Autonomous Learning Investment Strategies...

For investment professional use only. Not to be shown or distributed to the general public.

Autonomous Learning Investment Strategies (ALIS)

Adil Abdulali

The Next Investment Process Paradigm: The Third Wave


THE CONFLUENCE OF FIVE UNPRECEDENTED DEVELOPMENTS

HAS TRANSFORMED INTO THE NEW ALIS PARADIGM:

Enormous growth and

multiple structures of

data

New data science

and structuring

platforms to parse

and classify data

Machine Learning –

AI is finally working

Low cost on-demand and

scalable computing leads

to favorable investor fee

dynamics

Playing too close

to the information “Edge” is

a risk

Autonomous Learning Investment Strategies (ALIS): Why now?

1


Structured Unstructured

Financial

Non-Financial

Source: IDC Digital Universe Study, 2010

Four quadrants—the multiple structures of data

News

SEC Filings

Credit Card Data

Purchase Scan Data

Search Keywords

Advertising Data

Municipal Traffic Data

Shipping Manifest

Twitter

GPS Cell Phone Data

Blogs/Websites

Satellite Imaging

A L I S+C O M P U T A T I O N A L

F I N A N C E

2


c

Exponential data growth—doubling at the rate of Moore’s Law1

0

5

10

15

20

25

30

35

40

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

Structured Data Unstructured Data

Ze

tta

byte

s

1 Kelly, Kevin. The Inevitable: Understanding the 12 Technological Forces that will Shape Our Future (p.257). Penguin Publishing Group. Kindle Edition2 McKenna, Brian “What does a petabyte look like?” ComputerWeekly.com. March 2013. According to Michael Chui, principal at McKinsey, the US Library of Congress “had collected 235 terabytes of data by April 2011 and a petabyte is more than four times that.”3 One zettabyte = one thousand exabytes = one million petabytes = one billion terabytes = one trillion gigabytes (1,000,000,000,000 gigabytes)

Source: IDC Digital Universe Study, 2010

1 ZETTABYTE:more than four million times the

size of the entire US Library of

Congress2,3

3


Structured Unstructured

Financial

Non-Financial

DATA

Clean

How to interpret that data—new platforms

PLATFORM ORGANIZES DATA FOR

ANALYSIS

Data Science Platform

Parse

Classify Normalize

Categorize

Transform

4


Traditional Programming

Hypothesis driven: Programmers develop

models, based on which they code rule-

based algorithms. To these they feed data

to produce the desired outcome

What is Machine Learning?

1Phil Simon (March 18, 2013). Too Big to Ignore: The Business Case for Big

Data. Wiley. p. 89. ISBN 978-1-118-63817-0

Source: Sebastian Raschka, 2016

Machine Learning

Data driven: Data is provided to a learning

algorithm from which models are

developed. These can be used to solve the

problem task.

Programmer

Inputs (observations)

Model Computer Outputs

Inputs

Outputs

ModelComputer

“Machine Learning is the field of study that gives computers the ability to

learn without being explicitly programmed .”

ARTHUR SAMUEL, 19591

5


Human

Performance

Computer

Performance

Time

Pe

rfo

rma

nce

We are here.

Jeremy Howard TED talk

Now machines learn: AI is real, not artificial

6

http://www.ted.com/talks/jeremy_howard_the_wonderful_and_terrifying_implications_of_computers_that_can_learn


HOW MANY YEARS FOR A COMPUTER TO BECOME W ORLD MASTER?

Is the investment game too complicated? It’s been said before…

CHECKERS

“It’s (the computer’s) guidelines

for judgement are not nearly as

good as a human’s”

MARION TINSLEY, 1994 1

1994

Computer becomes Checkers World Champion

10^20 Permutations

”It may be 100 years before a computer

beats humans at Go – maybe even

longer”

DR PIET HUT FOR PRINCETON'S INSTITUTE

FOR ADVANCED STUDIES, 19973

2016

AlphaGo Computer beats human

World Champion (Lee Sedol)

GO

10^700 Permutations

”No computer will ever beat me”

GARRY KASPAROV, 1987 2

1997

Computer wins a series against a reigning World Champion

(Kasparov)

10^120 Permutations

CHESS

1Corcoran, E. “Squaring off in a game of Checkers”. The Washington Post, August 15,

1994 / 2Hsu, F.H. Behind Deep Blue: Building the Computer that Defeated the World

Chess Champion Princeton University Press (2002) / 3Johnson, G. “To Test a Powerful

Computer, Play an Ancient Game. The New York Times. July 29 1997

0 5 10 15 20 25 30 35 40

Go

Chess

Checkers 38 years

12 years

6 years

1950 2016201020001990198019701960

7


Making decisions under conditions of uncertainty with incomplete information

2011

IBM’s Watson machine defeats the highest ranking

Jeopardy champions in history

JEOPARDY

Uncertain information: the question is

unknown, and must be solved with a conviction

level bet faster than the opponent

“Defeating a chess champion is a

piece of cake compared to parsing

puns and analyzing language”

WSJ, 20111

1990 201720152010200520001995

Checkers Chess Jeopardy Go Poker

1Baker, S. “Can a Computer Win on ‘Jeopardy’?”. The Wall Street Journal, February, 5, 20112Wilson, C. “Jeopardy, Scheopardy”, Slate, February 11, 2011

POKER

Incomplete information: the state of the game

is unknown (cards are hidden), and the

opponent’s strategy is unknown (they bluff)

2017

Libratus computer beats four professionals at no-

limit Texas Hold ‘em poker by $1.7million

“Watson can win at Jeopardy, but how

would it do at poker? …This is outside

the realm of traditional game theory”

Slate, 20112

8


This isn’t a false dawn

0%

5%

10%

15%

20%

25%

2011 2012 2013 2014 2015

Computer error rates on ImageNet Visual Recognition Challenge

Sources: The Economist, Google, ImageNet, Stanford Vision Lab

RAPID MACHINE-LEARNING IMPROVEMENTS MEAN COMPUTERS HAVE SURPASSED

HUMANS AT PREVIOUSLY UNTHINKABLE TASKS

HUMAN LEVEL

9


GARY KASPAROV1

“Weak human + machine + better process was superior to a strong computer

alone and, more remarkably, superior to a strong human + machine + inferior

process.”

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Man and MachineMachine onlyMan only

Man and Machine dominates Machine–only and Man–only

Data from Kenneth W. Regan; Ranking by raw error as of 2010 using chess program Rybka 3;

http://www.cse.buffalo.edu/~regan/chess/fidelity/FreestyleStudy.html1Garry Kasparov “The Chess Master and the Computer”. The New York Review of Books, February

11 2010; http://www.nybooks.com/articles/2010/02/11/the-chess-master-and-the-computer/

% OF TOP 206 ALL-TIME HIGHEST RATED CHESS PERFORMANCES BY ‘CYBORGS’

(COMBINATIONS OF MAN AND MACHINE), MACHINES ONLY AND MAN ONLY

10

http://www.nybooks.com/articles/2010/02/11/the-chess-master-and-the-computer/


Machine learning (ML) frameworks and methodologies

SUPERVISED LEARNING

• Predicting Value

• Decision Tree Regression

• Nearest Neighbors

• Predicting Classification

• Neural Networks/Deep Learning

• Support Vector Machines

• Baysian Classifiers

• Genetic Algorithms

• Markovian Decision Processes • Gaming AI

REINFORCEMENT LEARNING OTHER

UNSUPERVISED LEARNING

• Clustering

• K-means clustering

• Association Rule Learning

• Dimensionality Reduction

• Principal Components

Analysis

Note: Many of these techniques can be used across multiple frameworks – The

above is just illustrative of potential types of ML. 11


c

Plummeting processing costs make ML methodologies achievable

0.000001

0.00001

0.0001

0.001

0.01

0.1

1

10

100

1000

1980 1985 1990 1995 2000 2005 2010

$ c

ost p

er

1b

n c

alc

ula

tio

ns (

log

sca

le)

MACHINES DON’T OWN HOUSES IN THE HAMPTONS!

Source: Nordhaus (2007); updated data through 2010 from Nordhaus, personal website,

http://www.econ.yale.edu/Nordhaus/homepage/”Two Centuries of Productivity Growth in

Computing”, authors calculations. Adjusted for year 2006 purchasing power

EXPONENTIAL

COST REDUCTION:

A million dollars of 1980

computing power costs

less than 4 cents today

12


The Cloud: on-demand and scalable computing costs plummet

0

0.05

0.1

0.15

0.2

0.25

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

$ p

er

GB

pe

r M

on

th

Amazon Microsoft Google IBM

The lowest price has

fallen 84% since 2010

March 2006: The first ‘big player’

to enter cloud computing

November 2010:

Enters the market with

Microsoft Azure

May 2011: Launches Google Cloud May 2013: Acquires Softlayer

Early entrant

August 2008Early Entrant October 2015:

Dell to acquire

early entrant EMC

June 2009 January 2011 June 2012November 2011

Source: Alex Teu, “Cloud Storage is Esting the World Alive” TechCrunch. 20 August 2014;

Trefis Team. “IBM Cloud Services Part-II” Forbes. 9 March 2015; Company Websites.

13


c

When costs plummet, investors gain

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

0.00% 5.00% 10.00% 15.00% 20.00% 25.00%

Old 2&20 New 1/10/20

OPPORTUNITIES FOR FAVORABLE FEE RESTRUCTURING

Gross Returns of Underlying Portfolio

% o

f G

ross R

etu

rn R

eta

ine

d B

y I

nve

sto

rs

% Change

125%

50.0%

26.6%

7.7%

Gross Return

2.5%

5.0%

10.0%

15.0%

14

Source: Hedge Fund Fees – A Perfect Solution by Jeffrey Tarrant, Adil Abdulali and Michael

Weinberg http://www.pionline.com/article/20170306/ONLINE/170309918/hedge-fund-fees-

8211-a-perfect-solution


Man + Machine + Data Science + Four Quadrants + Low Costs = ALIS

Model

Inputs Outputs

15


c MOV37 ALIS

Sweet Spot

Portfolio Holding Period

Sca

lab

ility

of S

tra

teg

y

HIGH

LOW

SHORT-TERM MEDIUM-TERM LONG-TERM

a

No High

Frequency

Trading

This is not high frequency trading

16


“Repeatable process”

Information source

Discretionary

Unique analytical processGood Judgement of

Portfolio Manager

Structured financial data

and “edge” through

relationships within

sectors, regions, and other

informational sources,

expert networks etc.

Top Tier MBAs &

Finance Professionals

Human intellectual

and intuitive pattern

recognition: 10,000

hours1

Massive, finite

quantities of expensive

structured financial

data

PhDs, algorithms and

expensive processing

power

Many PhD’s

assisted by machineQuant

ALIS

The Third Wave: The Darwinian Evolution of the “Repeatable Process”

Source: Malcolm Gladwell. “Outliers” 17


Pay a Fine

Meanwhile, the information “edge” game has changed

Back off the “Edge” Go to Jail

Regulation FD (Fair Disclosure)

–Company information to

everyone at the same time

2000 2013

“SAC Capital to pay

$1.8 billion, the largest insider

trading fine ever”1

2014

“SAC’s Martoma gets

nine years prison for insider

trading”2

1 Sheelah Kolhatkar, “SAC Capital to pay $1.8 billion, the largest insider trading fin ever”

Bloomberg News, November 4, 2013 / 2 Nate Raymond “SAC’s Martoma gets nine years

prison for insider trading” Reuters, September 8, 2014 / Source: Vintage Monopoy Mr

Pennybags Chance Card Vectors by robdevenney on DeviantArt.com

Investment players have three choices

18


The genealogy of existing quant managers

1982 RenTech

DE Shaw1988

2 Sigma

The Voleon

Group

Winton

AHL

(now Man AHL)

2001

1987

1997

Today

2007

Aspect

Capital

$36bn Closed $6.6bn $19.2bn $39bn $35bn $34bn $1.5bn

Medallion

Fund

$6.7bn

1993 PDT

1986APT (Morgan

Stanley)

2011PDT

Partners

Jim Simons establishes

Renaissance Technologies

David Shaw joins APT group

at Morgan Stanley

Mike Adam, Martin Lueck,

and David Harding form AHL

(now Man AHL)

David Shaw leaves APT

to form own fund

Pete Muller’s PDT Group spins

Out of Morgan Stanley

Pete Muller works for and is offered

a job at RenTech, but instead forms

PDT out of APT at Morgan Stanley

Lueck & Harding spin off to form

own companies

Michael Kharitonov & Jon

McAuliffe, spin off DE Shaw to form

Voleon

David Siegel & John Overdeck

spin off DE Shaw to form 2 Sigma

19

Flagships Close As Assets Grow


“I have one guy who has a PhD. in finance. We don’t hire people from

business schools. We don’t hire people from Wall Street. We hire people

who have done good science”

Existing Quant Managers Industry Connections Universities

Where’s the talent?

1Lux, Hal ‘The Secret World of Jim Simons’ Institutional Investor, 11/1/2000

JIM SIMONS1

INVESTORS

20


WE ARE

Pivotal Moves of Man and Machine

“Kasparov had concluded that the counterintuitive

play must be a sign of superior intelligence”1

“AlphaGo’s move was simply beyond the current

human understanding of the game...”2

1Murray Campbell, one of the IBM scientists who designed Deep Blue, speaking to Nate Silver for his

book The Signal and the Noise. Finley, Klint. ‘Did a computer bug help deep blue beat Kasparov?’

Wired.com, September 2012 / 2Esteban, Cristobal, “Move 37”, Medium. March 2016,

https://medium.com/@cristobal_esteban/move-37-a3b500aa75c2#.1d9d7vwcm 3MOV in assembly language means to replace the old with the new.

…YOUR MOVE3

KASPAROV RESIGNED ON HIS 37TH MOVE AGAINST DEEP BLUE

ALPHAGO’S 37TH MOVE FLUMMOXED GO MASTER LEE SEDOL

21


Autonomous Technical2,3,4,5

Growth of Unit Investment Performance in MSCI World Down Months

ANNUALIZED RETURN

BETA

MSCI WORLD

R

MSCI WORLD

R-SQ

MSCI WORLD

31.6% 0.24 0.29 0.09

1.0%

-2.4%

-8%

-6%

-4%

-2%

0%

2%

4%

Jan-13 Jan-14 Jan-15 Jan-16 Jan-17

Autonomous Technical MSCI World Index

Autonomous Technical Average MSCI World Index Average

$289

$145

$90

$110

$130

$150

$170

$190

$210

$230

$250

$270

$290

$310

Jan-13 Jan-14 Jan-15 Jan-16 Jan-17

Autonomous Technical MSCI World Index

Data estimated of February 28, 2017

Please refer to Disclaimer and Notes beginning on Page 44.

Underlying Manager Profile

STRATEGY OVERVIEW CLASSIFICATION REGION

FUND

LAUNCH

DATE

Autonomous technical

Inspired by video gaming AI, this strategy utilizes 1000’s of independent Jesse Livermore like autonomous computer

agents trading, adapting and learning from each trade in a simple portfolio with clear risk management limits and

protocols but implemented with no human intervention at all. The non-traditional background of the founder as well as

the unusual structure of the organization makes this the ultimate financial market hack.

Autonomous

TechnicalUS 2013

22


Flow Detection US Equities2,3,4,5


FUND

LAUNCH

DATE

Capital Flow detection by ML

algorithms

A two state statistical arbitrage strategy driven by prevailing liquidity conditions in the market based on measuring

investment flows in the equity markets using sophisticated mathematical techniques. These techniques allow the

machine to process large amounts of data across a large number of securities simultaneously to uncover patterns

that would not be possible for a human. One of the states is active in stress and driven by liquidation flows and the

other state is active in calm markets and is driven by sentiment indicators processed in a unique way. ML is used in

the construction of each strategy as well as the switching algorithm between the two

Flow Detection

US EquitiesUS 2013

ANNUALIZED RETURN

BETA

MSCI WORLD

R

MSCI WORLD

R-SQ

MSCI WORLD

4.8% 0.20 0.17 0.03

$123

$137

$80

$90

$100

$110

$120

$130

$140

$150

Jun-13 Jun-14 Jun-15 Jun-16

Flow Detection US Equities MSCI World

0.9%

-2.4%

-8%

-4%

0%

4%

8%

Jun-13 Jun-14 Jun-15 Jun-16

Flow Detection MSCI World Index

Flow Detection Average MSCI World Index Average



23Data estimated of February 28, 2017



Deep Learning Market Neutral2,3,4,5


FUND

LAUNCH

DATE

Deep Learning Market Neutral

Uses Deep Learning algorithms, Natural Language Understanding and large scale parallel computing on historical

data to predict prices. The founder is a Protégé of one of the more successful global stat arb funds and has used

machine learning to take it to the next level. R&D makes up a significant part of the investment process.

Deep Learning US 2016

ANNUALIZED RETURN

BETA

MSCI WORLD

R

MSCI WORLD

R-SQ

MSCI WORLD

5.4% -0.33 -0.46 0.22

$104

$111

98

100

102

104

106

108

110

112

Jun-16 Aug-16 Oct-16 Dec-16 Feb-17

Deep Learning Market Neutral MSCI World Index

2.1%

-1.5%

-3%

-2%

-2%

-1%

-1%

0%

1%

1%

2%

2%

3%

3%

Jun-16 Aug-16 Oct-16 Dec-16 Feb-17

Deep Learning Market Neutral MSCI World Index

Deep Learning Market Neutral Average MSCI World Index average






Genetic Algorithmic2,3,4,5

$155

$178

$80

$100

$120

$140

$160

$180

$200

Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan-17

Genetic Algorithm MSCI World Index

2.1%

-2.6%

-10%

-8%

-6%

-4%

-2%

0%

2%

4%

Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan-17

Genetic Algorithm MSCI World Index

Genetic Algorithm Average MSCI World Index Average


FUND

LAUNCH

DATE

Genetic Algorithmic

Seeks to identify an optimal, diverse collection of trading models to achieve investment targets using genetic

algorithms. These algorithms provide a framework to identify best solutions from a massive pool of potential solutions

and to evolve them over time. The managers background is from the data science technology industry

Genetic

AlgorithmicUS 2012

ANNUALIZED RETURN

BETA

MSCI WORLD

R

MSCI WORLD

R-SQ

MSCI WORLD

8.8% 0.10 0.16 0.02






ML Fundamental2,3,4,5


FUND

LAUNCH

DATE

ML Fundamental

This strategy uses the latest deep learning techniques from the world of artificial intelligence and applies it to stock

picking. It is a completely systematic strategy anchored in fundamental analysis based on processing a much larger

volume and breadth of information than a human ever could. The systematic pattern matching algorithm uses

fundamentals, technicals, news, governance and 3rd party research estimates simultaneously to make its stock picks.

The portfolio is liquid, has low turnover, does not use balance sheet leverage and targets 100 to 240 stocks in total.

ML Fundamental Global 2015

ANNUALIZED RETURN

BETA

MSCI WORLD

R

MSCI WORLD

R-SQ

MSCI WORLD

7.6% -0.14 -0.28 0.08

$113

$108

80

85

90

95

100

105

110

115

120

Aug-15 Nov-15 Feb-16 May-16 Aug-16 Nov-16 Feb-17

ML Fundamental MSCI World Index

1.0%

-2.7%

-9%

-7%

-5%

-3%

-1%

1%

3%

5%

Aug-15 Nov-15 Feb-16 May-16 Aug-16 Nov-16 Feb-17

ML Fundamental MSCI World Index

ML Fundamental Average MSCI World Index Average






Autonomous Learning Investment Strategies (ALIS)

Adil Abdulali

The Next Investment Process Paradigm: The Third Wave


Disclaimer & Notes

D I S C L A I M E R

The information contained in this document is confidential and is intended only for

the use of the person to whom it is given and is not to be reproduced or

redistributed. It is not a solicitation to invest in any investment product, nor is it

intended to provide investment advice. It is intended for information purposes only

and should be used only by sophisticated investors who are knowledgeable of the

risks involved. This document does not constitute an offer to sell or the solicitation

of an offer to buy any securities, nor will any sale of a security occur in any

jurisdiction where such an offer, solicitation or sale would be unlawful.

This document may only be distributed to persons who are “accredited investors”

and “qualified purchasers” within the meaning of U.S. Securities laws and to non-

U.S. persons. It is intended solely for the use of the person to whom it is given and

may not be reproduced or distributed to any other person. The MOV37™ Funds of

Funds have not been, and will not be, registered under the Investment Company

Act of 1940, as amended, and interests of the MOV37 Funds of

Funds have not been, and will not be, registered under the Securities Act of 1933,

as amended, and may only be offered in private placement transactions.

An investment in a MOV37™ Funds of Funds may be made only pursuant to the

applicable offering documents. Past performance is not a guarantee of future

results. Inherent in any investment is the potential for loss. This summary is for

discussion purposes only. It is not intended to supplement or replace a

confidential offering memorandum or related offering materials, which should be

the sole basis for making an investment decision.

Important information: Hedge funds are speculative investments and are not

suitable for all investors, nor do they represent a complete investment program.

Hedge funds are not subject to the same regulatory requirements as mutual funds.

MOV37™ Fund of funds are only opened to qualified investors who are

comfortable with the substantial risks associated with investing in hedge funds.

MOV37™ Fund of Funds’ investment programs are speculative and entail

substantial risks. An investment in the MOV37™ Fund of Funds includes the risks

inherent in an investment in securities, as well as specific risks associated with

limited liquidity, the use of leverage, arbitrage, short sales, options, futures,

derivative instruments, investment in non-US securities, “junk” bonds and illiquid

investments. In particular, the MOV37™ Fund of Funds may use leverage in

making investments in underlying funds, and underlying fund managers also may

employ leverage through a number of measures, either of which could increase

any loss incurred. The more leverage employed, the more likely a substantial

change will occur, either up or down, in the value of the investment. There can be

no assurances that the strategy pursued by MOV37™ (hedging or otherwise), or

the strategy of any underlying fund manager with whom the MOV37™ Fund of

Funds invest, will be successful or that MOV37™ (or any underlying fund

manager) will employ such strategies with respect to all or any portion of a

portfolio. Past performance is not indicative of future results.

This message contains information that is confidential and privileged. Unless you

receive prior authorization from MOV37™, you may not use, copy, print, forward or

disclose to anyone the message or any information contained in the

message. Thank you for your cooperation.

Although this summary has been prepared using sources, models and data that

MOV37™ believes to be reasonably reliable, its accuracy, completeness or

suitability cannot be guaranteed. Therefore, the information is supplied on an “AS

IS” basis, and no warranty is made as to its accuracy, completeness, non-

infringement of

third-party rights, merchantability or fitness for a particular purpose. Furthermore,

the information used to generate this summary includes third party unverified data

as well as MOV37™’s assumptions, made on a best efforts basis, with regard to

incomplete information, and independent of market conditions. The recipients of

this report assume all risks in relying on the information set forth herein.

NOTES:

[1] The annualized return since inception for each fund and strategy is reported net

of all underlying manager fees and gross of MOV37™ expenses, management

fees and incentive fees. Returns will vary based on the timing of investment.

Returns, allocations and other data are estimated through February 28, 2017.

[2] The data provided herein has been calculated based upon the most recent data

that MOV37™ has received from each hedge fund manager. MOV37™ cannot

confirm or guarantee the accuracy of the data provided.

[3] The index information is included for discussion purposes only to show the

general trends in certain markets in the periods indicated and is not intended to

imply that the MOV37™ funds will be similar to these indices in performance,

composition, or element of risk. While MOV37™ does not use benchmarks for its

funds, it tracks various indices for comparative purposes only. Index comparisons

for the underlying funds were chosen for comparative purposes only in MOV37™’s

good faith judgment and may not reflect actual fund benchmarks, if any.

The S&P 500 Index is compiled by Standard & Poor’s and includes 500 leading

companies in leading industries of the U.S. economy. The S&P 500 DRI stands for

dividend reinvestment index. This description has been obtained from

www.standardandpoors.com. The S&P 500 Index is included for informational

purposes only and is not representative of the type of securities invested in by the

MOV37™ funds.

The MSCI World Local Index is a free float-adjusted market capitalization weighted

index that is designed to measure the equity market performance of developed

markets. The MSCI World Index consists of the following 23 developed market

country indexes: Australia, Austria, Belgium, Canada, Denmark, Finland, France,

Germany, Hong Kong, Ireland, Israel, Italy, Japan, Netherlands, New Zealand,

Norway, Portugal, Singapore, Spain, Sweden, Switzerland, the United Kingdom,

and the United States. This description has been obtained from www.msci.com.

The MSCI World Index is included for informational purposes only and is not

representative of the type of securities invested in by the MOV37™ Funds or any

underlying hedge fund.

This description has been obtained from www.msci.com. The MSCI World Index is

included for informational purposes only and is not representative of the type of

securities invested in by the MOV37™ funds.

The HFOF Composite return is the average of the following five fund of funds

indices: EDHC, HFRI, INVESTHEDGE, Eurekahedge and Barclay FOF Indices.

[4] Growth of Unit Investment represents the growth of $100 had it been invested

in the fund or hypothetical portfolio for the time period indicated and has been

calculated net of the underlying managers’ fees and expenses and gross of

MOV37™’s expenses, management fee and incentive fee. Growth of $100 had it

achieved the return of the indicated index is shown for comparative purposes only.

[5] Annualized return has been reported by the manager and represents the

annualized return since the fund’s inception estimated through February 28, 2017.

[6] The proposed portfolio is for illustrative purposes only and is intended to show

how the hypothetical portfolio would have performed if it had been invested in

particular portfolio funds during the period commencing January 1, 2014. The

returns are reported net of all underlying manager fees and gross of MOV37™’s

expenses, management fees and incentive fees. Returns, allocations and other

data are estimated through February 28, 2017

The data is based on a hypothetical portfolio that has not be created and includes

7 underlying hedge funds, which are comprised of underlying hedge funds that

MOV37™ funds are not currently invested in. The data has been estimated by

MOV37™ for a hypothetical portfolio with $200 million in assets under

management based on information provided by the underlying hedge fund

managers, conversations with the underlying hedge fund managers and SEC

filings, among other sources of information, and MOV37™’s good faith

assumptions with respect to incomplete information. MOV37™ cannot confirm or

guarantee the accuracy of the information provided that was used to calculate the

portfolio breakdown and exposures. The underlying hedge funds and allocations

thereto for the proposed portfolio are hypothetical and may not be available at the

time of investment due to capacity constraints. Furthermore, any hedge funds that

are not currently in MOV37™’s portfolio will be subject to MOV37™’s operational

due diligence process. More parameters, other than investment strategy, are

considered when constructing a portfolio, such as risk levels, desired

transparency, liquidity requirements and asset classes, for example. Any future

portfolio created for the investor may not have the characteristics outlined herein.

THE PROPOSED MOV37™ PORTFOLIO IS HYPOTHETICAL AND DOES NOT

REPRESENT THE ACTUAL CHARACTERISTICS EXPERIENCED BY MOV37™.

Hypothetical performance is inherently limited and does not reflect actual trading

or decisions that may be made in response to market or other conditions. Even

though the hypothetical information does not represent past performance, any past

performance is not a guide to future results.

The proposed portfolio is equally weighted across all funds in the portfolio, and

rebalanced quarterly. For periods when the returns of funds are not available, the

allocation that would have been applied to that fund is distributed equally across

the remaining funds.

[7] MOV37™ has characterized the techniques employed by the ALIS managers,

as reflected in this slide, based upon information which was provided to it or made

available by the ALIS manager through extensive process-focused meetings and

due diligence. MOV37™ has not conducted an examination of the ALIS

managers’ codes. There can be no assurance that the characterizations are

correct or will accurately reflect the techniques in the future.

28

Autonomous Learning Investment Strategies...

Documents

Transcript of Autonomous Learning Investment Strategies...