stopped my preparation when I saw that I could kick

Black's queen around. At the board. I began to view the position more realistically.

19 ·- 1Wd7 20 d5 'ifd6

Later 011 said that 20 ... lt:la5 was OK for Black. I won't dwell on the opening too much since theory has subsequently developed quite rapidly in this line.

21 � I found this move after about 30

minutes' thought, the point being that ... 1Wg3 will not be check. After 21 �g2 lt:ld3 (21...hf2 22 �xf2 lt:ld3+ 23 �g2 lt:lxe1+ 24 lt:lxe1 is slightly better for White) 22 l:te2 ! (22 dxc6 is met by 22 ... ..txf2; for example, 23 J:te5 lt:lxe5 24 �xf2 lllxf3 25 �xf3 'lfh2! with a dangerous attack) 22 ... lt:ld4! (the simplest) 23 .i.e7 (23 lllxd4 .i.xd4 24 .i.e7 'ifb6 25 .i.xf8 lllf4+ is slightly better for Black) 23 .. . '1ff4 24 ltlxd4 1Wxd4 25 .i.xf8 lllf4+ Black will end up a pawn down, but his very active bishops and White's weakened kingside provide excellent compensation.

21 ••. li:las! (D) Not 21. ..lt:ld3 22 dxc6! .i.xf2 23

lt:le4! .i.xe4 24 :Xe4 and White wins. 22 ..te7 1Wr4 23 J:txaS?!

Since the game continuation is not very clear, 23 l:a4! was, objectively speaking, a better move. After 23 ... lt:lxb3 (23 . . . 'lfh6? loses to 24 h4! ll:lxb3 25 J:txb4) 24 J:txb4 'lfxf3


(24...�2+ 25 �g2 'iixf3+ 26 'iixf3 transposes) 25 'i'xf3 �2+ 26 �g2 itlxf3 27 �xf3 :feB an ending arises in which White has an extra pawn but Black can still put up a good fight. White is better, but I was afraid that his advantage might prove too small for a win. Seeing that White has lots of tactical resources, I decided to avoid liquidating into an endgame. However, it turns out that the rather speculative text-move offers Black even more chances than he would have in the endgame.

23 ••• haS 24 :te3

The black queen is running out of squares.

24 �b6! The alternatives are inferior: 1) 24 ... h5? 25 itle2 'inl6 26 g5

'i'h7 and Black saves his queen, but he is completely lost!

2) 24 ... :t"e8? 25 d6! �b6 (after 25 ... cxd6 26 �g5 ! the black queen is trapped) 26 d7 �xe3 27 dxe8'i'+

l:txe8 28 �g5 'i'xg5 Z!J ttlxg5 .bg5 is good for White.

25 ltle2 Wh6 26 �g2

Due to the weakness of h3,White must pause for this move. Remarkably, he still stands to trap the queen, but the price will be much higher.

B

26 �xe3 27 g5 (D)

27 Very pretty and I must admit that

it came as a shock to me. Neverthe less, White is better after this move. The alternatives are:

I ) 27 ... 'i'h5?! 28 itlg3 :fe8 19 �xb4 l:txb4 30 itlxh5 .ixh5 31 Pel l:txe3 32 d6! and now:

I a) 32 ... �xf3+ 33 'i'xO llxf3 34 dxc7 l:tg3+ with another branch:

lal) 35 �xg3 :Xb3+ J6 �4 l:tb4+ 37 �e5 llc4 38 � �! (38 ... f5? 39 gxf6 gxf6 40 b4 Acl 41 b5 l:tdl + 42 �e6 :tel+ 43 �xf6 :tfl + 44 �e5 wins for White) 33

ANAND - OIL, EUROPEAN CLUBS CUP, BELGRADE 1999 265

�d7 (39 b4 �eS 40 b5 :tel 41 b6 :dt+ 42 �c6 l:lcl+ leads to a draw) 39 ... ltd4+ 40 �c6 �e7 41 cS'if :C4+ 42 >\>b7 l:lxcS 43 �xeS �d6 (43 ... f5 44 gxf6+ gxf6 45 b4 �d6 46 �b7 f5 47 b5 f4 4S b6 t3 49 �cS f2 50 b7 f!'f 51 bS11r+ also draws) 44 b4 �c6 45 b5+ �xb5 46 �d7 and the king and pawn ending is a draw.

la2) 35 �f2 ! ! l:lf3+ 36 �xf3 l:lxb3+ 37 �e4 .l:tb4+ 3S �d3 and White wins.

lb) 32 ... cxd6 33 'ifxd6! (33 .td5 1xf3+ 34 .txf3 .l:txb2+ is less clear) 33. .. hf3+ 34 �f2 l:lbe4 35 .td5 :e2+ 36 �g3! (36 �xf3 l:l2e3+ leads to a draw) 36 ... .1:teS 37 �xf3 �b2 3S 'iWd7 l:tf8 39 h4 g6 (Black can't allow h5 since that would give White mating threats after a future h6 or g6; however, White now has a neat trick) 40 h5 ! gxh5 41 g6! hxg6 42 'fd6 :tbs ( 42 ... l:tb5 43 'ifxg6+ �h8 44 .txf7) 43 'ifxg6+ �hS 44 'l'xh5+ >\>g7 45 'l'xf7+ �h6 46 i.e4 :Zb6 47 �4 gives White a decisive attack.

2) 27 . .. .tc2 2S .txc2 and now: 2a) 2S ... "ifb6 29 .i.xb4 'ifxb4 30

�c3 'l'xb2 3 1 'ifd3 and White keeps a slight edge.

2b) 2S ... hg5 29 ll:lxg5 (alternatively, 29 .txg5 transposes to line '3' below) 29 ... ll:lxc2 30 'il'xc2 .l:tfeS 3 1 'l'xc7 'i'a6 looks OK for Black as White has poor pieces coordination.

3) 27 ... .txg5 2S .txg5 .tc2 29 .txc 2 'i'b6 is Black's best line. It is

very difficult to assess due to the unusual material balance, but I think that it's about equal.

28 �xb3 ll:ld3 29 'iWal!

Luckily, White can defend both bishops from the a3-square.

29 ••• ll:lxf2+ 30 �g2 :tfe8!

White wins after 30 ... l:lxb3 3 1 .txfS �xfS 32 'ifaS+ �e7 33 lbed4.

31 'iWa3 lbd3 (D)

Materially White is doing fine, but his pieces aren't coordinating very well.

32 �cJ? Mter this mistake White's pieces

get into a real mess. The correct line was 32 .ta4!, and now:

1) 32 ... .tc5 33 hc5 l:lxe2+ 34 �g3 .l:texb2 (during the game, I stopped my analysis after 34 ... .1:tbxb2, but I only had to see one move further: 35 .tc2! wins) 35 .ta7 should win for White.


2) 32 ... :txb2 33 "i'xb2 ll)xb2 34 ..ixe8 and White has a clear plus.

32 .•. ..id! 33 ll)dl

The knight had been looking forward to some heroic duty, but now it is forced back to defend a measly pawn.

33 ••• ..ie4! 34 ..ia4

34 �g3 .if4+ 35 �g4 is a critical line, but Black has a range of reasonable options:

1) 35 ... h5+ 36 gxh6 f5+ 37 �h3 ..ixf3 38 d6+ with a complete mess.

2) 35 ... :xe7 36 11fxe7 (White is genuinely threatening the e4-bishop because the d !-knight covers the f2-square) 36 ... ..ixf3+ 37 �xf3 ..id6 and after Black wins the bishop on b3 it looks about equal.

3) 35 ... l:l.a8 ! (to deflect the bishop from b3) 36 ..ia4 (D) and now:

3a) 36 ... h5+ was my first idea, but this tears open Black's kingside

and with an enemy queen and pair of bishops, it is perhaps not the wisest course of action. The main line runs 3 7 gxh6 f5+ 38 �h3 ..ixf3 39 d6 (39 'i'xd3 ..ig4+ 40 �g2 l:l.xe7 41 d6 l:l.e2+ 42 ll)f2 ..ie3 43 'i'd5+'t>h744 11fxa8 cxd6 45 hxg7 �xg7 is a draw) and now:

3al) 39 ... ..ig4+ 40 �g2 .ixdl (after 40 ... ll)el + 41 �fl liJO 42 d7 l:l.eb8 43 hxg7 Black looks to be in danger due to his expOSid king) 41 d7 l:l.eb8 42 1i'xd3 ..ixa4 43 'l'd5+ �h8 44 d81i'+ l:l.xd8 45 ..ixd8 and again White seems to be better.

3a2) 39 ... ll)el 40 M cxd6 41 lLixf5 is totally unclear.

3b) 36 ... l:l.eb8! is more solid be cause Black's king remains safe. Af. ter 37 d6 (37 b4 h5+ 38 gxh6 :b6!) 37 ... cxd6 (37 ... l:l.a5 38 Ql tiS+ 39 �h4 ..ixd7 40 "i'xd3 .ba4 41 b4 l:l.aa8 42 lLic3 favours White) 38 ..ixd6 ..ixd6! (38 ... lba4 39 'l'xa4 ..ixf3+40 �xf3 ..ixd6 41 'i'd4� 42 �g2 ..ic7 43 "i'c5 offers White winning chances) 39 "i'xd6 .id3+ 40 �xf3 lLixb2 41 .fub2 (41 .ic6 ll)xdl 42 ..ixa8 l:l.xa8 43 'i'xdl is a draw) 4l...:lxb2 the result is a draw. In contrast to line ' I b' of the note to Black's 27th move, here there is no h-pawn and White cannot break through with only the bishop and queen.

34 ••• lLieS 34 ... l:l.a8 is also possible. Then 35

�g3 ..if4+ 36 �g4 transposes to the

ANAND - Oil., EUROPEAN CWBS CUP, BELGRADE 1999 267

previous note, while 35 �xe8 :ctxa3 36 bxa3 �e5 37 �c3 �xf3+ 38 � is unclear; White is a pawn down but has a dangerous a-pawn. Although it is not easy to evaluate this line, it is clear that White is in no danger.

35 �xeS (D)

l:xe8? This error cost Lembit the game.

After 35 .. . �xf3+ he would have been very much in the game: 36 �fl �xdl 37 �b5 h5 ! (Black's king needs some air) 38 'i'c5, and now:

1) 38...�xb2 (White is left with very few pawns and therefore Black has some drawing chances) 39 d6 c6 (39 ... cxd6? loses to 40 �xd6 �a3 41 'lxe5 �xd6 42 'ilfxd6 l:xb5 43 ld8+ �h7 44 'ilfd3+) 40 �xc6 �g4 and it is doubtful if White can win. For instance, after 41 g6 fxg6 42 d7 �xd7 43 .i.xd7 �xd7 44 'i'd5+ �h7 45 'lxd7 .tel ! (intending ... �h6, when I don't see how White can ever break through) 46 'ilfc7 :cte8 Black

easily creates a forkess with queen vs rook and pawns.

2) 38 ... �f4 39 d6 c6 40 �xc6 (40 'i'd4 lZ.xb5 41 'ilfxf4 �g4 leads to a surprising fortress) 40 ... �e2+ (an amazing move; 40 ... �g4? 41 'ilfd4 favours White) 41 �gl (after 41 �xe2 lZ.xb2+ White's king can't escape the net) 4 l ...lZ.b3! (suddenly Black has counterplay against the white king) and now:

2a) 42 'ilfa5 lZ.g3+ 43 �f2 �c4! (not 43...�d3 44 ifa8+ liih7 45 �!) gives Black awkward threats against White's king.

2b) 42 d7 (to deflect the knight) 42 .. .c�:Jxd7 43 iff5 �e3+ 44 lith l �g4 45 'i'c2 �e6 and Black has reasonable compensation. If White takes on d7 Black forces perpetual check: 46 �xd7 �d5+ 47 �h2 M4+ 48 �gl lZ.g3+ 49 �fl lZ.f3+ 50 �e2 l:e3+ 5 1 �fl li:f3+.

36 �g3! (D)

36


Or 36 ... lLlxf3 37 'i'a4 and White wins.

37 'i'cJ .bdl 37 .. . :xe7 38 d6 :es 39 dxc7 :cs

40 'i'xe5 i.xdl 41 'i'd5! i.xg5 42 'i'xg5 is also decisive.

38 'i'xeS ..b4

39 b4! The final finesse, since now Black

can't reach an ending with 'i' vs .1:1.+8s. Both bishops will be lost, one for the d-pawn and one foc the bpawn, so Lembit resigned.

1-0

Despite this effort, my team lost the match and was eliminated.

After Belgrade, my next event was Wijk aan Zee, which turned out to be my most successful tournament of 1999. I always try to play in the top tournaments whenever possible, and there is no doubt that Wijk aan Zee, Linares and Dortmund are the leading traditional events, while Frankfurt is the outstanding rapid event. Thus I have a regular routine for at least part of the year, which is only broken under exceptional circumstances; for example, in 2001 I was touring India after my World Championship success and had to miss Linares. I feel that if a leading player doesn't want to play in one of these top tournaments, then he may be having a problem with his motivation. The following were my two best games from Wijk aan Zee .

Game 50

D. Reinderman - V. Anand Wijk aan lee 1 999

Sici l ian, Najdorf

1 e4 c5 2 liJI'3 d6 3 d4 cxd4 4 tllxd4 tllr6 5 ltlc3 a6 6 .ic4 e6 7 .ib3 bS 8 0-0 b4 9 ltla4 .id7

The same slightly unusual variation as was played in Game 41 (Lautiee-Anand).

10 f4 lllc6 11 f5 e5 12 lllf3?! (D)

Somehow this doesn't seem consistent with the aggressive f5. 1 2 oOe6 is the critical continuation, for which see the notes to Game 41.

12 ••• h6 After 1 2 ... .ie7 I was afraid that

White could seize control of d5 by 13 .ig5. Nevertheless, the d5-square doesn't mean much in itself (if the knight on a4 could land there, it would be a different story!), so this was a playable alternative. The continuation might be 13 ... '1rc7 (13 ... lllxe4? is impossible due to 14 'ird5 lllxg5 15 lllxg5, when White wins) 14 'lrd3 llla5 ! (14 .. . 0-0 15 :ad1) 15 .ixf6 .ixf6 16 llfd1 .ie7 and White has no follow-up since 17 .ixf7+ is insufficient in view of 17 .. .'�xf7 1 8 'ird5+ �6 19 g4 h6 20 h4 g5.

13 Wei .ie7 14 'iWg3?!

This seems to gain time, because Black cannot meet the attack on g7 by castling, but after Black's reply White is reminded of the weakness of the e4-pawn.

14 �! Now the move 'lrg3 does not serve

any real purpose, and White has to spend a tempo dealing with the attack on e4. Even if Black ends up playing .. .'�g8-h7, he has not lost any time.

15 'iWel


If 1 5 %:tel , then 15 ... ll:la5 followed by ... .lic6 steps up Black's pressure against e4, so White prefers to retract his previous move. Black already has a slight advantage owing to the awkward position of the knight on a4 and the weakness of the e4-pawn.

15 llb8 Black seeks to exploit the a4-

knight by ... 'ii'e8 followed by a move of the c6-knight, but first he must stop the knight escaping via b6.

16 .lid2 Or 16 a3 lLia5 (not 16 ... a5 due to

17 axb4 axb4 1 8 ..tc4, intending to reinforce the a4-knight by b3; if 18 ... lLia5, then 19 ..td3 is fine for White) 17 axb4 ll:lxb3 1 8 cxb3 i.b5 ( 18 ... ..tc6 19 lLic3 l:txb4 20 l:txa6 ..txe4 21 ll:lxe4 l:txe4 is just slightly better for Black) 19 l:tf2 i.d3 20 l:td2 ..txe4 21 ll:lxe5 'ilc7 22 ll:ld3 ..txf5 and Black has a large advantage.

16 ••• 'ile8! White is now forced to take action

to avoid losing the a4-knight. 17 a3 a5 (D)

Black needs to recapture with a pawn on b4 to stop the knight jumping to c3 and d5. After the text-move, Black intends to play ... �g8-h7, improving his position while leaving White handicapped by his vulnerable knight.

18 �h1 �g8 19 l:tg1?

White prepares to meet ... �h7 with the advance g4-g5, but this

move is artificial and gives Black the chance to seize the initiative. Note that in this position White is unable to support his knight by 19 axb4 axb4 20 i.c4 followe by b3, because 20 . . . ll:la5! attacks both c4 and a4. The best try is probably 19 h3, but after 19... �h7 20 axb4 axb4 White 's problems remain .

19 bxa3! A forcing solution, but Black coo1d

have also tried the quieter 19...�h7!?, playing to keep all the advantages of his position. Then White's intended attack 20 g4 (Black meets 20 'i'h4 by 20 ... l:tf8) can be met by 20 . ..lild4 21 ll:lxd4 exd4 22 g5 hxg5 23 :xg5! �g8! (not 23 ... i.c6? 24 'i'h4+ �g8 25 l:txg7+ and White wins) and now 24 :Xg7+ �xg7 25 "ifg3+ � 26 l:lg l fails to 26 ... ll:lg4 (26 . .. �d8? 27 "ifg7+ �e7 28 l:tg6! wins for White) 27 'ilxg4 ..tf6 and Black defends.

20 :Xa3 li:ld4! Black aims to exploit the weak

nesses at a4 and e4; the exchange on

REINDERMAN - ANAND, WIJK AAN ZEE 1999 271

d4 doubles Black's pawns, but open up new lines of attack such as the e-file 20 . .. d5 is wrong since 21 exd5 i.xa3 22 dxc6 i.xc6 23 lt:Jxe5 favours White.

21 li:lxd4 After 2 l lt:lc3 lt:lxb3 22 cxb3 i.d8,

followed by ... i.c6, White is in very bad shape.

21 exd4 (DJ

The pawn on b4 has disappeared, but the one on d4 does an equally good job of confining the a4-knight. The immediate threat is 22 . .. d5.

n :at The rook has to retreat, but now

Black turns his attention to the e4-pawn.

n ... .i.c6 22 ... i.f8 is the wrong plan. Black

tries to stop e5, but it's too slow and 23 .ixa5 l0xe4 24 :n .i.c6 25 l0b6 should enable White to generate sufficient counterplay

23 i.xaS

After 23 e5 dxe5 24 'ilfxe5 � 25 i.el J:l.b5 (25 ... d37! 26 i.d5 i.xa4 27 lha4 dxc2 28 J:l.c4! is not so clear) 26 'ilrxd4 J:l.xf5 Black is doing very well - the e 1-bishop can't go anywhere due to ... lt:Jg3+ and he can still continue his attack by ... h5-h4.

23 .•• fue4 Threatening, amongst other ideas,

... i.g5-e3. 24 :n bS! (DJ

It turns out that Black need not free his rook with ... 'it>h7 as he can use it actively on its original square. The advance of the h-pawn, together with Black's pressure on the long light-square diagonal, signals the start of a direct attack on White's king. With all White's minor pieces stranded on the queenside, he is in no position to repel Black's threats.

25 l!Jb6 Trying to block the dangerous di

agonal by occupying d5. 25 h4


26 �dS Or 26 !MS �gS 27 l0f'4 (this

doesn't help, but 27 f6 �xdS 28 �xd5 ttlg3+ is also lost for White) 27 ... h3 with an enormous attack.

26 �gS Now there is a threat of 27 ... l0g3+

28 hxg3 hxg3+ 29 �gl �e3+. 27 l:tr3

27 �xc6 does not stop 27 ... l0g3+, while 27 'llfxe4 'llfxe4 28 �xe4 he4 29 h3 �xc2 gives Black a winning ending.

27 h3! (D) The pressure on the long diago

nal reaches its zenith. 27 ... lixb6? is wrong because of 28 'ilfxe4!.

28 �xc6 28 g3 l::txb6! wins for Black.

28 ••• 'ilfxc6 29 'ilfe2 �f4!

Threatening to take on g2 and then h2. White has no defence.

30 ��1 30 l::txf4 lllg3+ 3 1 hxg3 hxg2++

32 �gl .l:lhl + 33 �2 :xal also wins for Black.

30 --· h3+ 31 �h1

Black's advantage is so large that he has several ways to win.

31 ... :es Threatening 32 ... M4 or 32 ... ll)g:;.

3 l ...hxg2+ 32 '11Vxg2 l::th5 followed by . . .l::tg5 would also I!Jve won.

32 11Fc4 lllf2+ 33 l::txf2 'ilfxc4

Another way to win is 33 ... '1!fe4 34 l::tf3 hxg2+ 35 �xg2 1bh2+ 36 �xh2 'llfxf3.

34 l0xc4 35 l0xd6 36 b4

bf2 l::te2 �! (D)

A cruel finishing blow. 37 gxb3 hd6

0·1

Game 51

V. Anand - J. Piket Wijk aan lee 1 999 Ruy Lopez, Chigorin

1 e4 Jeroen has a fondness for the Ruy

Lopez as Black and has played almost all its subvariations at one time or another. So I sat wondering which one it would be this time.

1 ... e5 2 �f3 �c6 3 ..tbS a6 4 h4 �6 5 0..0 ..te7 6 :e1 b5 7 ..tb3 d6 8 c3 0..0 9 h3 ll:\aS

Later in the tournament he went for the Breyer, but today he wanted something else.

10 ..tc2 cS 11 d4 lM7

An old favourite of Keres. 12 �bd2 cxd4 13 cxd4 �c6 14 �b3

14 �fl �xd4 15 �d4 exd4 gives White nothing.

14 .•• aS 15 ..td3!

Stronger than the more commonly played 15 ..te3; after the text-move Black has serious problems to solve.

15 ... h6 15 ... a4 1 6 hb5 1ib6 17 ..txc6

'lfxc6 18 �bd2 is known to be better for White.

16 dS �b4 11 ..tn (DJ

17 �xa5? is premature, since after 17 ... �xd3 (17 ... '1fxa5 18 ..td2 'lfa4 1 9 b3 'il'a3 20 ..tcl 'lfa5 2I ..td2 is a draw) 1 8 �6 �xel 19 �xd8 �xf3+ Black has too much material for the queen.

17 ... a4 The point of White's play is that

17 ... :c8 may now be answered by 18 �xa5!:

I ) 18 . . . lLJc2 19 �c6 is very good for White.


2) 1 8 ... l:l.xcl 19 l:l.xc l ( 19 'i'xc l 'i'xa5 20 a3 is also very good for White) 19 ... 'i'xa5 20 a3 lba2 2 l l:l.al and White wins.

3) 18 ... 'i'xa5 19 i.d2 and Black cannot stop a3, regaining the piece with a clear advantage for White.

4) 1 8 ... lbb8 and now: 4a) 19 lte2? 1Vxa5 20 i.d2 'i'a4

favours Black. 4b) 19 ltbl lbc2 ( 1 9 ... 'i'xa5 20

i.d2 'i'xa2 21 i.xb4 'i'a4 22 'i'd2 is good for White) 20 lbc6 (20 J:!.e2 'i'xa5 2 l l:l.xc2 'i'xa2 leads to equality) 20 ... l:l.xc6 2 l l:l.e2 lbd4 22 dxc6 lbxe2+ 23 i.xe2 and White has a slight advantage.

4c) 19 a4! is strongest; for exam-ple, 19 ... lbc2 ( 1 9 ... bxa4 20 l:l.xa4 and 19 ... 'iWxa5 20 i.d2! are also un-satisfactory for Black) 20 axb5 lbxal (20 ... i.xb5 loses to 21 i.xb5 lbxal 22 lbc6 lbxc6 23 i.xc6) 21 bxa6 'i'xa5 22 i.d2 'i'b6 23 'i'xal with fantastic compensation for the exchange.

18 lbbd4! This is still theory. I knew that it

was an idea of Shamkovich (later I found it was played in ShamkovichBenjamin, USA 1976). Jeroen knew of it as well, but we both thought it led only to a slight advantage for White. After the game (and perhaps because of it!) we agreed that Black's task was incredibly unpleasant. The alternative is 1 8 a3 lbxd5 19 'iWxd5 lbb6 20 'i'dl axb3 21 'l'xb3, which

theory considers to give White some advantage, but I think it is too little to provide real winning chances.

18 ••• exd4 19 a3 lbxdS

The alternative is 19 ... lbc5 20 axb4 lbb3 2 l l:l.bl i.f6 (21 ...'iWb6 22 i.f 4 is also clearly better for White) 22 i.f4 l:l.e8, but White can just play for i.d3-c2. The pawn-structure is just awful for Black, since d6, d4 and b3 are all weak, while Black cannot achieve .. .f5 any time soon.

8

20 exdS i.f6 21 lbxd4 (D)

White has regained his pawn. If Black could activate his light-squared bishop he would be fine, but this is impossible since b5 is so weak.

21 •.. <'UcS There is no really satisfactory

continuation; for example, 2I...'I'b6 22 i.e3 does not solve !he problem of the b5-pawn. Perhaps 2l...ixd4 is relatively best, eliminating the

ANAND - PIKEr, WIJK AAN ZEE 1999 275

knight before it anives on c6, but even here White retains an advantage after 22 W' xd4 ltlc5 23 'il'b4.

22 ltJc6! Now if the queen goes to d7, Black

will never be able to release the bishop from a6.

22 •.• 'ilb6 This defends b5 but the d6-pawn

becomes a big problem. 23 .U4 lUe8

23 ... .ixb2 24 :a2 .U6 25 .ixd6 is very good for White.

24 Wfc2! This move required some calcula

tion (see 25 ... ltld4 in the note to Black's 25th move). 24 J:l.bl is possible but would lose time compared to the text-move. 24 .ixd67 is totally wrong in view of 24 .. . lt:\e4! attacking d6 and f2.

24 ... ltlb3 lS l1adl! (D)

Thanks to 24 'i'c2!, this rook can at once move to an active square.

25 ... W'c7 After 25 ... J:I.xel 26 J:l.xel ltld4 27

1i'e4 White wins thanks to Black's back rank- a recurrent theme during this phase of the game. 25 ... ltld4 is the critical line, when White continues 26 J:l.xe8+ (26 ltlxd47 is ineffective due to 26 ... J:I.xel ! 27 J:l.xel .ixd4) 26 ... J:I.xe8 27 ltlxd4 .ixd4 28 J:l.xd4! (the point, since otheJWise Black is out of danger) 28 ... W'xd4 29 W'c6 W'e4 (after 29 ... J:I.d8 30 .ig5! f6 3 1 .ie3 White wins material) 30 .ixd6 .ic8 3 1 .ixb5 J:l.d8 32 W'c7 J:l.e8 33 .ig3! (after 33 .ixe8 W'xe8 White still has a long technical phase ahead of him) 33 ... .if5 34 d6 with a winning position.

26 J:l.e4! Since White threatens to double

on the e-tile and thus take control of it, Black now has no choice.

26 J:l.xe4 27 Wfxe4 hb2 28 J:l.el!

28 W'b4 .ie5 is less clear -cut. 28 W'd7

Black has to deal with the backrank threat.

29 W'b4 .U6 30 .ixd6

The battle is over because the difference between the activity of the two armies is too great - just compare the pieces on e I, c6 and d6 with those on b3. a6 and aS.

30 h6 31 .id3!


The threat is 32 �e7+ 'it>h8 33 'ilt'e4 with a decisive attack.

31 .•• lle8 32 llxe8+?

This doesn't spoil anything, but White could have finished the game at once with 32 �b8 !.

B

32 'li'xe8 33 i.e7! (D)

Threatening 34 'li'e4. 33 ..tb2

Black cannot save himself; e.g., 33...�2 (33 ... .i.xe7 34 ltlxe7+ lt>h8

35 'li'e4 is winning for White) 34 .i.xf6 'li'el + 35 \t>h2 'li'xt2 (35 ... ltlf3+ 36 \t>g3 and 35 ... gxf6 36 li:Je7+ �h8 37 111h4 'li' e5+ 38 f4 are decisive) 36 �e7+ 'it>h8 37 �g6+'it>h7 38 li:Je5+ lt>g8 39 'fle7 with an easy win foc White.

34 'fle4 g6 35 i.b4!

White could try to finish the game with a mating attack, but this is simplest - Black loses all his queenside pawns.

35 ••• 'l'xe4 36 he4

Black's pieces must rush back to stop the d-pawn.

36 .if6 37 d6 i.c8 38 i.d3! ..td7 39 ltlb8! i.e6 40 hb5 � 41 ha4 1-0

lbis game secured my second public prize of the tournament foc the best game of the day.

I managed to make +6, which turned out only to be sufficient foc second place, as Kasparov made an amazing + 7. Although I did not win the tournament, it was a great success for me, measured not only by the score but also by my standard of play; foc example, I won four public prizes.

At the beginning of 1999, I was still luxuriating in a run of almost uninterrupted success which had started with the Krarnnik game frocn Belgrade 1997 (see Game 42). However, all good things must come to an end, and foc

me the tide turned during Linares 1999.

I started the tournament with five draws, and then won the foUowing game.

Game 52

V. Anand - P. Svidler Linares 1 999

Grunfeld Defence

1 d4 2 c4 3 � 4 rn 5 'ii'b3

llJr6 g6 d5 .i.g7

The Russian System, in which White expends time with his queen in order to build up a strong centre, but Black has a lead in development to compensate. Play is usually very

sharp in this line and the current game was no exception.

5 dxc4 6 'ifxc4 0-0 7 e4 a6

Svidler repeats the so-called Hungarian system, which he had played against Kasparov a month or so before this game. 1 .. lDa6 and 7 .. .i.g4 are alternatives.

8 e5 b5 9 'iWb3 llJrd7

10 b4 Kasparov played 10 e6 and went

on to win, but I assumed that Sv idler had an improvement ready over that game. I myself had a new idea ready in one line, but Peter pre-empted me by playing adifferent a sub-variation 0 I...c4).

10 ... c5

11 e6 c4 (D)

I had briefly acquainted myself with the theory of this move, but had concentrated on Black's alternatives. Now I had to check my memory of theory.

12 1i'd1 13 exf7+ 14 b5 15 bxg6 bxg6 16 .i.e3

This is all theory. The main move is 16 ... .i.g4, but Svidler had something else in mind.

16 ... .i.fS! (D) Svidler said that he was surprised

this natural move had hardly ever been played.


17 tLlgS Obviously, this possibility is the

key difference between 1 6 ... ..tf5 and 16 ... i.g4. Here I could have transposed into the game Lputian-I.Sokolov, Sarajevo 1998 by playing 17 ..te2 tLlb4 (the move-order of Lputian-Sokolov was 1 6 ..te2 ..tf5 17 .i.e3 tLlb4) 18 l:.cl , when Sokolov continued 1 8 ... tLl6d5. However, I thought it would be better not to follow Lputian-Sokolov blindly, so when I saw the alternative plan of 17 tLlg5, I decided to go for it. Of course, attacking without developing all your pieces is risky, but this system calls for some aggression! Later I discovered that Sokolov's play had been improved by 18 ... e5 ! in Kroeze-Van Haastert, Wijk aan Zee 1998 (no, not Hoogovens, but Sonnevanck!). I haven't checked everything, but it looks good for Black.

The text-move isn't totally illogical. White wants to play g4 and ..tg2 instead of placing the bishop on e2.

However, with accurate play Black should be OK.

17 -- l'mi After 17 ... tLlxd4 Black seems to

gain sufficient compensation for the exchange; for instance, 18 fiJxfT � 19 .:tc 1 e5 (the later game Sashikiran-Ahmad, Vung Tau City 1999 went 19 .. .lla7 20 g4 ..tc8 21 i.g2 ltd7 22 ..te4 l:.d6 23 <MI and White won, but at this point it's quite unclear) 20 g4 .i.e6 21 ..tg2 L7 with active play.

18 g4 ..tefi! The bishop heads for d5. 18 ... .td7

19 ..tg2 gives White an edge. 19 llJce4

19 i..g2 i..d5 20 liJxd5 tLlxd5 is slightly better for Black.

19 ... .ids 20 'ird2(D)

20 tLlxf6+ exf6 21 tLlf3 gives Black too much compensation after 21 . ..1i'd7 or 2 I . ..lba4.

20 ltd6?

ANAND - SVIDLER, LINARES 1999 279

White's attack on the h-file more than compensates for the weak pawn on d4 and in fact Black never manages to capture this pawn. Black should have played 20 .. ."ifd7!, as Svidler suggested after the game.

The idea is that Black can now meet 21 f3 with 2 l ...�xf3. It would take us too far afield to analyse this in detail , but it's clear that this was a betthan 20 ... l:td6.

21 f3! Now the rook will be sorely missed

on the f-file. 21 .•• �xe4

2l...ltlxd4? 22 hd4 he4 loses to either 23 "ifh2 or 23 �xg7 J:txd2 24 i.e5.

22 fxe4 (D)

22 ltld7 22 ... ltlxd4 is a critical alterna

tive : I) 23 'l'h2? was a move I consid

ered during the game. White intends 24.ixc4+ ltlxc4 25 "ifh7+ �8 26

hd4 l:txd4 27 0-0+, but Black appears able to defend by 23 ... e5:

Ia) 24 0-0-0 ile7 25 'irh7+ (25 l:!.d2 �f6 26 llJIJ7 �g7 is a draw, while 25 �g2 is met by 25 ... �f6) 25 .. .'itf8 26 l:td2 �e8 defends.

lb) 24 �d3 clears the fl-square, but doesn't allow the knight to come to c4 with tempo. However, it isn't check, so Black can ignore it by 24 ... �f6!, when White has no followup.

!c) 24 hc4+ ltlxc4 25 'irh7+ Wf8 26 0-0+ J:tf6 27 lbf6+ (after 27 'irxg6 �g8 28 'irh7+ �f8 the interpolation of1ixg6 doesn't help White; for example, after 29 J:txf6+ 'irxf6 30 �xd4 'lrxg5 31 l:tfl + �e7 the black queen defends the g7-bishop) 27 ... 'ifxf6 28 �xd4 'irxg5 and there is nothing clear. One continuation is 29 :n + �f6 30 �c5+ �e8 3 1 'lrg8+ �d7 3 2 'irxa8 'lrxg4+ 33 �hi "ifh3+ with a draw.

2) 23 �xd4! is the right move: 2a) 23...�xd4 24 'lfh2 �g7 (or

24 ... �8 25 e5 winning) 25 e5 ! l:tc6 26 �e2 and White wins.

2b) 23 ... l:txd4 24 'lff2! (after 24 'lfh2 'ifd6! 25 �xc4+ltlxc4 26 "ifh7+ Wf8 White has nothing) 24 ... 'lff8 (24 ... �6 25 l:th8+! and 24 ... 'IreS 25 11'h4 are also decisive) 25 'lrh4 'iff6 26 �e2! and Black seems to have no defence against 27 'lfh7+ followed by 28 :n.

23 'ifhl (D) 13 l"flf8


If White gets 'l'h7+ in, then it's usually all over; e.g., 23 ... 'Wb6 24 �xc4+ bxc4 25 'l'h7+ � 26 0-0+ �6 27 'l'xg6 (27 J:l.ae 1, intending 28 e5, is also good) 27 ... �g8 (27 .. .l2:le5 loses to 28 dxe5 'l'xe3+ 29 �h1) 28 'l'f7+ �h8 29 l:tf3 lllxg4 30 l:th3+ lllh6 (30 ... J:I.h6 3 1 J:l.xh6+ lllxh6 32 'l'g6 �g8 33 l:tfl mates) 3 1 'l'f5 �g8 32 'l'h7+ �f8 33 J:l.fl+ and White wins after 33...�e8 34 'l'xg7 or 33 ... J:I.f6 34 llle6+.

24 eS! 24 d5 llle5 allows Black to keep

fighting. 24 ... J:l.d7

24 ... J:I.xd4 fails to 25 'l'f2 'l'a5+ 26 �e2.

25 llle6! 'I' aS+ Or 25 ... lllxe6 26 �xc4! (a recur

ring theme) 26 ... '1'a5+ (26 ... bxc4 loses to 27 'l'h7+ �f7 28 0-0+) 27 .id2 'l'xd2+ 28 �xd2 l:txd4+ 29 �e3 bxc4 30 'l'h7+ �f7 3 1 J:l.afl+ and White has too much material.

26 �d2 lllxeS! (D)

26 ... '1'b6 1oses to 27 lllxg7.

27 �e2? Missing a clear win by 27 i.g2!

which, like .ie2, covers the f3-square but also attacks the aS-rook; after 27 ... c3 (27 ... fug4 28 baS ltlxh2 29 .ixa8 is hopeless for Black) 28 bxc3 lllxg4 29 'l'h3 the game is over. On the other hand, 27 i.xa5? isn't so clear since 27 ... lllo+ 28 � lllxh2 29 lllxfB J:l.xf8 30 J:l.xh2 i.xd4 3 1 J:l.bl gives Black has a fair amount of play for the piece.

27 ... c3 (D) 27 ... 'Wb6 requires an accurate re

sponse: I ) 28 lllxf8? only complicates

matters after 28 ... J:I.xd4. Black is fighting on after either 29 i.c3 J:l.xl8 or 29 llle6 'l'xe6 30 'l'h7+ � 31 :n + �6 32 o-o-o llln 33 l:l.xf6 '1'xt6 34 :n 'l'xfl+ 35 .bn .

2) 28 lllxg7! is a clean kill; after 28 ... lllf7 29 �5 gxf5 30 gxf5 there is no defence to 31 :SI+.

ANAND - SVJDLER, LINARES 1999 281

28 .bc3?! A slight slip. 28 bxc3 ! li:Jf7 29

&g7 �xg7 is better, as the d4-pawn is more secure than in the game.

28 •.. b4 l9 l2Jxg7 bxc3?

Svidler thought he was already lost and didn't find his last chance: 29 ... li:Jf7! 30 .td2 �xg7. Then the open h·file and two bishops promise White some advantage, but White has to watch the d4-pawn so Black is still fighting.

A last joke. 31 0-0+

'M1 (D)

The reason why Peter continued so long is that he wanted to play a game where White wins with 0-0+. Unfortunately, the game was published on the Internet as continuing '3l llfl + 1-0' and it took a while for this to be corrected. So, for the record, 31 0-0+ was the move played.

1-0

After another three draws, I faced Kasparov with the white pieces. Round about this time there had been a lot of talk about a possible title match against Kasparov. I think this started to affect me during 1999; instead of simply playing freely and easily, my mind was on other matters - making a good impression for the match, and so on. I think that I have now learned to shut these extraneous matters out of my mind, but at the time I wasn't able to do this. The game against Kasparov therefore assumed a particular importance, since it could not only decide the destination of first prize at Linares, but also influence our morale before a potential match. The game proved remarkable enough. I ran straight into some very deep opening preparation by Kasparov. Although I had analysed the same line, it was soon clear that my preparation had been lacking. In the critical position, I had analysed several continuations,


but missed the one which he actually played, which forced White to find a very narrow path to a draw. I missed it, and was driven into a very bad endgame. However, I then started to defend well and, aided by some inaccuracies from Kasparov, I managed to reach a drawn position. Unfortunately,just at this moment I started looking for too clean a forced draw and I made a falal blunder.

If I had salvaged this game, then 1999 might have been a very different year for both Kasparov and myself. I would have been given a huge psychological boost by saving a game against an important rival Kasparov, on the other hand, would have known that he had failed to score a full point despite being handed a winning position from the opening. As it happened, our careers went in opposite directions: he went on to have one of the best years of his life while I went on to have one of the lousiest years of mine.

Despite the disappointment against Kasparov, I still managed to win another nice game in the tournament, against Topalov.

Game 53

V. Topalov - V. Anand Linares 1 999

Caro-Kann Defence

1 e4 c:6 2 d4 dS 3 eS ..trs 4 ltlc:3

4 00 had been the main line for some time, but Black had been doing OK against that. It was inevitable that White would return to 4 ltlc3.

4 e6 5 g4 ..tg6 6 ltlge2 ltle7

The old main line was 6 ... c5, but by the time this game was played 6...l1Je7 had become equally popular.

7 ltlr4 8 h4 9 ltlbS

10 hS 11 f3 (D)

c5 c:xd4 ltlec:6 ..te4

All this is familiar theory. In previous games, Black had now played ll.....ixf3 and got a whole mass of for pawns for the piece, but in return White got a strong attack with gS and h6. My trainer Ubilava and I wondered if Black could get these moves in himself.

11 12 ltld6+ 13 exd6

a6! bd6 gS!

This allows Black to set up a strong pawn-front on the kingside and also misplaces White's knighi. 1 3 ... e5 is inferior due to 14 ltlg2! 'Wxd6 IS fxe4 dxe4 16 ltle3; the knight is very much in the game and so Black doesn't have enough compensation for the piece.

14 ltlh3 14 ltlg2 makes no sense with

Black's pawn on gS. 14 h6 15 fxe4 dxe4 16 ..tg2 rs 17 0-0 0-0! (D)

It looks safer not to 'castle into it', but after 17 ... llf8 18 gxfS exfS 19 ..ixe4 fxe4 20 llxf8+ �xf8 21 .ixgS


hxg5 22 'ii'g4 White's attack is quite dangerous. By castling, Black is able to recapture on f8 with his queen.

18 c3? White may already need to take

drastic measures and force a draw, since the game continuation looks quite convincing for Black. Therefore the correct line is 1 8 gxf5 exf5 19 he4 (19 i.xg5 hxg5 20 he4 'ii'xd6! gives Black chances of an advantage) 19 .. 1xe4 20 i.xg5 ! (20 :xrs+ 'ii'xf8 !) 20 ... llxf1+ (20 ... hxg5 21 'ii'g4 is more dangerous for Black) 21 'ii'xfl hxg5 22 'ii'f5 'ii'd7 (22 ... 'ii'f8 is also a draw) 23 'ii'g6+ (23 'ii'xe4 'ii'xh3 24 'ii'g6+ �h8 is another draw) 23 ... 'ii'g7 24 'ii'e8+ 'ii'fS with perpetual check. This draw has since occurred in at least four games !

18 1Vxd6 19 gxfs exrs 20 1Vb3+

After 20 i.e3 'M7! 21 i.xd4 (21 cxd4 lDb6 favours Black) 2I ... lDxd4

22 cxd4 lDb6, followed by ... �. Black has good compensation for the piece.

20 -· �h8 21 he4

White finally sacrifices the piece. However, the alternatives 21 11'xb7 lDd7 22 1Vb3 lDde5 and 21 cxd4 lDxd4 22 1Vc3 lDbc6 are also good for Black.

21 ... fxe4 22 :xt8+ 9ixf8 (D)

23 1Ve6! Black is two pawns up but still has

to find a way to develop his queen· side. White threatens simply 24 i.xg5 followed by 25 llfl .

23 lbd7! Funnily enough, Black has to t<ES

back the piece that he just recovered. Honestly speaking, this was a fairly intuitive sacrifice, but I felt that Black was better and didn't see any other way to get more than a draw. Note that 23 ... e3? fails to 24 i.xe3!

TOPALOV - ANAND, LINARES 1999 285

0d7 25 :n 'I' g7 26 :n lOts 21 'lfS and White wins.

24 'l'xd7 24 .1xg5? hxg5 25 'l'xd7 l:td8! 26

'le6 :e8 27 'I' g6 lLle5 wins for Black.

24 .:td8 (D)

25 'l'g4! Topalov chooses the correct de

fence. In a practical game it is hard to play amove such as 25 'i'xb7, after which Black has a sure draw and chances to try for more. Indeed, it turns out to favour Black:

I) 25 ... '1'f3 (it was when I reallized that ithis move gives Black at least a forced draw that I decidet to play 23 ... lLld7; however, it is not the best move) 26 lLlxg5 hxg5 27 'i' xc6 1g4+ 28 �h2 'i'xh5+ 29 �g2 'i'g4+ 30 'i>h2 'ilih4+ 31 �gl (31 �g2? loses to 3\,.:f8!) 3 I . .. .:tf8 (the al-ternative 31... 'i' g3+ 32 'ito>h I 'ilth3+ 33 'i>gl .:tf8 34 .i.d2! transposes to the main line of this note) 32.1d2

1fg4+ 33 �hi 'it'h3+ 34 �gl e3 35 'iie4! 'l'g3+ (D) and now:

Ia) 36 �hi? and then: lal) 36 ... .:tf6 37 'it'xd4 (37 'ife8+?

�g7 38 1!t'd7+ �h6 39 1!t'xd4 1!t'f3+ 40 �h2 exd2 ! is winning for Black) 37 ... 'it'f3+ 38 �h2 'it'f2+ 39 �h 1 exd2 40 1!Vxf2 .:txf2 41 �gl ! and White draws.

la2) 36 . . . exd2! 37 'ifxd4+ �h7 38 1!t'd7+ (38 1!t'xd2 .:th8! wins for Black) 38...�g6 39 'it'e6+ �h5! (39...�g7 40 1!t'd7+ .:tf7? 41 1!t'xd2 is a draw) 40 'ife2+ 'i'g4! 41 'ifh2+ (41 '6'xg4+ �xg4 42 �g2 .:td8 43 .:tdl �f4 is decisive) 4l ...�g6 42 1i'd6+ .:t6 43 1i'xd2 .:n wins for Black.

lb) 36 'it'g2 'it'xg2+ 37 �xg2 .:tf2+ 38 �g3 .:txd2 39 cxd4 .:txb2 and now:

lbl) 40 �g4? e2 41 �xg5 l:txa2 42 .:tel �g7 (42 . . . a5? 43 �6 �g8 44 �e7 is a draw) 43 <M4 a5 44 �e3 a4 45 <M2 a3 and Black wins.


lb2) 40 .l:el ! e2 41 a4 �g7 42 �f2! draws.

2) 25 ... 'iid6! 26 �g2 (D) with two possibilities for Black:

2a) 26 ... e3 27 cxd4 (27 'iixa6? loses to 27 ... g4!) and now:

2al) 27 ... 'iid5+ 28 �h2 'iif3 !? 29 'iic7 (29 .be3 .l:b8 30 'iid7 is also playable) 29 ... .1:xd4 (29 . . . g4 30 'iif4 'iixh3+ 31 �gl is roughly equal) 30 'iic8+ .l:d8 31 'iie6 'iie2+ 32 �g3 'iiel+ 33 �g2 with a draw.

2a2) 27 ... g4!? causes White more trouble:

2a21) 28 ..txe3? just seems to lose on the spot to 28 ... gxh3+; e.g., 29 �xh3 'iie6+ 30 �h2 'iixe3 31 'iixc6 'iif4+ 32 �hl .l:g8.

2a22) 28 �g5 hxg5 29 ..txe3 'iVd5+ (White seems to be fine after 29 ... .1:b8 30 'iin .l:xb2+ 31 �fl) 30 �gl 'iif3 with another branch:

2a221) 31 ..txg5 �xd4! wins. 2a222) 3 1 .l:el 'iig3+ 32 �fl

'iih3+ and White is losing after 33

�gl g3! or 33 �e2 'l'f3+! 34 i>d2 �xd4.

2a223) 31 'iib3 'iig3+ (Black has nothing after 3 l ... .l:e8?! 32 .if2) 32 �hl 'iih3+ 33�gl g3 34 'iic2 g235 llel 'iVh 1 + 36 �2 ID"8+ (36 .. .loe5? 37 'iif5! is fine for White) 37 �2 l:lfl (37.../t)e5 38 dxe5 'iif1+39�1 .l:d8+ 40 'iid2 "i'f3+ 41 �cl lxd2 42 �xd2 'i!rd5+ favours Black, but is less effective than the text-move) 38 'iic3 /t)e7! (38 ... gl'i!V? 39 hgl Wg2+ 40�dl 'iVxg1 41 d5+ lLlli442•c8+ �h7 43 'iVb7+! �h8 is a draw) 39 d5+ �h7 40 h6 �5 41 "i'c7+ i>xh6 42 'iic6+ �h5 43 'iie8+ � with a winning position for Black.

2a23) 28 �gl 'iie6! (28 ... • d5+ 29�g3 and theredoesn't seem to be more than a draw) 29 'fic7 'fie4+ 30

�g3 e2 gives Black a large advantage.

2b) 26 ... 'iif6! (this is even more effective) 27 cxd4 leads to arDher branch:

2bl) 27 ... .1:£8? 28 �h2 (28 l'xa6? 'iif3+ 29 �h2 g4 30 0f4 g3+ 31 �h3 g2+ 32 �h2 'l'fl 33 li:lg6+ �g8 wins for Black) 28 ... 'i'd6+ 'B �g2 .l:f3 30 'iic8+ �g7 31 'l'g4 'iixd4 is unclear.

2b2) 27 ... 'iif3+! 28 �h2 li:le5! (28 ... g4? 29 "ifxc6! g3+ 30�gl "i'dl+ is a draw) 29 'iic7 (29 dxe5 'l'e2+) 29. ..�g4+ 30 �gl 'fidl+ 31 i>g2 .l:f8 ! and Black wins.

2S eJ (D} 26 b3

TOPALOV - ANAND, UNARES 1999 287

Once again, White had a plausible alternative, namely 26 cxd4 l:txd4 27 'l'e2, and now:

I) 27 .. . 'i'd6 28 1he3 (28 �g2? loses after 28 ... 1i'd5+ 29 �h2 l:tdl) 28 ... :g4+ (28 ... l:tdi + 29 �g2 l0e5 30 'l'e4 �3 is unclear) 29 �fl ldi+30Wt"21lfc2+ 31 ..td2 l:td4 32 :dl l'xd l 33 16'e8+ �g7 34 ..tc3 with a murky position.

2) 27 ... l:h4 and now: 2a) 28 �xg5? 16'f4! is winning

for Black. 2b) 28 ..txe3 l:txh3 29 �g2 l:th4

30 if2 l:tf4 31 ..tg3 :.n and White has some compensation due to the exposed black king. ·

3) 27 ... 1ff5 and now: 3a) 28 .i.xe3 l:tg4+ 29 �h2 1i'e5+

30 �hi l:tg3 ! 31 1i'fl 1i'e6! ! (Fritz suggest this;31 ...:.Xe3 32 1ff8+

�h733 'i'f7+ 16'g7 34 1i'f5+ �h8 is better for Black but not winning) 32 lOg! l:xe3 and wins.

3b) 28 1i'xe3 is a tough nut to crack:

3bl) 28 .. .l:td1+ 29 �g2 (29 �h2? loses to 29 ... l0e5 !) 29 ... 1i'd5+ 30 1i'f3! (30 �f2? l0e5 31 16'c3 l:ld3 32 1i'c8+ �g7 33 1i'c7+ Wf6! 34 1i'b6+ �f5 is decisive) 30 ... g4 31 16'xd5 gxh3+ 32 �xh3 l:txd5 33 b3 with a draw.

3b2) 28 ... l:te4 29 16'c3+ (White loses after 29 1i'f2? 1i'g4+ 30 16'g2 l:tel+ 31 �h2 l:te2 32 l0f2 16'h4+) 29 ... l0d4! 30 l0f2 (the only move) 30 .. .l:tel+! 3 1 �g2 (3 1 1i'xe1?�+ 32 �fl t0xe1 33 �xe1 g4! gives Black a clear advantage) 31 ... 11'd5+ 32 �h3 leads to a draw.

3b3) 28 ... l:tg4+ 29 �h2l0e5 (not 29 ... l:te4? 30 1i'f2 1i'e5+ 31 �h1 l:th4 with a draw) gives Black a very strong attack, but as a final test I turned on Fritz to see whether it could defend. It found 30 1i'b6! (30 lOg! loses on the spot to 30 ... l:te4! 31 16'c3 l:tc4 !) 30 . . . l0f3+ 31 �h i, but Black still wins by 31 ...1i'd5 ! (this is the key - Black sets up a mating battery) 32 1i'xh6+ �g8 33 1i'g6+ �f8 34 1i'f6+ �e8 35 1i'h8+ �d7 36 1i'h7+ �c8 37 1i'c2+ �b8 38 1i'b3 l:tc4 ! 39 ..tf4+ gxf4 40 t0xf4 16'e4 and White has no defence.

Thus 26 cxd4 is inferior to the text-move.

26 27 1i'e4 28 �g2?

I was a bit confused when I saw this, but then I saw what to do. The alternative was 28 ..ta3 !, and now:


1) 28 ... dxc3 29 llfl •e6 30 l:tf5 (30 �b4 c2! 3 1 �c3 l:tdl 32 'i'xc2 l:txfl+ 33 �xfl ... xh3+ 34 �gl 'i'g4+ 35 'i'g2 'ifdl + 36 �h2 'i'xh5+ 37 �gl is a draw) and now:

Ia) 30 ... l:ld 1 + is the obvious continuation, but White can defend by 31 �g2 e2 32 l:txe5 el• 33 l:txe6 l:td2+ 34 �3 'i'fl+ 35 �g3 l:td3+ 36 'i'xd3 ... xd3+ 37 �h2 �g7 38 l:te7+ �6 39lLlf2 'i'f3 (39 ... 'i'd240 �g3 'i'f4+ 41 �g2 'i'd4 42 lLle4+ �5 43 lLlg3+ �f6 draws, while 39 ... 'i'c2? 40 �g3 �5 4l lLlg4 even wins for White) 40 ll:,e4+ �5 41 lLlg3+ �6 42 lLle4+ with another draw.

lb) 30 ... l:td5 ! 3 1 �fl (31 �hl e2 !, 31 �g2 'ilfc6 32 �g3 e2 and 31 �b4 e2 32 �xc3 l:tdl+ 33 �g2 e1'i' favour Black) 31 . ..'ilfc6! 32 l:tfll+ �g7 33 lLlxg5 l:tdl + 34 �e2 l:td2+ 35 �xe3 'ilfb6+ 36 �4 hxg5+ 37 �xe5 is a draw.

2) 28 ... d3 29 l:tfl 'ilfe6 30 'ilfxe3 •g4+ 31 �2 (3 1 �h l? 'ilfe2! wins

for Black) 31 ...... 5+ 32 �g2 Wg4+ 33 �f2 is a draw.

3) 28 ... lLlf3+! is the best chance, and now:

3a) 29 �hi lLld2 ! (29 ... Wc6 30 •xc6 bxc6 31 cxd4 e2 32 i.b4 defends) 30 •xb7 (30 1fg6 'i'xg6 31 hxg6 d3 is also very good for Black) 30 ... 'Wf5 ! 31 'i'g2 d3 and White is in serious trouble.

3b) 29 �g2! ltld2 30 'i'g6 (not 30 'Wxb7? ¥f5 with a large advantage for Black) 30 ... 'ifxg6 31 lug6 d3 (3 l ... dxc3 32 �e7 l:tc8 33 i.f6r �g8 34 l:tcl c2 35 .i.b2 defends) 32 �c 1 lDxb3 33 axb3 d2 34 .ixd2 l:txd2+ and Black is slightly better, although a draw is more likely than a win for Black.

28 e2! (D)

29 .i.xgS White is in serious difficulties,

whatever he plays: l ) 29 9xe2 d3! 30 'i'f2 'i'c6+ 31

�g3 ¥e6! 32lDgl d2 33 .i.xd2 •g4+

TOPALOV - ANAND, LINARES 1999 289

34 �h2 'i'xh5+ 35 �g3 �g4 ! wins for Black.

2) 29 J..b2 :C8 30 1rxe2 dxc3 31 ..ia3 (31 J..c l 'i'c6+ 32 �h2 'ire6 33 'i>g2 'i'd5+ 34 �gl %lf3 is winning for Black) 31 ... 'i'c6+ 32 �h2 �f3+ 33 �hl l:td8 34 'i'fl (34 'irg2 g4) 34 ... .1:1.d2 with a decisive attack for Black.

3) 29 J..d2 l:tf8! (29 ... d3 30 c4 �f3 31 1rxf3 'i'xal 32 �f2 is less clear) 30 �gl (30 1rxe2 1rc6+, 30 lxd4 l:td8! 31 'i'e3 l:txd2! and 30 cxd4 'ilffl+ 3 1 �h2 �f3+ 32 �g3 �d2 are all hopeless for White) 30 ... d3 (30 ... 1rf2+ 3 1 �hl �f3 32 �xf3 'ilfxf3+ 33 1rxf3 l:txf3 34 l:tel dxc3 35 J:be2 cxd2 36 l:txd2 gives Black an extra pawn in a rook ending, but White can probably draw) 3l .l:l.e1 (31 'l'd4 �h7 ! is awkward, as 32 'i'e4+ �g8 33 Wd5+ l:tf7 only helps Black) 3 1 ...'ii'f2+ 32 �hl

'ii'g3 33 'i'g2 'ii'h4+ 34 'ii'h2 'ii'e4+ 35 'l'g2 �f3 36 �xf3 l:txf3 37 �gl

'ii'f5 wins for Black. l9 hxgS (D) 30 cxd4 'i'c6 31 dS

The alternative 3 1 'ii'xc6 �xc6 32 �gl .l:l.e8 33l:tel �xd4 also wins for Black.

31 ••• 'ii'xdS 3l. .. .l:l.xd5 32 l:te1 doesn't lead

any where for Black. 32 'i'xd5 l:txd5 33 l:te1 l:td2 34 <M2! l:txa2

35 .l:txel?! This makes life easy for Black,

but 35 �e3 m 36 l:txe2 (36 �f2 �g7 37 l:txe2 l:txe2+ 38 �xe2 �h6 also wins for Black) 36 ... l:txe2+ 37 �xe2 �g7 38 �f3 �h6 39 �g4 a5 40 l!:lf2 b5 41 lLle4 a4 42 bxa4 bxa4 43 lllc3 a3 is also decisive.

35 ••• l!:ld3+ 36 �e3 l:txe2+ 37 �xe2 g4! (D)

A nice intermezzo, avoiding the trap 37 ... lllf4+ ?? 38 lllxf4 gxf4 39 �3 with a draw.


38 itJgS 39 �e3

itJc1+ itJxb3 (D)

40 h6 aS 41 Wf4

41 �e4 g3! is also lost foc White. 41 llJd4!

Taking away the f5-square. 42 �g4 a4 43 �bS

Or 43 lDe4 a3 44 itJc3 b5. 43 ... ltX6

It's still not toolate to be careless - after 43 ... a3?? 44 'it>g6 White would even win. However, after the textmove 44 �g6 ltle5+ 45 'it>f5 a3! 46 �xeS a2 wraps it up.

0-1

In Linares I made heavy weather of most of my games, and the Svidler and Topalov encounters were no more than isolated flashes of brilliance. In some events the moves just flow, but at Linares everything seemed an uphill stroggle. After this mediocre result, I went on to play in Dos Hermanas, which was an unmitigated disaster. Matters did not improve much during the rest of 1999, though it has to be said that I played relatively little. However ,matters then took a tum for the better. During 1999, I had spent a month and a half preparing for Kasparov, and this work proved a useful foundation foc my play in 2000. In general, I have found that work you do always pays off in the end, although perhaps not in the game you want or the tournament you want.

The new year started with an immediate success - I won the world blitz championship in Warsaw. After this I went on to the traditional tournament at Wijk aan Zee. Although I perhaps made too many draws in this tournament I was still quite satisfied with my joint second place, especially after the troubles of the previous year. The following game was my best of the event, and is an object lesson in how to play against the Stonewall Dutch.

Game 54

V. Anand - P. Nikolic Wijk aan lee 2000

Dutch Defence

1 d4 rs 2 g3

In the Dutch Defence, it's quite handy to delay developing the g)knight because it can be advantageously placed on h3 in some lines.

2 .•• li::l£6 3 .i.g2 e6 4 c4 dS

Predrag goes straight for the Stonewall, a line in which the knight is indeed best placed on h3.

s li::lh3 c6 Black's aim is to set up a solid

chain of pawns in the centre. While it is not easy for White to break this down, Black's pawn-structure is rather inflexible.

6 0-0 .i.d6 7 ,..c2 0-0 8 li::ld2 .i.d7 9 li::lr3 .i.e8 (D)

A standard idea in the Stonewall this bishop hopes to emerge via h5.

10 .i.£4! I had seen various games where

White goes li::lf4-d3 followed by .i.f4 and wondered why he couldn't save time by developing the bishop directly.

10 h6

Black intends to take advantage of the position of the knight on h3 by pushing his g-pawn.

11 'Wb3! White forces Black to make as

many concessions as possible in order to develop his pieces.

11 b6 Black must weaken his queenside.

If he plays I l ... lln then he allows li::le5 with gain of tempo (before or after .i.xd6 ). Also, defending the pawn from f7 means that Black can't develop with ... lC!bd7.

l l ...g5 is an interesting try. After 12 .i.xd6 'Wxd6 White can continue:

I) 13 lC!e5 enables White to keep a safe edge.


2) 13 'i'xb7 is more ambitious, but also more double-edged. After 1 3 ... g4 White has:

2a) 14 liJf4? gxf3 15 exf3 (15 .ixf3 .if? 16 'i'xa8 'i'c7 17 cxd5 exd5 and the queen does not escape) 1 5 ... .if7 ! ( 15 ... .id7 16 c5!) 16 'i'xa8 ( 1 6 c5 'I'd? 17 'ifxa8 liJa6 also favours Black) 16 ... 'i'c7 17 cxd5 exd5 1 8 l:acl liJfd7 followed by 19 ... liJb6 winning the queen.

2b) 14 'i'xa8 gxh3 and now: 2bl ) 1 5 .ixh3 file? (again it's a

bit tricky - how does White extricate his queen?) 16 llfcl (after 16 a4 liJbd7 17 cxd5 liJxd5 the white queen is doomed) 16 ... �d7, followed by ... liJa6, and White will have to surrender his queen.

2b2) 15 'ifxa7 ! (it is more important to save the queen than the g2-bishop) 15 ... hxg2 16 llfc l gives White the advantage. He has considerable pressure on the queenside and a rook and two pawns for a bishop and a knight - and that is being generous in counting the g2-pawn as a 'live' pawn.

Therefore White has the pleasant choice between line ' 1 ' and line '2b2'.

12 mct A classic case of the problem of

which rook to play to a particular file. With hindsight, l:lacl would have increased the force of a later f4 to such an extent that Black might not even have gone in for ... g5.

12 ... �e7 (D) 12 ... g5 13 �xd6 1Wxd6 14 liJe5

gives White a clear advantage.

13 cxdS! This is the accurate moment to

take on d5, when Black can't recapture ... cxd5.

13 ... liJxdS 13 ... cxd5? loses to 14 !i.e?, while

13 ... exd5 14 liJe5 prevents the development of the b8-knight

14 !i..d2 White can already speak of a clear

advantage since the 'Stonewall' has been breached.

14 ... g5 Sealing the h3-knight out of play,

but potentially weakening Black's kingside.

IS liJeS aS! A good move. The weaknessofc6

makes ... �d7 impossible, so Black has to find another way to activate his rook.

16 e4

ANAND - NIKOUt, WJJK AAN ZEE 2000 293

Opening up lines and chipping away at the pawn-wall in front of Black's king.

16 fxe4 17 .be4 (D)

17 D.a7 17 ... a4 is a more active possibil

ity. After 18 'iid3 a3 19 b3 the critical line is 19 ... ll:la6 20 ll:lxc6 (if White wants to avoid complications then 20 'ife2 is safe and strong) 20 . . .. bc6 21 l:l.xc6 ll:lab4 22 .txb4 ltlxb4 23 'ifc4 l:l.f6, and now:

1) 24 l:l.xe6? b5 25 l:l.a6+ bxc4 26 l:txa8 'ifxa8 27 .txa8 cxb3 28 axb3 :a6 and the a-pawn is too dangerous.

2) 24 l:l.c7 b5 25 'ifc3 is not very clear as the attack on a2 makes it hard todevelop the a1-rook.

3) 24 l:l.xb6! l:l.c8 25 l:l.xb4 l:l.xc4 26l:l.xc4 is promising for White even though the knight is out of play for the moment; e.g., 26 ... 'if aS 27 l:l.d1 or26 ... .td6 27 f3, followed by ll:lf2.

18 f4!

Further eroding the enemy king's defences.

18 ... gxf4 19 �b1

Since Black can hardly take on g3 (19 ... fxg3?20 'ifxg3+ �h8 21 ll:lg6+ gives White a decisive attack), White keeps his options open as to how to recapture on f4.

19 .tf6 Clearing the way for the rook to

switch to the g-file in case White plays gxf4.

20 ll:lxr4 Therefore White takes the chance

to activate the h3-knight. 20 'ild6 (D)

21 ll:lrg6 Forcing the exchange of a defen-

sive bishop. 21 22 ll:lxg6 23 .tf4!

.bg6 l:l.fT7

Black must take the bishop, but this both opens the g-file and ties


Black down to the defence of the c6-pawn.

23 �4 24 gxl'4 .i.g1 (D)

The greedy 24 ... .i.xd4 25 'ilfh3 .i.xb2 is punished by 26 l:tdl ! .i.d4 27 1rxh6, when White has too many threats against the poorly defended kingside.

25 'li'h3 Moving the rook to gl should

come later, in order to keep the enemy knight fixed on b8.

25 •.• l:t'6 26 l:tc3!

Playing 26 l:!.c3, followed by l:tgl and l:tcg3, doubles rooks just as quickly as 26 l:tgl , followed by l:!.g3 and l:!.ag I , but it fixes the knight on b8 longer.

26 ·- Wfxd4 Black grabs a pawn, but the game

will be decided by White's attack along the g-file.

27 11Fg2 l:td7

Threatening 28 ... 'I'd I+, so 28 Zlg3 would now be a blunder.

28 l:tgl White's major pieces slot neatly

into place on the g-file. 28 ... b5

There is little Black can do. 29 :gJ 'fka7 30 ltJeS l:te7 (D)

31 'fid2 White has a choice of winning

lines. The text-move aims to penetrate along the d-file; 31 Zlg6 is also sufficient, while 31 .i.xc6 foUowed by 32 .i.d7 is perhaps most brutal.

31 ... 1lc7 32 l:td3

The rest is fairly simple. 32 ... lle8 33 :d6 cS

33 .. . �h8 loses to 34 'i'g2!, fol· lowed by 35 �g6+.

34 lDg4 :m 35 ltlxh6+ lilh8 36 ltlg4 :l.d8

:r1 'i'g2 38 'ii'b3+ 39 'i'h7+ 40 ..ig6+

ANAND - NIKOUt, WIJK AAN ZEE 2000 295

l:!.xd6 �g8 'M7 1-0

As usual at Wijk aan Zee, the public were able to vote on the best game of the day. This game won the public prize for round 2.

During 2000, in addition to traditional events at Wijk aan Zee, Linares and Dortmund, I played in some rapid events and took part in an 'advanced chess' tournament at Leon, in which the players could use a computer during the games. Adding in blitz and blindfold events, 2000 offered quite a variety of chess activities. All these forms of chess present different challenges, and they add the spice of variety to what would otherwise be a somewhat monotonous progression of tournaments. For professional players, it is very important to keep the creative juices flowing, and playing different forms of chess is a good way to achieve this. An additional point is that these different forms of chess add to the popularity of the game. You could argue that if Frankfurt, for example, didn't organize a rapid event then they might organize a classical tournament, but I think it is the other way around - if they didn't organize a rapid tournament then they wouldn't have a chess tournament at all. I think that if you can sell a particular type of event to a sponsor, then you should just do it, even if the result if a proliferation of different types of chess.

Playing against computers is a more marginal case. At Dortmund 2000, a computer took part in the top event and scored 50%. It is, of course, possible to use the same argument - that having a computer in a tournament might attract sponsor and enable an event to take place which would otherwise be impossible. However, I have some reservations about this and I think that mixed human/computer events should be solely human vs computer events in other words, the players shouldn't be expected to play a human one day and a computer the next. Humans can't switch their styles this way, and without wishing to sound too much like making excuses, I think that an event such as Dortmund will automatically create good results for the computer.

Also, I think that it is necessary to look at the rules for human vs computer games. It can't be right that a computer can access a vast opening database at the boars.Since the computer also has an array of endgame databases at its disposal, the battleground has essentially been narrowed to the middlegame, which makes it very tough for the human player. It may well be that in I 0 years the computer will win regardless of any rules or reslrictions, but just at the moment, when human vs computer contests are still interesting, we have


to ask whether the rules are really fair. The problem is that the current line are a hangover from the days when humans could beat computers with their eyes shut, and at that time nobody cared whether the computer had an openings book. Now the situation is very different, and the rules deserves a further look.

If they wanted to, humans could probably improve their results against computers with three or four months of special training. Your though processes need to be completely different when playing computers; you have to be much more alert and focussed tactically, while at the same time you can take a more relaxed approach to strategic matters. However, at the moment the infrequency of such events would hardly justify such an expenditure of time.

Linares 2000 was another disappointment However, soon after this it became clear that the match with Kasparov wouldn't happen, since we were mable to agree terms for the contract.l was disappointed, but at least the uncertainty had gone. I felt quite motivated to do well in the tournaments that followed and two months of rest (April and May) had the required effect My play started to show the old freshness again. The recovery in my play gathered pace throughout the year, and you could see it not only in terms of results but also in the style of the games. During my best years, 1998 and 2000, my games had a creativity and vigour which was generally lacking in my games from 1999. As 2000 progressed the successes mounted-I won the advanced chess event in Leon and the Frankfurt rapid, and I was joint first at Dortmund. In August I won the FIDE World Cup in Shenyang, and the following game from this event saw me entering the Najdorf labyrinth again

Game 55

V. Anand - A. Khalifman FIDE World Cup, Shenyang 2000

Sici l ian, Najdorf

1 e4 c5 A surprise. However, since the

Najdorf was so topical in the qualification tournament at Polanica Zdroj, it was to be expected that some other players might want to join the discussion.

2 it:lf3 d6 3 d4 cxd4 4 it:lxd4 it:\r6 S it:lc3 a6 6 i.e3 e6

This line is all the rage now. Loek van Wely is, as always, leading the charge, but Gelfand has joined in.

7 f3 7 g4 was played a lot at Polanica

Black had no problems theoretically speaking and Shirov's win over Van Wely was decided in the middlegame.

7 ... bS 8 g4

Both of us are heading straight for the main line. It is very difficult for

of us side to deviate, so I guess both

of us were already thinking about move 16.

8 9 'i'd2

10 0-0.0

h6 it:lbd7 i.b7

11 h4 b4 12 lDa4 'i'aS 13 b3 it:lcS 14 a3 (D)

14 l:tc8 This move was introduced in the

game Anand-Gelfand, Amber rapid, Monaco 2000. Previously 14 ... lDxa4 15 axb4 'ilc7 16 bxa4 had been played, but the practical results from this position until then had favoured White. A few months later, Gelfand went on to resuscitate that line as well!

15 'i'xb4 'i'c7 16 �bl!

16 lDxc5 has been tried a few times, but without too much success.


My second Ubilava and I spent some time looking for alternatives and pretty soon we discovered 16 �b I .

16 li:lcd7 Khalifman played this after a long

thinJc. There are various other possibilities for Black. After 16 ... d5 17 li:lxc5 .ixc5 18 1Va4+ li:ld7 19 b4 the position is unclear, while in a later game de Ia Riva-Van Wely, Zonal tournament, Mondariz 2000 Black played 16 ... li:lfd7 and went on to win.

17 1i'd2 After 17 'ifc4 d5 18 'ifxc7 lb.c7

1 9 exd5 li:lxd5 Black regains his pawn, since 20 .te l is answered by 20 . . . <t)7b6!. Returning the pawn by 20 .id2 .ixa3 Ieads only to equality.

17 -·· dS 18 .th3

In this line White usually has to meet ... d5 with .ih3, aiming for a breakthrough by g5-g6 and possibly a sacrifice on e6. The fact that the queenside looks different doesn't change that!

18 19 gS 20 hxgS 21 fxe4?!

dxe4 hxgS li:ldS (D)

21 g6? is wrong due to 21. ..l:txh3! (after my game with Ljubo in Buenos Aires 1994, I won't forget this resource - see Game 26, note to Black's 17th move, for more about this) 22 l:txh3 li:lxe3 23 'ifxe3 e5 and Black is clearly better.

However, the correct line was 21 .ixe6! l:l.xhl 22 .ixd7+ 'i'xd7 23 l:l.xh I exf3 24 .if2, as pointed out by Ubilava and Mikhalchishin. Material is equal, but Black's king remains exposed.

21 ... 22 'ifxe3

Now Black's excellent knight on e5 and pressure against the queenside give him good compensation for the pawn.

23 l:thfi Round about here I was having

second thoughts about this position. What if Black doesn't allow the sacrifice on e6?

23 il..xa3 This move doesn't lose, but it

does allow White to stir up dangerous complications. Black should have considered 23 ... g6 (when 24 J:f6 is answered by 24 ... .ie7!) if only because his position is easier to play than White's.

24 g6

ANAND - KHAUFMAN, WORW CUP, SHENYANG 2000 299

No exclams or question marks, since after any other move White is lost.

24 ltlxg6 25 �xe6

I had seen that White has a dangerous attack, but hadn't realized how strong it was.

25 fxe6 (D)

26 ltlxe6 26 'l'g5 is less accurate: I ) 26 ... :Jt6 27 ltlxe6 'l'e7 28 'l'g4

quite dangerous for Black, but he can escape by 28 .. ..:1.c6 (28 ... l:l.c7? 29 l:d8+ 'l'xd8 30 ltlxd8 �xd8 31 l:l.d1+ �e8 32 'l'g3 l:l.c6 33 b4 J.xb4 34 '1'b8+ J.c8 35 'l'xb4 wins for White) 29 l:l.d8+ 'it'xd8 30 ltlxd8 �xd8 31 l:l.d1+ �c7 (31...l:d6 32 �a2 J.b4 33 l:xd6+ �xd6 34 ltlb6 M 35 'l'xg7 :Jt7 36 'l'f6+ and only White has winning chances) 32 l:d7+ �b8 33 l:d8+ J.c8 34 ltlb6 ltle7 35 llX:4 and I don't think White has winning chances.

2) 26 ... �xe4! is the best defence, aiming for a counterattack against White's king. After27 ll'lxe6 (D) (27 l:l.de I �e7 28 'I' d2 l:h2! wins for Black) Black has a choice of good lines:

2a) 27 ... �xc2+ 28 �a2 �xb3+ 29 �xa3 �xe6 30 'l'xg6+ �f7 and White's king is now the more exposed.

2b) 27 ... 'it'e7 28 'l'g4 �xc2+ 29 �a2 (29 �a! is met by 29 ... l:l.h4! because 30 ltlxg7+ 'l'xg7+ is now check!) 29 ... J.xb3+ 30 �xb3 l:l.b8+ 31 �a2 (3 1 �c2 l:h2+ is also decisive) 3 1 ...l:l.h2+ with a winning position for Black.

26 ••• 'l'e7? 26 ... 'l'xc2+? is also wrong as 27

�al is winning for White. The only move is 26 ... '1'e5, which I intended to meet with 27 'itb6 J.xe4 28 l:l.d8+ �e7 29 'l'a7+ l:l.c7 30 ll'lxc7 �xd8 when (so I thought) Black couldn't survive. However, back in the hotel

300 VISIIY ANAND: MY BEST GAMES OF CHESS

our German friend (Fritz) showed us that this line should be a draw! After 26 ... '1'e5 27 'il'b6 (27 'l'a7 'l'xe6 28 'l'xb7 liJe7 29 tbb6 'l'c6 defends) 27 ... .i.xe4! White can play:

1) 28 .fu.g7+ �e7 29 lt:lf5+ .i.xf5 30 l:l.del .i.e4 3 1 l:l.xe4 (31 'i'xg6 .i.xg6 32 l:l.xe5+ �d8 is distinctly better for Black) 3 1 ... 'l'xe4 32 'l'f6+ �d7 33 l:l.d1+ 'iti>e8 and Black defends.

2) 28 l:l.d8+ 'iti>e7 (D) with another branch:

2a) 29 tbf8 and now: 2a1) 29 ... l:l.xd8? 1oses to 30 'I'a7+

Wd6 3 1 l:l.dl +. 2a2) 29 ... .i.xc2+ 30 'iti>a2 tbxf8

(30 ... l:l.xf8?! 31 l:l.fxf8 favours White, while 30 ... l:l.xd8 31 1i'b7+ 'iti>d6 32 'i'xa6+ is perpetual check) 3 1 l:l.xc8 tod7 32 'l'xa6 l:l.xc8 33 'I' xeS .i.xb3+ 34 Wxb3 'in>5+ 35 Wxa3 'l'xfl 36 tbc3 with a drawn ending.

2a3) 29 ... tbxf8! 30 l:l.xc8 tbd7 31 'i'xa6 l:l.xc8 32 'l'xc8 'l'b5! 33 'ilfc4

'iW xc4 34 bxc4 with the better ending for Black.

2b) 29 'ilfa7+ l:l.c7 30 tbxc7 (after 30 l:l.e8+, 30 ... Wxe8 31 liJxc7+ �d8 transposes to variation '2b2',while 30 ... l:l.xe8 31 tbxc7 .i.xc2+ 32 �a2 l:l.f8 33 tbb5+ Wd8 34 'i'a8+ We7 is perpetual check) and now:

2bl) 30 ... l:l.xd8 3 1 liJd5+ �e6 32 'iWn+ Wd6 33 '6'c7+ ltixd5 (33. . . � loses to 34 'i'c6+ 'i'd6 35 liJc7+ �e7 36 'i'xe4+ tbe5 37 l:l.f5) 34 liJb6+'iti>d4 35 'ilfxd8+ and Black will lose his queen.

2b2) 30 . . . Wxd8 31 l:l.d1+ .id6 with a further branch:

2b21) 3 2 tbxa6? l:l.hl ! (the tempting 32 ... 1Lxc2+ 33 �xc2 l:l.h2+ 34 'iti>b1 'ilfe4+ 35 �a1 l:l.d2 36 1rb6T �e7 37 l:l.xd2 'i'e 1 + 38 Wa2 'l'xd2+ 39 tbb2 is a draw, but 32 .. . :b2 is also good for Black) and the attack collapses since Black's centralized pieces control too many squares.

2b22) 3 2 liJe6+ 'lfxe6 33 liX:5 .i.xc2+ 34 �xc2 'iVf5+ 35 �I � 36 l:l.xd6 l:l.h1+ 37 l:l.d1 will be a draw.

2b23) 32 lbc5 .i.xc2+ (32 ... .if5? 33 tbb5 ! axb5 34 tbb7+ We7 35 tbxd6+ favours White after 35 ... .id7 36 tOeS+ or 35...�6 36 'fin+ �g5 37 fuf5) 33 Wa2 .i.xb3+ (33 ... .if5 34 l:l.xd6+ 1fxd6 35 liJb7+ Wxc7 36 tbxd6+ 'iti>xd6 37 •xg7 is another draw, while 33 ... .i.xd1 34 liJ7 e6+ 'ilfxe6 35 ltlxe6+ 'iti>c8 36 lild4 is about equal) 34 'iti>xb3 (34 �a3?


i.xdl ! wins fey Black) 34 ... 1lh3+ 35 �a4 llh4+ 36 �b3! (36 �a5? ilc3+ is winning for Black) and the complications peter out to perpetual check.

27 'ii'b6! (D)

Now White breaks through.

Or: 27 ltlf8

I) 27 ... l:th6 28 l:!.d8+ llxd8 29 lix7+ 'ffxc7 (29 .. .'�d7 30 'ii'xb7 �d6 31 'l'b6+ 'it>d7 32 l:!.dl+ wins since 32...�c8 33 'ffxa6+ �xc7 34 'i!Va7+ mates) 30 1!hc7 lld7 3 1 'i!Vb8+ �e7 32 1'g8 and White wins.

2) 27 ... ltle5 28 l:td8+ llxd8 29 ltlc7+ �d7 30 'ffxb7 'i!fg5 31 ltlb6+ (better than 3 1 ltlxa6+ �e8 32 ltlc7+, when 32 ... 'it>e7 hangs on for a draw) 31...�d6 32 ltlbd5! ltlc6 (32 ... 'ii'g6 33 �b5+ mates, while 32 ... 1lh6 33 'i!Vxa6+ �d7 34 •xa3 also wins for White)33 'ifxa6 il..c5 (if White is allowed to take the bishop,then he has a crushing attack for no sacrifice) 34

ltlb5+ and the c6-knight falls, with a winning position for White.

28 lldB+ Not 28 ltlxf8 llc6 29 ltlg6 llxb6

30 ltlxe7 lte6, which only leads to a draw.

28 ••• lhd8 29 llJc7+ 'fllxc7

29 ... �d7? loses immediately to 30 'ii'xb7.

30 11xc7 lld7 (D)

31 1l'b8+ 3 1 11t'e5+ is also strong; fey exam

ple, 31 ... �d8 (after 31 ... il..e7 32 ltlc5 White wins material straight away) 32 ltlb6 i.d d6 33 il 5+ c7 (a 33 ... 1le7 34 lldl White picks up the bishop) 34 'i!Va5 and Black must surrender material.

After the text-move, the queen is extremely powerful and Black must also contend with the possibility of ltlb6.

31 •.• �e7 32 1ifeS+ ltle6


33 .:tgl 'lif7 The main alternative is 33 ... :.gs,

and now: J) 34 lbb6 ..id6 (after 34 ... :.d6?

35 lbc4 �b4 36 .l:g6 White should win) 35 'i'f5 (35 'i'h5 is unclear after 35 ... l:tdd8 36 lbd5+ ..ixd5 37 exd5 lbf4, but not 35 ... �c5? when 36 :.n is s•ong) 35 ... g6 (35 ... ..ih2? loses to 36 l:tg6!) 36 l:lxg6 l:txg6 37 'i'xg6 :.c7 and Black is hanging on.

2) 34 l:tg6! l:td6 35 'lia2 (35 lbc3 'lid7 36 lbd5 �xd5 37 exd5 lbc7 38 lbg7+ l:txg7 39 'i'xg7+ 'ilrc8 is only a draw) 35 ... �c1 (D) (the only move) and now:

2a) 36 lbc5 �c8 37 lbxe6 �xe6 38 'irc5 'lid7 39 e5 �xb3+ 40 cxb3 :.xg6 41 'l"d5+ 'lie7 42 'irxg8 �h6 looks like a draw.

2b) 36 :Xe6+ :Xe6 37 'irc7+ �6 38 'i'xb7 is unpleasant for Black. The vulnerable black king means that White's queen has the edge over Black's rooks.

2c) 36 &3 ..if4 (36 ... 'ilrd7 37 lbd5 and 36...�c6 37 lbd5+ 'ilrd7 38 c4 are winning for White) 37 lhe6+l:l.xe6 38 'irxf4 :.c8! 39 'irg4! (not39 'lib2 'lieS and now White can't play 40 lbd5 because of 40 ... .ixd5 41 exd5 lle2 1atching on to the c2-pawn) 39...'ilrf7 40 1Vf5+ 'ilre7 41 �5+. followed by 42 c4, with a large advantage for White. In view of Black's exposed king, his rooks will have a hard time against the queen and actively placed knight. If Black exchanges on d5, then White obtains two connected passed pawns.

34 lbbei muiB 35 'ilra2 (D)

I didn't want to allow ... l:ld1+, but 35 1Vf5+ 'lie7 36 :.g6 :.dl + 37 �a2 J:l.8d6 38 lbc4 should also win.

35 ••• .if8 36 �d7

36 1Vf5+ 'ilre7 37 l:tfl would have been more precise, with a quick win.

36 :Xd7


37 'i'fS+ By now White has a choice of good

lines; for example, 37 l:l.fl + <j;e7 38 l:l.xf8 �xf8 39 'l'xe6 is also decisive.

37 rl;e7

38 :n �c8 39 fff7+ �d6 40 e5+ 1-0

Afler 40 ... <j;d5 41 lldl+ White wins a piece, so Black resigned.

Afrer Shenyang, I won a rapid event in Corsica and I came to the FIDE World Championship very motivated. To play 21 games in an event of this strength without a single loss says it all - I was on top form. Apart from a scare against Khalifman, my play was convincing throughout. A pattern developed where I would win with White and draw with Black. While there is probably an element of coincidence here, I think my black openings have become more solid over the past few years.

Game 56

V. Anand - M. Adams FIDE World Championship, New Delhi 2000

Ruy Lopez, M0ller

1 e4 eS 2 �f3 �6 3 �bS a6 4 �a4 �f6 s o-o �Q 6 c3 bS 7 �c2 d6 8 a4 �g4 9 h3 �hS

10 d3 0-0 1 1 �bd2 b4

After 1 l ...d5 12 axb5 axb5 13 l:l.xa8 'irxa8 14 exd5 �xd5 15 ll'le4 �b6 16 ll'lg3 �g6 17 ll'lh4 White was slightly better in Gild.GarciaBenjamin, Toronto 1998.

12 aS l:l.b8 (D)

13 g4!?

White wants to bury the bishop on g6. but I spent some lime looking at the sacrifice on g4. The alternative is 1 3 l:l.el 'ireS 14 ll'lfl b3 15 .ibl l:l.b5, with an edge for White, as played in Adams-Benjamin, World Team Championship, Lucerne 1997.

13 ••• �g6

The sacrifice leads to complex play: 1 3...�xg4 14 hxg4 bg4 and now:

I ) 15 �b3 and here: Ia) 15 ... f5 with two lines: Ia!) 16 ll'lxc5? and now: !al l) 16 ... dxc5 17 �b3+ �h8

18 �d5 ( 1 8 �e6 'ireS wins for Black) 18 ... l:l.f6 19 �xc6 l:l.xc6 20 exf5 (20 �xe5 �xdl 2I ll'lf7+ �g8 22 ll'lxd8 l:tg6+ 23 �h2 .if3 is very good for Black) 20 .. .'ti'f6 21 Ael 'irxf5 22 ltJh4 'irh5 is good f<r Black.

lal2) 1 6 ... fxe4 17 ll'le6 (17 .ig5 'ireS and 17 dxe4 �xf3 18 1i'd5+ �h8 are hopeless for White, in the latter case because there is no defence to 19...1i'h4) 17 ... 1i'c8! ( 17...1i'f6 18 �g5 'ilt'g6 is less clear due to 191Dh4! 'ti'xe6 20 �b3) 18 ll'leg5 (18 dxe4 �xf3 19 'lid5 ll'le7 20 1i'c4 l:l.f6! wins) 18...�xf3 19 �b3+ �h8 20 � 'ife8! with a decisive attack.

ANAND - ADAMS, FIDE WORLD CH., NEW DELHI 2000 305

la2) 16 d4! iLa7 and now: la21) 17 exf5? d5 18 1We2 bxc3

(18 ... e4? 19 iLxe4 dxe4 20 1Wxe4 is unclear) 19 dxe5 l:txb3 20 iLxb3 {\d4 21 o!Dxd4 iLxe2 22 o!Dxe2 cxb2 23 .ixb2 1i'g5+ and Black wins.

la22) 17 dxe5 threatens 18 'iid5+, and after 17 . . . �h8 1 8 1i'd3 ! I don't see anything obvious for Black.

I b) 15 ... iLa7! (or even 15 ... bxc3 16 bxc3 L7 !) is stronger. After 16 �g2 f5 White's knight i s misplaced on b3 and will have to return to d2.

2) 15 �g2 f5 1 6 il.b3+ �h8 (D) and now:

w

2a) 17 l:th1 fxe4 18 dxe4 1i'f6, followed by ... o!De7-g6, is awkyard for White.

2b) 17 exf5 and here: 2bl) 17 ... 'iif6 18 iLe6 iLxf5 19

iLxf5 'iixf5 20 ll:le4 'iig4+ 21 ll:lg3 is unclear.

2b2) 17 ... l:txf5 18 il.d5 (18 iLe6 l:tf4 19 iLxg4 l:txg4+ 20 �h1 l:th4+! 21 �xh4 'iixh4+ 22 �g2 'iig5+ is a

draw, while 18 l:th1 d5 ! is unpleasant for White) 1 8 ... 1We8 1 9 l:th 1 W'g6 (19 ... o!De7 20 il.e4 'iig6 21 � comes to the same thing) 20 �fl ll:le7 2 1 il.e4l:th5 ! (2l ...d5 22 iLxf5 'iixf5 23 d4 exd4 24 cxd4 il.h3+ 25 �e1 favours White) 22 l:tg I iLh3+ 23 �e2 'iif7 with a messy position. Of course, it is one thing to reach this position in the calm of your study and another to have it in a practical game when elimination from the world championship is hanging over your head! I might have gone for line '2c' had the sacrifice arisen at the board.

2c) 1 7 iLe6 was the manoeuvre I was hoping would save me. Later analysis bore this out - White is in no danger; e.g., 17 ... 'iie8 ( 1 7 ... 'iif6 1 8 iLxf5 iLxf5 19 exf5 1ixf5 20 ll:le4 is slightly better for White) and now:

2c l) 18 iLxf5 iLxf5 ( 18 ... l:txf5 looks better, but it falls short after 19 exf5 'iih5 20 l:th 1 'iixf5 2 1 ltle4; for example, 2 1...l:tf8 22 ltlh4! and I don't see anything for Black) 19 exf5 l:txf5 transposes to line '2c2'.

2c2) 18 exf5 iLxf5 19 iLxf5 :Xf5 20 ltle4 W'g6+ 21 �hi ! (the key move; 2 1 ll:lg3 l:tbf8 looks fine for Black) 21 . . .l:tbf8 22 ll:lh2 iLxf2 23 W'g4! and White has beaten back the immediate attack, though Black is very much in the game with three pawns for the piece.

14 li:.c4 (D)


Not l4 ltJh4?bxc3 15 bxc3 ltJxe4.

14 bxc3 15 bxc3

Intending .i.e3. 15 ... "i'c8

Aiming for a sacrifice on g4, but Black never manages to execute it. 15 ... ltJa 7!? 16 ..ie3 .1xe3 17 ltJxe3 h6 was a possible alternative.

16 ..ta4 ltJa7 Alternatively, l6 ... ltJd8 and now: l) 1 7 d4 .i.a7 is good for Black as

the e4-pawn is too weak. 2) 1 7 .i.e3 ..txe3 18 ltJxe3 (if in

stead 1 8 fxe3, then l 8 ... ltJe6 followed by ... ltJd7) l 8 ... ltJe6 is the point of ... ltJd8 - Black's knight is heading for f4.

3) l 7 ltJh4! ltJe6 18 ltJf5 'iid8 19 'li'f3 ltJd7 and White is slightly better, but nothing more. Black's position is quite resilient.

17 .i.e3! ..txe3 The sacrifice still doesn't work:

after l 7 . . . ltlxg4 18 .i.xc5 dxc5 1 9

hxg4 '6'xg4+ 20 �hl ..ih5 2 1 �h2 the attack collapses.

18 ltJxe3 (D)

Now White is slightly better. 18 c6 19 'li'd2 'ilc7 20 c4

Intending to disrupt Black's pawns with the temporary pawn sacrifice c5.

20 ... cS? (D) Preventing White's advance but

weakening the d5-square. Normally speaking, one never expects a positional error from Michael. I thought I saw him wince as soon as he made the move - a slip of the hand, perhaps? 20 ... ltJd7! is better, since after 21 ..tc2 Black can go for ... lt:lc8-e7. He is slightly cramped, but his posi· tion is solid.

21 ltJh4? Dubious, as playing the knight to

f5 gives B lack the chance to rid himself of the passive bishop on g6.

ANAND - ADAMS, FIDE WORW CH., NEW DEIRI 2000 307

Also, this is the tempo Black needs to swing his knight to e7. White should have played 2 1 g5 ! lllh5 22 llld5 'ii'dS 23 :fb I , when the critical line runs 23 ... f6 (Black has nothing else) 24 :b6 fxg5 25 :abl ! :as (25 ... l:l.xb6? loses to 26 axb6 lllcS 27 b7) 26 lllxg5 lllf4 27 h4 h6 (it looks like Black has broken out, but White can sacrifice the knight on g5) 2S :tb7! (not 2S lllxf4? l:l.xf4 29 llle6 :tg4+ 30 �l -xh4 31 :bS+ :xbS 32 :txbS+ �f7 33 �2 •hi + 34 �e2 .txe4! and Black is winning) 28 .. .hxg5 29 :d7 •es 30 :bb7 (30 :txg7+ �xg7 31 .txeS :axeS is unclear) 30 ...• e6 31 :xg7+ �hS 32 .td7! •xd5 33 exd5! (33 cxd5 �xg7 and there is nothing clear for White after 34 � �h6! 35 'ifb lllb5mr 34 hxg5 l:l.f7) 33...�xg7 34 hxg5 (D) and now:

1) 34 ... :f7 35 'ifb2 :hs 36 .te6 llle2+ 37 �fl lllf4 38 :xf7+ .txf7 39 �g1 be6 40 dxe6 and White wins easily.

B

2) 34 ... �gS 35 1Wb2 with another branch:

2a) 35 .. .l:l.f7 loses to 36 .te6lllxe6 37 dxe6.

2b) 35 ... lllcS 36 l:l.bS! l:l.xbS (or 36 ... llle2+ 37 �g2 lllf4+ 3S �h2 l:l.xbS 39 •xbS .txd3 40 .txc8 .txc4 41 .. c7 and White wins) 37 •xbS .txd3 3S .txcS .tf5 39 •xd6 .txcS 40 •xe5 should win for White.

2c) 35 ... bd3 36 •b6 llle2+ (or 36 . . . .txc4 37 .te6+ lllxe6 3S •b1 mating) 37 �fl lllf4+ 3S �e1 .txc4 39 •c7 and after a couple of checks Black will have no defence to the threat of .te6+.

3) 34 ... l:l.abS 35 �2 l:l.xb7 (White wins after 35 . . . lllxd3 36 .te6+) 36 •xb7 :f7 37 •xa7 .tf5 3S •xa6 :xd7 39 •c6 llle2+ 40 �g2 llld4 4 1 •cs and the passed a-pawn i s very hard to stop; e.g., 4 Le4 42 dxe4 .txe4+ 43 f3 .txf3+ 44 �h2 :e7 45 a6 l:l.f7 46 •b7 !.

Now we return to the position after 21 lllh4? (D).

JOB VISHY ANAND: MY BEST GAMES OF CHESS

21 1oh8? Too slow. 21 ...ll:lc8! was correct,

and now: 1) 22 g5 ll:lh5 23 ll:ld5 '6'd8 is no

longer so effective since Black can challenge the white knight with ... ll:le7.

2) 22 ll:lhf5 i.xf5 (after 22 ... ll:le7 23 ll:lxe7+ '6'xe7 24 l:l.fbl Black is handicapped by the poorly placed bishop on g6) and now White should be content to keep an edge by 23 ll:lxf5 ll:le7 24 ll:le3, as 23 exf5 ll:le7 24 l:l.fbl (or 24 g5 ll:lh5) 24 ... d5 ! 25 g5 lLlhS gives Black counterplay.

3) 22 ll:lhg2 is the safest move; after 22 ... ll:le7 23 f4 exf4 24 ll:lxf4 White has an edge but cannot claim more.

22 gS! ll:lhS 23 ll:lds WdB 24 J:l.lbl! (D)

White reverts to the correct plan -exploiting the outpost on b6 and the offside a7-knight.

24 ll:lf4

The other possibility is 24 ... l:l.xbl + 25 l:l.xb 1, and now:

1) 25 ... ll:lf4 26 etJxf4 'i'xg5+ (or 26 ... exf4 27 'irxf4) 27 ll:lhg2 exf4 28 l:l.b7 ll:lc8 29 i.d7! (29 'i'xf4 transposes to line '2a') and White has a large advantage.

2) 25 ... ll:lc8 and now: 2a) 26 l:l.b7 (this wins, but it's

quite complicated - however, the variations are very attractive, so it is worth analysing in detail) 26 ... lllf4 27 ll:lxg6+! (even stronger than 27 ll:lxf4 Wxg5+ 28 ll:lhg2 exf4 29 Wxf4 WdS 30 Wd2 ll:le7 31 ltlf4 and White stands well) 27 .. ixg6 28 h4 (D) with another branch:

2al ) 28 ... ll:le6 29 l:l.d7! 'liteS 30 Wd 1 is decisive.

2a2) 28 ... ll:lxd5 29 cxd5 l:l.f4 (or 29 ... ll:le7 30 i.d7!) 30 .id7 l:l.xh4 (30 ... ll:le7 31 .ie6 l:l.xh4 32 'l'b2 ll:lc6 33 dxc6 Wxg5+ 34 <j;>fJ l:l.hl+ 35 <j;>e2 Wh5+ 36 <j;>d2 wins f<r White) 31 .ie6 l:l.f4 32 l:l.b8 'i'xgS+ 33 <j;>f) and Black loses.

ANAND - ADAMS, FIDE WORLD CH., NEW DELRI 2000 309

2a3) 28...ltJh3+ 29 �g2! (not 29 'ii>hl? l:txf2 30 'ire3 'irf8 with problems for White) 29 .. .l:txf2+ 30 'ifxf2 l2Jxf2 (White has only a rook for a queen, but remarkably he still wins) 31 l:b8! h6 (forced, as after 3L.h5 32 ttlb6 �h7 33 l:xc8 'ire7 34 ltJd7 White wins on the spot) 32 ltJb6 �h7 33 J:xc8 'ire7 34 ltJd7! (White cannot pause to capture the knight on f2, since after 34 �f2 hxg5 Black's queen slips out, when it will be impossible to avoid perpetual check) 34 . .. hxg5 35 ltJf8+ �h6 36 ltJe6!! (the remarkable point) 36 . . . �h5 (the only move) 37 �xf2! (now is the correct time to take the knight; 37 l:h8+ ?? is a mistake because Black wins after 37...�g4 38 �xf2 'irf6+!) 37 . . .'1'f6+ (37...�g4 38 R.dl+ �h3 39 hxg5 ! and 37 ... 'ifxe6 38 R.dl+ are also winning for White) 38 �g2 'lxe6 39 R.dl+! (and not 39 l:lh8+?? �g4 40 R.dl+ �4 since here e3 Isn't covered)and White wins the queen

2b) 26 l:lb8 is a simpler win: 26 ... ltJf4 27 ltJxf4 exf4 (27 ... 'ifxg5+ 28 ltJhg2 exf4 29 R.d7! is decisive) 28 'irxf4 (a clearer continuation than 28 ltJg2 f3 29 ltJh4 ilc7 30 l:lb2) 28 ... 'irc7 (28 ... 1Wxa5 loses after 29 '1Vxd6!) and now:

2bl ) 29 l:lbl 'iixa5 30 R.c6?! (30 R.d7 'ifc7 3 1 R.g4 would still be good for White) 30 ... ifc7 3 1 R.d5 ltJe7 32 ltJxg6+ ltJxg6 (32 ... hxg6 33 R.xf7!) 33 'ilfg3 l:lb8 34 l:txb8+ 'ilfxb8 35 R.xf7 a5 and the a-pawn gives Black counterplay.

2b2) 29 l:la8! is simple and crushing, since White can meet 29 ... '1Vb7 by 30 'ilfxd6!.

25 lLlxf4 'l'xg5+ 25 ... l:lxbl+ 26 l:lxbl 'ilfxg5+ 27

ltJhg2 exf4 28 'irxf4 'l'd8 29 l:lb6! leaves White with a clear advantage.

26 ltJhgl (D)

26 ••• exf4?! After this White has an intermezzo

that allows him to keep both rooks.


Black could still have exchanged on b l .

27 J:.b6! Occupation of the outpost brings

Black's a- and d-pawns under fire. 27 ••. J:.bd8 28 'i'xC4 (D)

28 •.• 'fie7 29 l:lab1

There is no rush to take the pawn on a6.

29 ••. ltX:8 30 J:.b7 'fie6

By now there are many ways to win.

31 'figS h6

3 l ...'i'e5 32 h4 and 3l ....xh3 32 iilf4 'fih6 33 'i'xh6 gxh6 34 l:lc7 are also hopeless for Black.

32 'fig3 W'r6 3 3 it:lf4 l'i:Je7 34 'it>g2! (D)

34 l:lxe7? ! ilxe7 35 'i'xg6 l:lb8 is unnecessarily complicated.

34 •••

35 lMS 36 i.dl! 1-0

There is no defence to the threats of 37 i.g4 and 37 l:l lb6, because 36 ... lt:lf6 runs into 37 l:le7. Therefore Michael resigned, giving me the lead in our semi-final match.

Game 57

V. Anand - A. Shirov FIDE World Championship Final (4), Teheran 2000

French Defence

1 e4 Needing only one point from three

games to become World Champion, I decided just to play normally and forget about the score. There was a rest day after the third game, which I spent checking all the openings Alexei plays.

1 .•. e6 During 2000, Alexei and I played

a lot of French Defences. Now he tries it again, based on an improvement he had found over our game in Frankfurt.

2 d4 d5 3 �c3 l0f6 4 e5

In Sydney, 1 went for 4 i.gS, but didn't get much out of the opening.

4 ... l0fd7 5 �el

During 2000, I used this variation quite successfully, gaining two wins against Shirov in Leon and Frankfurt. In contrast, against Bareev in Shenyang I didn't get much from the opening. However, I included this variation in my preparations for New Delli. am felt ready to use it again.

5 cS 6 f4

Against 6 c3 cxd4 7 cxd4, Black can tty to blast White's centre immediately with 7 ... f6. 8 f4 is a possible reply, but this contains a lot of danger for White.

6 ... � Played fairly quickly. Black can

go for 6 ... cxd4, but Alexei was heading for his prepared improvement, so he saw no reason to enter a sideline.

7 c3 ....,6 8 �f3 (D)

8 ... f6 He played 8 ... i.e7 in Frankfurt,

but we soon transpose in any case. One can already make out the contours of the battle; White is trying to


maintain his pawn wedge in the centre ( d4 and e5) while Black will try to breach it ( ... f6, ... c5).

9 a3 iJ..e7 10 h4 0-0 11 l:th3 aS!

The earlier game went I I . ..llla5? 1 2 b4 cxb4 1 3 axb4 lllc4 14 lllg3 and now that the pressure on the centre has disappeared, White can direct his pieces to the kingside (AnandShirov, Frankfurt Rapid 2000). The text-move is much stronger, since by not allowing b4, Black is able to maintain the tension in the centre.

12 b3 Or 12 lllg3 cxd4 (Black tries to

open the centre as much as possible) 1 3 cxd4 fxe5, and now:

1 ) 14 dxe5 lllc5 1 5 lllg5 (15 iJ..d3 lllxd3+ 16 'l'xd3 iJ..d7 17 lllg5 'l'g l + 18 lllfl :C5 is good for Black) 15 ... lllb3 16 'l'd3 (16 iJ..d3 h6 favours Black) 16 ... iJ..xg5 (not 16 ... g6 17 lllxh7) 17 hxg5 g6 18 l:tb1 lllxcl 19 J:txcl J:txf4 defends.

2) 14 fxe5 :Xf3! (a typical exchange sacrifice) 15 gxf3 lllxd4 (the alternative 15 ... '1'xd4 16 f4 lllc5 also gives Black lots of play for the exchange) 16 f4 lllb3 and Black has enough compensation.

12 'l'c7 There are two games featuring

12 . . . 'l'd8, but this move looks much better; for a start, there are possibilities of ... 'l'c3+ in some lines.

13 l0eg1 (D)

13 a4? Alexei wants to sacrifice a piece

to blast open the centre, but his idea falls short. There was no need to take such drastic measures, since Black can simply try to swap light-squared bishops by l 3 . . . b6!, and now:

1 ) 14 iJ..d3?! cxd4 15 cxd4 (15 iJ..xh7+? �xh7 16 lllg5+ fxg5 17 hxg5+ �g8 and Black defends after 18 l:th8+ �xh8 19 'l'h5+ �g8 20 g6 iJ..h4+ 21 'l'xh4 lllf6 or 18 Wb5 llldxe5 19 fxe5 'l'xe5+ 20 lt)e2 i.a6) 15 ... fxe5 16 fxe5 ( l 6 dxe5 lllcxe5 17 fxe5 'l'c3+ 18 �e2 'l'xa1 19 'l'c2 iJ..a6 again defends) 16 ... lllxd4 is very good for Black.

2) 14 iJ..e3 iJ..a6 15 iJ..xa6 :Xa6 looks about equal.

14 b4 fxeS 14 ... b6 no longer works since 15

iJ..e3 iJ..a6 is met by 16 b5. 15 rxeS lLldxeS

In this position, drastic measures are necessary. If White can consolidate his centre, then his kingside

ANAND - SHIROV, FIDE WORW CH., TEHERAN 2()()() 313

attack will be overwhelming; for example, 15 ... cxd4 16 cxd4 liJdxe5 17 dxe5 liJxe5 18 .ltb2.

16 dxe5 liJxe5 17 liJxe5!

The best move. White's knights would keep tripping over each other's toes, so it's useful to exchange one of them.

17 •••

18 '�Vel 'IV xeS+ .ltxh4+?

He played this quite fast, but it's an error. Mter 18 ... '1Vc7! Black follows up with ... e5, which should give him some play. Still, the extra piece should favour White after 19 .lg5.

19 Wd1! (D)

Now Black has to swap queens or make some other concession.

19 ... '1Vf6? Mter 19 ... '1Vxe2+ 20 .ltxe2 .ltf2

21 .i.e3 e5 22 .ltxf2 J:l.xf2 23 J:l.g3 White has a comfortable edge in the ending. The situation reminds me of

one of the main lines of the Slav, where White also has a piece for three pawns (I d4 d5 2 c4 c6 3 liJc3 liJf6 4 liJf3 dxc4 5 a4 .ltf5 6 liJe5 e6 7 f3 .ltb4 8 e4 .ltxe4 9 fxe4 liJxe4 10 .ltd2 1fxd4 l l liJxe4 '1Vxe4+ 12 'IV e2 .ltxd2+ 13 �xd2, etc.). In the current position, Black's centre would be too loose.

20 �f3! 'Wxc3?! 20 ... g5 2 l liJxh4 'IVxfl+ 22 'IVxfl

lbfl + 23 �e2 J:l.xcl 24 J:l.xc I gxh4 25 .l::txh4 is very good for White, as his rooks have all the open files they need!

21 .ltb2 Now White wins a second piece.

21 ... 11b3+ 22 �cl e5 (D)

The only move - otherwise lLie5 wins the queen.

w

23 J:l.xb4 Obviously a pleasant choice - to

win Black's queen or be two pieces up. I saw that after the text-move I


could swap queens and decided not to bother with the alternative. However, taking the queen would have finished it earlier: 23 ll:ld2 .i.xh3 24 ll:lxb3 :t2 (after 24 ... .i.g4 25 'ifxg4 .=:xn + 26 �c2 axb3+ 27 �xb3 c4+ 28 �a2 White just collects all the pawns) 25 'il'b5! .i.g5+ (25 ... .i.d7 also loses, to 26 'il'xd7 .=:xn + 27 �c2 axb3+ 28 �xb3) 26 �bl .i.f5+ 27 .i.d3 axb3 28 bf5 l::txf5 29 bxc5 is winning for White.

23 ••. .i.f5 24 'il'd1 e4 25 111xb3 axb3 26 ll:ld2

Harvesting all Black's central pawns is going to take a while, but basically the evaluation of the position is clear - White is winning.

26 ••• e3 26 .. . c4 27 .i.d4 wins for White.

27 llJr3 27 o!llxb3 .i.g6 28 .i.e2 :n 29

�dl l::txg2 30 l:g4 is also sufficient. 27 l:ae8 (DJ

28 �1?! 28 bxc5 e2 29 �d2 exn• 30

l:xfl wins as well, but there is no reason to return material. However, 28 .i.e2! would have gained a tempo; after 28 ... c4 (28 ... d4 29 .i.c4+ �h8 30 bxc5 d3 3 1 .i.c3 keeps everytbing under control) 29 .i.d4 White can play �b2 immediately.

28 c4 29 .i.e2 .i.e4 30 �1 l:e6 31 .i.cJ 1lg6 32 .1:112 .i.d3 33 hd3 ad3 34 �b2 d2 35 �xb3

By now there are alternative routes to victory; e.g., 35 .=:dt e2 36 l:xd2 :Xf3 37 l::txe2 or 35 .i.xd2 exd2 36 ltd! .

35 ••• :g3 36 �b2

Yes, 36 �c2 is quicker. I guess I was getting nervous!

36 g5 (D)

ANAND - SHIROV, FIDE WORW CH., TEHERAN 2000 115

37 �c2 ltc8 White also wins after 37 ... g4 38

.ie5(38 li)d4 l:lf2 39 l:ldl e2 40 li:\xe2 J:be2 41 l:bd2 is also good) 38 ... gxf3 39 i.xg3 f2 40 l:lh4 l:lf5 (40 . . . l:lc8+ loses to 41 'it>d3 l:lc l 42 :hi) 41 :l.g4+ �f7 42 i.xf2 :l.xf2 43 �d3.

38 �d3 39 i.eS 40 :l.h1 41 lLlli4

g4 :l.cl l:lxg2

Avoiding Black's last trick in the position: 41 li)d4?? d lli'+! 42 l:lxdl l:ld2+!, which only works because the move ll'ld4 cuts off the bishop's guard of the rook on a I .

After the text-move, I left the stage. The situation is hopeless for Black, so when I carne back Alexei graciously congratulated me on becoming World Champion.

1-0

Winning the world championship was the highlight of my career so far. The knockout format used in the FIDE championship is relatively new, but I think that it is a legitimate format and anyone who wins this event deserves the title of World Champion.

The impact of my victory in the world championship was extraordinary. Chess has become steadily more popular in India over the past few years, and the media coverage of the world championship was excellent even during the New Delhi segment, but it reached a whole new level after I won the final On my return from Teheran to Delhi, I was met at the airport by thousands of people and had a motorcade through the city, with banners hanging everywhere. In Chennai (Madras), my hometown, I was taken in a horse-carriage through the city and also honoured by the State Government. The publicity was not confined to India; there was also plenty of press coverage in other Asian countries such as Malaysia, while on my return to Spain I was, for the f1rst time, met by a media corps. I think the game can only benefit from the developmentof chess in India and other countries which are not currently considered major chess nations. I am sure that chess will continue to make progress in India and we may well see a new generation of players emerging there.

Having won the world title I feel content with what I have achieved on the chessboard and look forward to new challenges in the future.

Combinations

A. Hamed - V. Anand Thessaloniki Olympiad 1984

Although White's king is in an awkward situation, for the moment the bishop on g2 is providing an adequate defence. How did Black press home his attack?

V. Tukmakov - V. Anand Delhi 1986

White's bishop is horribly placed and Black is effectively a piece up on the rest of the board, but does he have anything better than simply exchanging on f2?

COMBINA110NS 317

White chose to meet the attack on his queen by 31 .:td2 (3 1 'lre2 would have been a better chance). How did Black finish the game at a stroke?

J. Levitt - V. Anand Match (3), London 1987

White is threatening to make further progress on the queenside with 29 l:lcc6. How did Black pre-empt this with a kingside strike?

5 B

J. Gdanski - V. Anand World Junior Championship,

Baguio City 1987

Although Black's protected passed pawn on a2 is very dangerous, White has an extra pawn and threatens to start eliminating the queenside pawns by li'ld4+ and lLlxb3. How did Black make use of his advanced pawns? Note that it is not Black's first move but his fourth which is the difficult one!


V. Anand - M. Adams Uoyds Bank, London 1987

Black has just played 14 ... ll:lf6. dS, preparing to answer 15 b5 by 1 5 ... ll:lxc3 followed by 16 ... �xb5. What is the flaw in this plan which allowed White to score a quick win?

7 B

V. Anand - P. Thipsay Coimbatore 1987

Anand had gone into this ending deliberately, even though his king is now far away from the action. Is the position a win?

This was a lucky escape for An· and. White has sacrificed a piece for an enormous attack, and now he could have won by 35 '1'116+ Wg8 36 J:e6, when there is no answer to the threat of 37 'il'g5+ Wf8 38 J:f6. Instead, he rushed in with 35 J:e6, but Black replied ... what?

COMBINATIONS 319

9 B

V. Anand - R. Gerber Bie/ 1988

After 44 . . . �e3, confining White's king, Black has good drawing prospects, for example 45 tDd6 Wc7 46 li)xf7 (46 tDf5 �cl) 46 ... Wxc6 47 h6 .i.d4 48 h7 Wd5 49h8W' �xh8 50 ft:lxh8 Wd4 51 ft:lf7 Wd3 52 tDe5+ "'c2. However, Black was impatient to attack the b2-pawn and played 44 ... �g7. Why was this a mistake?

D. Campora - V. Anand Thessaloniki Olympiad 1988

Black has a formidable line-up along the g-file, but White is threatening to force exchanges by 1i'fl. How did Black make use of the aggressively posted rook on g2?

V. Ivanchuk - V. Anand Reggio Emilia 1988/9

Ivanchuk now played the serious error 17 Whl? (instead 17 M4 would have been very good for White). How did Black then win?


V. Ruban - V. Anand Palma de Mal/orca GMA 1989

Material is equal, but White has several threats. If Black does not have something special, then he will be in big trouble. What did Anand play?

V. Anand - Ye Rongguang Manila 1nterzonal 1990

White would like to advance his pawn, but at the moment this would leave his rook undefended Is 5 1 l:tc2+ followed by 52 b6 the right way to proceed?

V. Anand - A. Dreev Candidmes match (2),

Chennai (Madras) 1991

The position seems to be rather awkward for White as his king is in serious danger. Not only does the possibility of a back-rank mate mean that White's queen has restricted mobility, but Black is also threatening to win the h2-pawn, after which his own h-pawn will become very dangerous. If White defends it by 42 "irgl , then 42 ... :tg2 will drive the queen away. Can you see a clear-cut way for White to save the game?

COMBINATIONS 321

V. Anand -J. Tilnman Linares 1991

White has an extra pawn, but he will have to work hard to exploit it. However, the Dutch grandmaster made the awful blunder 27 ... 'l'd7??. Why was this so bad?

A. Beliavsky - V. Anand Reggio Emilia 199112

Black has an advantage, but it appears that considerable work will be needed to win. Anand made it look easy by winning a vital pawn. How?

L. Ljubojevic - V. Anand Roquebrune Rapid 1992

A truly chaotic position. Black 16 ... l:td8 17 'i'f3 and 16 ... l:te8 17 can reach a favourable ending by

f4 are winning for White, so Sa1ov playing 29 ... 'ifxg5 30 'ifxg5 l:txg5 decided to give up the exchange by 31 �xfl lbd3, but how did Anand 16 .. .d6. What happened next? win far more convincingly?


V. Anand - M. IUescas Linares 1992

In this position one might expect protracted strategic manoeuvring by White, but Anand was able to utilize his queenside pressure to launch an explosive combination. How?

V. Anand - K. Robatsch Manila Olympiad 1992

In return for the sacrificed paw White has a large lead in develo] ment. How did he use it to launch deadly attack? The detailed vari tions are quite complex, but the rna point is to get the basic idea right!

C. Garcia Palenno - V. Anand Manila Olympiad 1992

White's king is very exposed, b he has a knight ready to hop into d How did Black win material?

COMBINATIONS 121

L. van Wely - V. Anand Tilburg 1992

White is a pawn up and he is attacking Black's rook, but even so Anand found a combination leaving him with a clear advantage. What was it?

White has very strong threats on the kingside in return for the piece, but first of all he has to deal with the attack on his rook. Which move is correct: 42 f6 or 42 .i.f6?

V. Anand - A- Khalifman Las Palmas 1993

White's h-file attack is certainly dangerous, but Black's queen has just leapt into White's position and is attacking the b2-rook and c3-pawn. How did White proceed with his attack?


V. Anand - M. Adams V. Anand - B. Gel'and PCA Candidates match (1), Dos Hermanas 1996

Linares 1994 White has a decisive positional

White has a slight advantage as advantage in that he is a pawn up and Black's e-pawn is slightly weak and his knights dominate the centre. Gelhis bishop is more active than the fand therefore decided to stake eveenemy knight. On the other hand, rything on a last gamble by 40 • ...ixh3 Black's doubled rooks on the d-file 41 gxh3 l:txf3. How did Anand recould be useful. Adams now contin- fute this last desperate attack? ued with 31 ... 'iff7. Why was this a mistake?

COMBINATIONS 325

V. Anand - J. Polgar Dos Hemwnas 1997

Material is equal, but White has a dangerous passed pawn . How did Anand utilize this pawn to win game?

V. Anand - V. lvanchuk 6th Amber Rapid, Monaco 1997

White is already a pawn up, but Black will gain a Iitle counterplay if he has time for .. . etJc5, attacking b3 and opening the d-file. How did Anand finish the game vigorously?

Despite material equality, White is winning due to the imprisoned bishop on h2. What is the most efficient method of finishing the game?

In this innocuous-looking position, White found a way to win a pawn. How?

Solutions

1) 42 .. .llld1! Threatening to win the queen with

43 ... lllf2+ and thus forcing White's reply.

43l:a2 J:al! A neat way to eliminate the defen

sive rook and so gain access to f2. 44 J:b2 44 l::l.xa1 lllf2+ 45 'il'xf2 'il'xf2 is

hopeless as White will lose the epawn immediatelY..

44jllxb2 45 Wxb2 'ifb1 46 1IM J:a3 47 .in :a2 o-1

2) 26 ... llld3! This is much stronger than taking

on f2. Black ends up in the same type of position, but with an extra pawn.

27 We2 This is forced because 27 fxe3

Wal+ 28 Wg2 llle1+ and 27 'ilfxd3 'ilfxf2+ 28 Wh1 e2 are dead lost for White.

27 ••• W:xf2+ 28 W:xfl e:xfl+ 29 Wfi wg7

Not only has Black won an important pawn, but his king has a clear run into the heart of White's position.

30 .ibS Wf6 31 cS ll\xcS 32 W:xf2 WeS 0-1

White will soon lose the d5-pawn, when he will be two pawns down f<r nothing.

3) 31...ll\f3.+ 0-1 A typical combination. After 32

gxf3 Wg5+ Black wins material.

4) 28. •• .ixg3 A surprising but strong combina

tion based on the immobility of the e2-bishop.

29 fxg3 'il'h3 30 ..i.d1 White is unable to defend because

his bishop gets in the way wherever it moves. The alternatives are:

1) 30 J:bc6 Wxg3+ 31 Wh1 li:lg8 32 l:[fl g4 33 'il'd2 Wh3+ 34 Wgl g3 and wins.

2) 30 l:[fl Wxg3+ 31 Whl li:lg4 32 J:xf8+ l:[xf8, again winning.

3) 30 ..i.fl Wxg3+ 31 1i'g2 (31 .ig2 and 31 Whl lose to 3 I ...li:lg4) 31 ...Wxe3+ 32 Wh2 Wxb3.

30 ••• llld7 White no longer defends fl and so

Black can win a whole rook. 31 Wg2 Wxgl+ 32 Wxg2 li:lxb6

33 ll\a5 :acs 34 l:[cS :XeS 35 bxcS ll\c4 36 lllxc4 bxc4 37 c6 :CS 311 ..i.a4 c3 0-1

SOLUTIONS 127

5) 48 .... b3! The only possibility, but the game

is not over yet. 49 lDd4+ Or 49 l:th I l:.d6+ 50 ltld4+ l:xd4+

5 1 cxd4 ..ixb2 and the pawns are too dangerous, e.g. 52 �c4 .te l ! 53 J:l.xcl b2 or 52 l:th6+ �f7 53 l:txh7+ ot>g6 54 l:a 7 .i.e I .

49 •.• �eS SOliJxb3..ixb251 .l:l.hl White takes aim at Black's last

pawn 51...J:I.a3! Of course 5 l ...al'ii' 52 lbxal, fol

lowed by .l:txh7, is a draw. 52 lba1 After 52 Wc2ltxb3 53 Wxb3 a I 'I'

54 l:xal ..ixal the h-pawn decides as the white king cannot reach h 1 (55 Wc2 Wxe4 56 Wd2 Wf3 57 wei �g2).

52 ... J:I.xc3+ 53 �dl ..ixa1 The simplest solution. Now there

is no question of a rook's pawn and wrong bishop draw.

54 :Lxa1 J:l.aJ 55 �c2 h5 56 �b2 :SS 57 f4+ �xe4 58 J:l.el + �xf4 59 �a1 h4 60 J:l.e7 J:l.h8 61 �xa2 0-1

6) After 15 b5 lbxcJ White played

16 'ird3!, threatening mate on h7. This interpolation defends the b5-pawn with gain of tempo, and so win a piece. The finish was 16 .. J!,6 17 ..ixc3 'irc7 18 bxa6 liJaS 19 ..ixa5 1-0.

7) Finding the winning line (without

moving the pieces!) is a good test of your powers of visualization. It is long but absolutely forced.

5J ... �c5 After 53 ... �d5 54 lbb6+ �c5 55

lbxa4+ Wc4 56 Wg5 Wb3 57 �f5 Wxa4 58 �e5, White's king is just in time to win the game.

54 lbaJ �d5 The best chance. White can no

longer win the a-pawn, so Black can ultimately play his king round to b3.

55 �g4 �e4 56 c4 �d4 57 �f3 �cS 58 �eJ �b4 59 �dJ!

This knight sacrifice leads to a winning ending of 'I' vs '1'.

59 ... �xa3 60 c5 �bl Or 60 .. .'�b3 (60 ... �b4 61 c6 a3

62 �c2) 61 c6 a3 62 c7 a2 63 c8'ir al'ir64 'irc4+ Wa3 65 'ira6+ Wb2 66 'irb5+ We I , winning as in the game.

61 c6 a3 62c7a2 63 c8'ir a1'1V64 'irc2+ �a3 65 'ireS+ Wal 66 'irc4+ �a3 67 'lfa6+ Wb2 68 'ifb6+ �cl 69 'ireS+ Wb2 70 'ifM+ 1-0

After 70 ... Wa2 7 1 Wc2 Black is mated.

8) After 35 ... 'irxe6 White resigned

immediately, due to 36 fxe6 ..ie4#.

9) Black had overlooked a tactical

point and White won by 45 c7+ Wd7 46 lbd6 Wxc7 47 llJe8+ Wd7 48


!iJxg7 �e7 49 � 'H6 SO b6 �g6 51 tClS 1-0.

10) 33 •. ."it'g3! Sometimes the best move is not a

spectacular combination but a simple liquidation to an ending. At first sight Black has not obtained very much from his position, but a closer inspection shows that the penetration of a rook into the heart of White's position has fatal consequences.

34 .bg3 l:be2 The pawns on a2, c3, d3 and f3

are all vulnerable to attack and in this blocked position White's bishop is unable to develop any activity.

35 .i.b2 l:le3 36 lbg8+ �xg8 37 d4

A desperate attempt to bring the bishop to life.

37 ... £4 Black shuts the bishop out com

pletely. 38 dxeS dxeS 39 l:l.gl+ �b7 40

:n Or 40 l:l.gS (i)f7 41 l:l.xhS+ �g7

42 l:l.fS l:l.xc3 and the pawns will fall one by one.

40 ••• :Xc3 41 :r.r2 (l)r7 42 �g2 bS 43 � b4 0-1

11) 17 ••• (i)f2+! 18 l:l.x£2 (the alterna

tive 18 'l'xf2 .i.xf2 19 gxh3 .i.e3 is hopeless) 18 ... .i.xg2+ 0-1 leads to a forced mate.

12) 48 ••• !iJf3+ 49 � (i)b4+! An unusual case in which a knight

forks another knight SO �xg3 Or SO li'lxh4 .i.xdS+ 51 f3 (51

�xg3 gxh4+ wins a whole rook) SI ...l:l.xc3 S2 li'lg6+ �h7 and Black's material advantage is decisive. Thus White has to surrender a piece.

so_!iJxrS+ s1 �h2 (l)e7 Defending the rook and thus un-

pinning the bishop. 52 e4 .i.e2 The rest is easy. S3 lhc8+ (i)xc854 �g3 .bhS 55

(i)d8 b6 56 f4 bxaS !IT bxaS �58 eS .i.f7 59 fxgS hxgS 60 exd6 (i)xd6 0-1

13) Definitely not! 51 �g2! Avoiding Black's cunning trap,

51 l:l.c2+ �f3 52 b6? 'Wh3+!! forcing a draw by stalemate. White would have to return by 52 l:l.c6 or 52 :Cs, but of course this loses time.

51-.Wrs s2 l:l.c2+ �dl 53 l:l.cl+ Again White must take care. 53

l:l.d2+ �el 54 b6? again runs into a stalemate after S4 . .. Wf3+ SS �gl 'l'g2+.

S3 ••• �e2 54 b6 Now the pawn can safely advance.

Black's one check on f3 is harmless. 54. •• Wits ss l:l.c7 'ires 56 :C2+

�d1 57 l:.cl+ �e2 58 l:l.b1 1-0

so unioNs 329

14) White continued: 42 'ii'g1 l:tgl 43 aS! Not 43 'ii'xg2? hxg2+ 44 �xg2

�5 45 h3 �4 46 a5 .i.b8 47 a6 i.a7 48 �fl �g3 49 �e2 �xh3 50 � .tb8! 5 1 a7 .txa7 52 �4 and one way to win is the attractive manoeuvre 52 ... .te3+ ! 53 �5 .i.g5. Paradoxically, White needs to preserve the black pawn on h3 to set up a stalemate.

43 'l!Fbl + �g7 44 a5 l:txh2+ 45 �gl may also draw, but it is certainly a less foccing line than the text-move and offers more chances to gJ wrong, for example 45 ... l:tg2+

46 �hl l:txg447 a6? (47 'ffb7+ �g6 48 'i'c8 is correct) 47 ... h2 48 a7 l:tg!+ 49 'ifxgl+ hxgl'if+ 50 �xgl i.d4+ and wins.

43,,,l:txg1+ 44 �xg1 Now White only needs to give up

his g-pawn with g5 and his a-pawn with a7 to force a draw, but he has to make sure that Black cannot win by meeting g5 with .. .f 5.

44 ... �g7 After 44 . . . �g5 45 a6 J.d4+ 46

�hi �4 47 g5, to be followed by 48 a7, Black is forced to stalemate White.

45 a6 ..tb8 46 �h1 �g8 47 �g1 47 g5 f5 48 g6 f4 49 g7 f3 50 a7

would also draw. 47...�f8 48 �h1 �e8 49 gS f5 50

R6 f4 51 g7 �7 52 g8'ii'+ �xg8 53 a7 ha7 1/z.lh

15) The game finished 28 ..txf4+! 1-0

since 28 ... exf4 29 'iff6+ forces mate and 28 .. .'�g7 29 l:tdgl leaves Black two pawns down with a bad position.

16) 17 .i.dl A surprise. White ignores the ex

change and, making use of the fact that d6 is now blocked, traps the black queen instead!

17 ••• 'ifc5 18 l:tcl 'ii'xcl 19 ..txcl ..td7 20 ..th6 l:tre8 21 'iff3 1-0

17) 33 ••• 'ifc5! This creates an awkward double

threat: 34 ... �6. epping the queen, and simply 34 ... 'ifxe3+.

34 lbd4 This rescues the queen, but at the

cost of giving up the e5-pawn. 34 ... lbr6 35 'iff3 'if xeS Not only has Black won a pawn,

but he also retains a large positional advantage.

36 l:td1 �g8 37 'iff4 'ifdS 38 l:ta1 e5 39 'ii'rs l:tc4

Further material loss is now inevitable.

40 b3 exd4! 0-1 As 4 1 'ifxd5 l:tc2+ 42 �3 l2Jxd5

wins a piece.

18) 29 ... ..txh3!


A very surprising move. Ignoring the attacked rook, Black takes a pawn which is defended three times!

30 "tfxh3 The other lines lead to crushing

material loss, e.g. 30 gxf5 'ilig4+ 3 1 �f2 'l'g2+ 3 2 �el 'ilfxhl+, 30 l:txh3 'l'xg4+ 3 1 l:tg3 'ilfxdl + or 30 �xh3 'ifxg4+ 3 1 �h2 �f3+.

30 ... 'i'el + 31 oi>gl 'l'e2+ 32 oi>g3 l:tf3+

32 .. . l:txg5 33 'ii"xh7+ �7 would have been even more devastating, but the move played is more than sufficient.

33 �3 'ii"xf3+ 34 1ii>h2 34 oi>h4 g5+ 35 Wh5 'ilff7+ 36

�xg5 '1Wf6+ 37 ltrh5 'l!t'h&t. 34 ••• �xg4+ 35 Wg1 'l'xd1+ 36

Wg2 'ife2+ 37 WgJ 'iffl+ 38 Wxg4 hS+ 39 oi>g5 'ilffti# ( 0-1)

19) 32 �xb6! Hedgehogs are not normally run

over like this! 32 ... �b6 33 lDaS 'l'a7 34 cS

dxc5 35 bxc5 �8 White also wins after 35 ... .i.xc5

36 .i.xc5 �d7 37 l:txd7 l:txd7 38 .i.xb6 or 35 ... �fd7 36 cxb6 fub6 37 'ifg3.

36 c6 l:tb6 36 ... �b6 37 l:tbl is similar. 37 l:tb1 l:txb1 38 l:txb1 1·0

20) 14 l:td8+!

1bis unusual sacrifice traps the black king in the centre long enough for the active white pieces to mount a lethal attack.

14.-.i.xd8 15 .i.xcS .i.d7 Or 15 ... .i.e7 16 'ifd3 f5 (16 ...• f6

1 7 'irb5+ leads to mate) 17 .i.xe7 �xe7 1 8 .i.xf5! 'trb6 (18 ... exf5 19 'trd6+ mates, while 18-.-� 19 .i.xh7 gives White an enormous attack for almost no sacrifice) 19 "llfc3! Wf8 20 .i.e4 e5 (or else 21 �5) 21 fue5 .i.e6 22 'ii"a3+ Wg8 23 "llfe7 l:tf8 24 .i.B (threatening 25 li)d?) with a winning attack.

16 'ifd3 'ii"b5 After 16 ... .i.a4 White has a very

at.-active win by 17 0.114! (17 "l'd6 -'.f6!) 17 ... .i.e? (17 ... l:tc8 18 llX6! mates and 17 .. .f5 also loses to 18 �c6) 18 ll:lc6!! .i.xc6 (18 ... .i.xc5 19 11fd8+ l:txd8 20 l:txd8#) 19 .i.xc6+ Wf8 20 'ifd8+ l:txd8 21 l:txd8#.

If Black tries 16 ... .i.b5, then 17 'li'd6 wins.

17 c4 '1fxc5 17 ... 'ifa4 18 �5 -'.!6 19 �d7

0-0-0 20 -'.xb7+ We? (20 .. .'�xb7 21 l:tbl+ mates) allows a beautiful mating continuation: 21 'l'd6+ Wxb7 22 l:tbl+ WaS 23 l:tb5! l:tc8 24 "llfd3! l:tc6 25 'ii"B 'ifa6 (25 .. .l:thc8 26 'ifxc6+ l:txc6 Tl l:tb8#) 26 l:ta5 l:tb8 27 'ii"xc6+ 'li'xc6 28 l:txa7#.

18 "ifxd7+ Wf8 19 'ifxb7 Not 19 -'.xb7 l:tb8 20 .i.c6 g6 and

Black escapes. 19 ... g6

SOLUTIONS 331

After 19 ... l%c8, White finishes by 20.ic6! 'i'xc6 2 l 'i'xc6 (21 %lxd8+?

%lxd8) 2l ....I:Xc6 22 %lxd8+ We7 23 :Xh8 with a clear extra piece.

20 'i!Fxa8 <Jig7 1-0 Black is too much material down

after 21 l:.xd8 1i'xc4 22 lai2.

21) 29 .•• lt:JxeS! 30 fxeS After 30 ll'ld6 'ife7 3 1 l%xd5 (or

31 fxe5 'i'xh4+ 32 Wgl %lc2! 33 'irxc2 'i'hl + 34 Wf2 'i'g2+ winning the queen) 3l . ..'i'xh4+ 32 Wg2 ll'lg6 33 %ld4 %lc 1 Black has a winning attack.

30 ••• .if3 3l 'i'd3 31 'ii'f2 .ixdl 32 ll'ld6 'i'c7 33

ll'lxc8 'i'xc8 is similar to the game: Black has an extra pawn and a large positional advantage.

31 ... .ixdl 32 ll'ld6 .ic2! The key point. 33 'i!Fe2 'i!Fc7 34 ll'lxc8 'i!Fxc8 White's king is exposed and his

pawns are weak, making Black's task straightforward

35 ll'lb6 'liVeS 36ll'lc4 .ibl 37 'i'b2 .td3 38 ll'ld6 'i'c3 0-1

White is losing more material.

2244 .%lf3!

Threatening both 45 ... lt'Jg4+ and 45. .. h4. Note that the alternatives 44 ... ll'lf3+? 45 gxf3 :Xf3 46 'i'g5 and 44 ... h4 45 .ixfl hxg3+ 46 Wh 1 'i'xa5 47 'i'el are unsound.

45 .ie2 45 gxf3 lbxf3+ costs White his

queen. 4S ... lt'Jg4+ Once again, there are various al

ternatives which fail, for example 45 ... ll'lc4? 46 .ixc4+, 45 ... %lxb3 46 'li'd5+ and 45 .. .lbg3 46 'iii>xg3 ll'lc4+ 47 1i'f4 h4+ 48 Wg4ll'le5+ 49 Wxh4. Black can play 45 ... h4, but it only draws after 46 gxf3 ll'lxf3+ 47 .ixf3 'li'xg3+ 48 Whl 'li'xh3+ 49 Wgl 'li'xf3 50 'i!Fd8+.

46 hxg4 :Xg3 47 Whl? 47 'iii>gl would have been a better

chance, but47 ... '1i'c5+ 48Whl l%xb3 gives Black excellent winning prospects, for example 49 'l'a2 'l'cl + 50 Wh2 'li'c7+ 51 'iii>hl 'li'n.

47 ... hxg4 Thanks to the position of the king

on h 1 , Black has greater attacking possibilities down the h-file. 48 'i!Fxb4 Wh7! 49 'ifc4 'i!Fe7 SO 'ilfdS l%e3 51 .ixg4 'i'h4+ 52 .ih3 'i'el+ 53 �h2 'ifg3+ 0-1

23) 42 .if6! 42 f6 is certainly tempting, but

Black can reply 42 ... ll'lce3 ! 43 fxe3 (43 hxg4 is also met by 43 ... '1i'c3) 43 ... '1i'c3 44 'ifxc3 (44 'li'fl 'li'xe3+) 44 ... dxc3 45 hxg4 c2 (the bishop on g7 may not be well placed if White cannot actually give mate) 46 l%c7 ltal + 47 Wf2 cl'IIV 48 %lxcl l:xcl with a drawn ending.


42 ..• \t>rs 42...0d2 43 l:.g7+ 'it>f8 44 .i.e7+

wins White's queen. 43 hxg4 'ii'd2 44 'ifxd2 Not 44 11Vxc4? J:l.al + 45 'it>h2

'iVh6+ 46 'it>g3 'iVf4+ 47 'it>h3 l:.hl#. 44 �xd2 45 J:.d7 l:laS 46 f3 &4

47 g5 l:.al+ 48 'it>h2 l:.a6 49 J:l.c7 49 g4 would have been simpler as

Black's pawns cannot move at the momen.t . .

49 .•• 't)e) SO .i.g7+ 'it>g8 Sl .i.xeS

d3 52 ..tcJ White must not try to be too

clever: 52 f6 d2 53 g6 hxg6 54 f7+ 'it>f8 55 .i.f4 g5! 56 .i.xe3 J:l.h6+ 57 'it>g3 dl 'iV 58 .i.c5+ l:.d6 would be embarrassing for White.

52 ... d2 53 .i.xd2 li)fi+ 54 'it>h3 li)xd2 55 f6

White wins this ending fairly easily thanks to the poor position of Black's knight.

SS ... l:.al 56 'it>g4 l:.gl 57 'it>g3 l:.hl 58 'it>r4 J:.h4+ 59 'it>CS l:la4 60 l:.g7+ 'it>f8 61 l:.xb7 1-0

24) 23 J:I.xa2! The first sacrifice defuses Black's

counterplay ... 23 ... 'ifxa2 24 li)xh7! ... and the second breaks through

on the kingside. 24 ... li)xb7 25 'ifxg6+ 'it>f8 26

J:.xb7 J:.xb7 27 'ifxh7 The liquidation has left White with

a large advantage. Black's king is

permanently exposed while White has sufficient pawns nearby to be relatively safe. White must try to bring his knight up to support his queen.

27 ••• 'tfg8 After 27 ... 11Vxa4 28 li)g41!t'al + 29

'it>f2 'iVa2+ 30 'it>gl 'iVf7 3 1 11Vh8+ 'it>e7 32 'ifxe5+ 'it>£8 33 11t'xa5 White has enough material to win, while 27 ... 11Val + 28 li)dl 11t'xa4 29 � 'ifal + 30 'it>f2 gives White a decisive attack.

28 'lfh4 lbd7 29 lLlg4 'i!fb3 30 11Vh6+ 'it>e7 31 'ifxc6 'ifbl+ 32 � 'ifb2+

The alternative was 32 ... 'ifb6+ 33

'iVxb6 lLlxb6 34 .i.b5 J:lc8 35 lL!xe5 'it>d6 36 f4 l:.xc3 37 g4, although this ending should be a win, for example 3 7 ... 'it>c5 38 g5 'it>d4 39 g6 'it>xe4 40 g7 J:l.c8 41 .i.d3+! 'it>d4 42 .ib7 li)xa4 43 lLld7, followed by ltlf8.

33 .tel l:.b8 34 ltlxeS! ltlxeS 35 'ifcS+ 'it>e6 36 11Vd5+ 'M6 37 'it'd6+ 'it>r7 38 'if xeS 'ifb6+ 3IJ 'ifd4 'ifxd4+ 40 cxd4 l:.b4 41 .i.bS! llxd4

Or 4I ...l:.xb5 42 axb5 'it>e7 43 'it>e3.

42 'it>e3 l:.b4 43 g4 'it>e6 44 gS l:.b3+ 45 'it>( 4 J:l.b4

45 ... l:.xb5 46 axb5 a4 47 b6 'it>d7 48 g6 a3 49 b7 'it>c7 50 g7 will promote with check.

46 .i.c6 'it>d6 47 .i.dSl:lbl 48 � lt>e7 49 f4 llal SO 'it>g6 llxa4 51 'it>g7 :at sz rs m 53 r6+ 'it>d6 54 g6 1-0

SOLUTIONS 111

25) 32 �dS! Cutting off Black's defence of the

d4-rook. Adams should have reconciled himself to the loss of the exchange and played 32 ... ll4xd5 33

60 • .Jlxd7 61 ltJr6+ Wg6 62liJxd7 ..tr7 63 'iii>d2 'iii>e6 64 liJb6 hS 65 h4 f4 66 hxgS <US 67 liJc4 �gS 68 liJd6 e3+ 69 fxe3 h4 70 lDe4+ 'iii>g4 71 1->e2 �s 72 ltJr2 1-0

cxd5 exd5 34 :te5, although e-M) here White would retain a large ad- 22 liJbS! axbS vantage. However, Black decided to 22 ... e5 (22 ... �xb2 23 ltJd6 costs defend his rook diagonally... Black his queen) 23 'ifh4 liJf6 24

32 .•• 'ii'g7? 33 :e5 liJc3 is rela•vely best, but now the ... only to have the defence cut off win is much simpler for White as he

again! There was an equally good a!- can just drop his knight in on d5. temative solution by 33 .be6+ lii>h8 23 �xg7 Wxg7 24 'iWh6+ 34 .ixf5. Oeaning up the kingside before

33 ••• :t4xdS 34 cxdS cxdS 35 g4! trapping the bishop by taking back Opening up more lines for the on b5.

rooks makes the win easy. 24,,,\ii>g8 2S'if:xh7+ � 26'ifh8+ 3S ••. lDe7 36 %:txe6 fxg4 37 'iWh4 lii>e7 27 'ifh4+ gS

J:d7 38 :tbe1 1ii>rs 39 fS 1·0 27 ... lii>fS 28 axb5 leaves White After 39 . . . liJxf5 40 :tf6+ %:tf7 41 well ahead on material, position and

lhf5 %:txf5 42 'ifd8+ lii>f7 43 lle7+ attack! Black loses his queen. 28 'ifxgS+ f6

26) By far the simplest solution is 42

.!Llef6+! gxf6 43 :g4+ 1-0, as White will be a clear rook up.

27) 56 l:td1 liJcS 57 d7 Winning a piece, after which the

remaining task is purely technical. 57 ... liJe6 58 d8'6' liJxd8 59 :Xd8

gS 60 l%d7 A neat echo of the earlier combi

nation. The exchange of rooks makes the task simpler.

28 ... liJf6 29 :tf3 is deadly. 29 llh7+ 'iii>d6 30 'ifd2+ lii>c7 31

cxbS 1-0 3l...�xe4 32 be4 1i'xe4 33 l:tcl+

lii>b8 34 l:txd7 is devastating.

29) 45 1i'xe6 1-0, since 45 ... l:txe6 al

lows mate, while 45 .. .fxe6 46 l:txh7 lii>xh7 47 l:thl wins the bishop in perfect safety.

30) 14 liJxd4! �xg2 15 liJfS 'ife6 16

1VgS


Defending the knight on f5 with gain of tempo.

16 ••• ltJe8 17 'ittxg2 h6 18 'ii'g4 �f6 19 'ii'f3

The extra pawn has been consolidated and, as a bonus, the bishop on

b2 has been activated. The rest is straightforward.

19 •• .e4 20 dxe4 �e4 21 :Cell 'itth7 22 bxcS bxcS 23 l:tdS �f6 24 .hf6 �6 25 :XeS lhb8 26 l:td1 l:tb2 27 ll:ld4 fie7 28 l:tbS 1-0

Index of Opponents

Numbers refer to pages. A bold number indicates that Anand was White.

Adams 1 30, 318, 304, 324 Agdestein 16 Andersson 325 Bareev 81, 110 Beliavsky 54, 136, 321 Benjamin 23, 142 Campora 319 Dreev 320 Fta�nik 119 Garcia Palermo 322 Gdanski 3 1 7 Gelfand 106, 189, 324 Gerber 319 Gil 3 1 7 Hamed 316 Hiibner 87 lllescas 322 lnkiov 7 lvanchuk 92, 102, 207, 253, 319, 325 Izeta 115 Kamsky 98, 148, 157, 164, 171 Karpov 59, 211, 244 Kasparov 67, 74, 184 Khalifman 297, 323 Kramnik 225, 257, 325 Kuijf, M. 40

Lautier 216, 220 Levitt 3 1 7 Ljubojevic 321 Morovic Fernandez 49 Nikolic 234, 291 Ninov 12 Oll l26, 262 Petursson 43 Piket 273, 323 Polgar, I. 153, 198, 325 Reinderrnann 269 Robatsch 322 Ruban 320 Salov 321 Shirov 238, 311 Sokolov, I. 90 Spassky 34 Svid1er 277 Tal 29 Thipsay 318 Timrnan 176, 321 Tomczak 318 Topa1ov 202, 249, 283 Tukmakov 316 Van Wely 323 Ye Rongguang 320

Index of Openings Numbers refer to pages. A bold number indicates that Anand was White.

Caro-Kann Defence 283 Dutch Defence 291 English 29 French Defence 74, 81, 110, 311 Griinfeld Defence 277 Petroff Defence 87, 102, 257 Pirc Defence 54, 115, 198 Ponziani Opening 40 Queen's Gambit

Queen's Gambit Accepted 106, 211

Semi-Slav 59, 225, 234 Slav Defence 126

Queen's Pawn 43 Ruy Lopez 16, 207

Arkhangelsk 157

Symbols + Check ++ Double check # Mate

Good move ! ! Excellent move ? Bad move ?? Blunder

Breyer 34 Chigorin 273 Flohr-Zaitsev 164 MBIIer 238, 249, 262, 314 Worrall Attack 98, 176

Scandinavian Defence 216 Sicilian

Grand Prix Attack 189 Kan 12, 130 Maroczy Bind 49 Najdorf 119, 136, 148, 153,

171, 220, 269, 297 Richter-Rauzer 7, 23, 92, 142,

253 Scheveningen 67, 90, 184, 202

Trompowsky 244

! ? Interesting move ?! Dubious move 1 -0 White wins 0-1 Black wins 1h-1h Draw (n) nth match game (D) Diagram follows

Anand my best games of chess (expanded edition)

Education

Transcript of Anand my best games of chess (expanded edition)