See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/328128353

Pacing profile in the main international open-water swimming competitions

Article  in  European Journal of Sport Science · October 2018

DOI: 10.1080/17461391.2018.1527946

CITATIONS

7READS

255

3 authors:

Some of the authors of this publication are also working on these related projects:

Monitoring Training in elite athletes View project

nfluence of training on the response to exercise of hormoneς plasma View project

Roberto Baldassarre

Italian University of Sport and Movement "Foro Italico"

18 PUBLICATIONS   53 CITATIONS   

SEE PROFILE

Marco Bonifazi

Università degli Studi di Siena

113 PUBLICATIONS   2,409 CITATIONS   

SEE PROFILE

Maria Francesca Piacentini

University of Rome Foro Italico

131 PUBLICATIONS   3,087 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Roberto Baldassarre on 29 January 2019.

The user has requested enhancement of the downloaded file.

Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=tejs20

European Journal of Sport Science

ISSN: 1746-1391 (Print) 1536-7290 (Online) Journal homepage: http://www.tandfonline.com/loi/tejs20

Pacing profile in the main international open-water swimming competitions

Roberto Baldassarre, Marco Bonifazi & Maria Francesca Piacentini

To cite this article: Roberto Baldassarre, Marco Bonifazi & Maria Francesca Piacentini (2018):Pacing profile in the main international open-water swimming competitions, European Journal ofSport Science, DOI: 10.1080/17461391.2018.1527946

To link to this article: https://doi.org/10.1080/17461391.2018.1527946

Published online: 06 Oct 2018.

Submit your article to this journal

View Crossmark data

ORIGINAL ARTICLE

Pacing profile in the main international open-water swimmingcompetitions

ROBERTO BALDASSARRE1, MARCO BONIFAZI2, & MARIA FRANCESCA PIACENTINI1

1Department of Movement, Human and Health Sciences, University of Rome Foro Italico, Rome, Italy & 2Department ofMedical, Surgical and Neuro Sciences, University of Siena, Siena, Italy

AbstractPurpose: Different aspects of pacing in endurance events have been investigated, however, there are very limited informationon pacing strategies during open-water swimming. The aim was to describe and compare the pacing profile used by male andfemale open-water swimmers (OW-swimmers) during the 5-, 10- and 25 km races in the main international competitions.Methods: A total of 438 performances were analysed for 5 km, 579 for 10 km and 189 for 25 km, from 2012 to 2017.Swimmers were divided into four groups based on finishing time. G1 whose finishing times were within 0.5% of thewinner’s time, G2 between 0.51% and 1% slower than winner’s time; G3 between 1.1% and 2% slower than winner’stime; G4 over 2% of winner’s time. Kolmogorov–Smirnov test was used to verify the normal distribution of data andrepeated measures ANOVA was performed. Results: G1 adopted a negative pacing and significantly increased the speedin the last split compared with the other groups during the 5-, 10- and 25-km races in both males and females (p< .001).During the 5- and 10-km race, the last split speed of G1 was significantly faster compared to the other groups in bothmales and females (p< .05). Conclusions: OW-swimmers that used a conservative approach remaining in G1 until thefinish of the race, increase the possibility to win a medal in the main international competitions.

Keywords: Endurance, performance, competition

Highlights. Fastest open-water swimmers adopted a negative pacing during the 5-, 10-, and 25-km races in the main international

competitions.. The last split in both female and male fastest group was significantly faster compared to the slowest groups.. A conservative pacing (remaining in the lead pack for the majority of the race and increasing speed only in the last split)

allows to optimize the benefits of drafting, to reduce the energy cost of swimming and to increase the possibility to wina medal.

. The coaches should include training sessions on negative pacing to improve the ability of athletes to increase speed in afatigued condition.

Introduction

Open-water swimming (OWS) is an endurance sport,present at World and Continental Championshipswith the 5, 10 and 25 km events, with the 10 km theonly event present at the Olympic games (OG)since 2008. A multi-lap 2500-m long course isusually used for all distances to allow regularfeeding during the race (Baldassarre, Bonifazi,Zamparo, & Piacentini, 2017; FINA, 2017). Duringthe Budapest World Championship (2017), the bestmale and female swimmers completed the 5-km

race in 54:31.4 and 59:07.0, the 10-km race in1:51:58.5 and 2:00:13.7, the 25-km race in5:02:46.4 and 5:21:58.4, respectively (h:mm:ss.0;http://www.omegatiming.com). An optimal pacingprofile has been shown to be important for successin endurance races (Hanley, 2014a, 2015, 2016;Mytton et al., 2015; Renfree & Gibson, 2013).Pacing strategy is the process whereby humans regu-late their rate of energy expenditure in a way thatallows them to complete a task (such as an endurancerace) in the minimal time while controlling the

© 2018 European College of Sport Science

Correspondence: Maria Francesca Piacentini Department of Movement, Human and Health Sciences, University of Rome “Foro Italico”,P.za L. De Bosis, 15, 00135 Rome, Italy. E-mail: mariafrancesca.piacentini@uniroma4.it

European Journal of Sport Science, 2018https://doi.org/10.1080/17461391.2018.1527946

magnitude of homeostatic disturbance (Foster et al.,2012). Pacing strategies can be categorised as positivepacing (the athlete’s speed decreases during the com-petition), negative pacing (increasing speed), evenpacing (a stable speed is maintained), variablepacing (speed is varied throughout), parabolic U-shaped pacing (speed is greater at the beginningand in the last portion of the race), parabolic J-shaped pacing (finishing speed is greater than thestarting speed) and parabolic reverse J-shapedpacing (starting speed is greater than the finishingspeed) (Abbiss & Laursen, 2008).Several factors have been shown to influence

pacing adopted during a race, e.g.: the difficulty ofaccelerating at the beginning of the race, the magni-tude of slowing down resulting from the loss ofpower output due to fatigue, the power losses to theenvironment, and the amount of essentially wastedkinetic energy at the end of the race (de Koninget al., 2011).Despite several aspects ofpacing inenduranceevents

have been investigated in recent years (Hanley, 2014a,2015, 2016; Renfree & Gibson, 2013), scientificresearch examining the pacing strategies of swimmersemployed during official competitions is scarce(Abbiss & Laursen, 2008). Specifically, mainly poolevents have been investigated. International level ath-letes have been shown to adopt a parabolic U-shapedpacing during 800-m (female) and 1500-m (male)swimming events (Lipinska, Allen, & Hopkins,2016a, 2016b; McGibbon, Pyne, Shephard, &Thompson, 2018) with a fast start and a final endspurt. The ability to produce an end spurt in the finallap(s) are key in pool events (McGibbon et al., 2018).Moreover, pacing patterns have been shown todisplay a low variability from heats to finals, indepen-dently on final time. Therefore pacing profile in eliteswimmers seems reproducible in training and withindifferent competitions, because the athletes are isolatedin their own lane, compared to track races where com-petitors share the same lane (Skorski, Faude, Caviezel,& Meyer, 2014).Mytton et al. (2015) compared the pacing of 1500-

m running and 400-m swimming. Although the dur-ation and the net energetics of both events are compar-able, the authors showed that the 400-m swimmersadopted a U-shaped pacing as previously reportedwhile the 1500-m runners adopted a J-shapedpacing. Another interesting finding was that medallistsand non-medallists of both events differed in theability to increase speed in the last part of the race.In competitive swimming, most of the pool events

are in the range of about 21 s to 15 min, and all ofthem are supported by some combination of phos-phate energy, anaerobic glycolysis, and aerobic com-bustion of carbohydrate, fat, and protein (Capelli,

Pendergast, & Termin, 1998; Pyne & Sharp, 2014).Whereas open-water swimming includes eventsfrom 1 to 5 h and the main substrates oxidizedduring the race are carbohydrate and fat (Zamparo& Bonifazi, 2013). The specific contributions of theenergy systems depend on both the length of therace and the intensity (Pyne & Sharp, 2014). Theopen-water swimming events longer then 5 km maybe considered glycogen depleting (Shaw, Koivisto,Gerrard, & Burke, 2014). A depletion of muscleand liver glycogen are associated with a decline inthe swimming economy at a given speed (Zamparoet al., 2005). Considering the different energydemands of pool swimming compared to open-water swimming, a different distribution of workand energy through the exercise should be adopted.Three studies (Vogt, Rüst, Rosemann, Lepers, &

Knechtle, 2013; Zingg, Rüst, Rosemann, Lepers, &Knechtle, 2014a, 2014b) have analysed the swim-ming performance of elite open-water swimmers(OW-swimmers) in the main international compe-titions (World Cups, European championships,World Championship and OG). These studiesshowed that gender differences during OWS aresmaller compared to other endurance or ultra-endur-ance performances of the same time duration.It has been demonstrated that performance in

longer-duration events is likely to be optimized by anegative pacing strategy with an increase in poweroutput or speed at the end of the event (Fosteret al., 1993; Renfree & Gibson, 2013). In fact, half-marathon runners who won a medal during inter-national events, have been shown to maintain aneven pacing increasing the speed during the final1.1 km, while the slower runners showed a parabolicreverse J-shaped pacing (Hanley, 2015). Marathonmedallists have been shown to adopt an even pace,while the slower marathon runners show a positivepace both in World Championship (WC) and OG(Hanley, 2016). During longer events such as ultra-endurance races (>4 h) athletes progressivelyreduce speed, resulting in the adoption of a positivepace (Abbiss & Laursen, 2008).Differently, from running, a common tactic

observed during OWS competitions is to stay in theback of the lead pack or behind the leader of therace benefiting from drafting (Chatard & Wilson,2003) and increasing speed only during the last lap(Munatones, 2011). Only one study, (Rodriguez &Veiga, 2017) investigated the pacing profile ofopen-water swimmers in the 10-km race during theWC of Kazan (2015). The fastest swimmersadopted a conservative starting strategy with a nega-tive pacing profile, increasing the chances of racesuccess. The fastest swimmers of both genderswere, in fact, able to increase speed by ∼3% during

2 R. Baldassarre et al.

the last quarter of the race (2.5 km) compared to thefirst. Chatard and Wilson (2003) showed a reductionin the energy cost of swimming of 11–38% swimmingbehind another swimmer at a distance of 0–50 cm.Remaining in the pack for the majority of the race isadvantageous to save energy for the critical stages.Moreover, swimmers utilizes most of the energy toovercome water resistance (Zamparo, Capelli, & Pen-dergast, 2011), because the energy cost of loco-motion in swimming is higher compared to othertypes of human locomotion (di Prampero, 1986;Holmer, 1979). The specific characteristics of waterlocomotion suggest that the pacing adopted by ath-letes of other land disciplines of the same time dur-ation (e.g. running and cycling) cannot betransferred to OWS competitions.Moreover, the optimal pacing strategy for a particu-

lar exercise incorporates knowledge of the endpointwith the memory of prior events of similar distance orduration and knowledge of external (environmental)and internal (metabolic) conditions (Gibson et al.,2006). Considering that OWS is a tactical race, andpacing can be influenced by external factors such ascurrents, tides, waves and water temperature (Baldas-sarre et al., 2017), the results obtained inonly one inter-national competition (Rodriguez & Veiga, 2017)cannot be generalized. Although only the 10 km is anOlympic event, the 5 and the 25 km are present atworld and continental championships however nostudy analysed pacing profiles of these distances.Therefore, the aim of the present study was to

describe and compare the pacing profile used bymale and female swimmers, according to perform-ance level, during the 5-, 10- and 25-km races inthe main international competitions (OG; WC andEuropean championships, EC). We hypothesizedthat all OW-swimmers, independently on perform-ance level, would adopt a negative pacing duringthe 5-, 10- and 25-km race.

Methods

Subjects

The official finishing and split times of OWS compe-titions were obtained from the web site of Ligue Eur-opéenne de Natation (LEN, http://www.len.eu),Fédération Internationale de Natation (FINA, www.fina.org) and International Olympic Committee(IOC, www.olympic.org). Only the races where thesplit times were available in the official results wereincluded in the analysis. A total of 7 events wereincluded: OG of 2016 in Rio (23-males and25-females for the 10 km); WC of 2013 in Barcelona(52-males and 42-females for the 5 km; 62-males and49-females for the 10 km; 32-males and 18-females

for the 25 km), 2015 in Kazan (47-males and38-females for the 5 km; 69-males and 51-females forthe 10 km; 24-males and 17-females for the 25 km)and 2017 in Budapest (61-males and 57-females forthe 5 km; 65-males and 59-females for the 10 km;25-males and 18-females for the 25 km); EC of 2012in Piombino (34-males and 19-females for the 5 km;35-males and 20-females for the 10 km; 9-males and4-females for the 25 km), 2014 in Berlin (24-malesand 22-females for the 5 km; 37-males and28-females for the 10 km; 14-males and 13-femalesfor the 25 km) and 2016 in Hoorn (24-males and18-females for the 5 km; 31-males and 23-females forthe 10 km; 10-males and 10-females for the 25 km).Athletes who did not start or finish or who were

disqualified were not included in the analysis. Atotal of 438 (242-male and 196-female) performanceswere analysed for the 5 km, 579 (322-male and 255-female) for the 10 km, and 180 (114-male and 80-female) for the 25 km from 2012 to 2017. Becausedata are public and available on the internet, noformal ethics committee approval was necessary.

Data analysis

The study was designed as an observational research todescribe thepacingprofile ofOW-swimmers in themaininternational competitions. The final race times of eachswimmer were converted in speed (m s−1). Split timeswere obtained for each 2.5 km (1st=2.5 km,2nd=5 km, 3rd=7.5 km, 4th=10 km, 5th=12.5 km,6th=15 km, 7th=17.5 km ,8th=20 km, 9th=22.5 km,10th=25 km) and the mean speed (m s−1) for eachsplit was calculated.A positive pacing was defined when the swimmers

decreased significantly speed in the last split comparedto the first, and a negative pacing when the swimmersincreased significantly speed in the last split comparedto the first.An evenpacingwasdefinedwhen the speeddid not change throughout the race.Swimmers in each racewere divided into four groups

based on finishing time, and each swimmer assigned inone grouponly.Group1 (G1; 61-males and 24-femalesfor the 5 km, 112-males and 85-females for the 10 km,and 26-males and 25-females for the 25 km)whose fin-ishing times were within 0.50% of the winner’s time,Group 2 (G2; 40-males and 27-females for the 5 km,51-males and 24-females for the 10 km, and 11-malesand 4-females for the 25 km) between 0.51% and 1%slower than winner’s time; Group 3 (G3; 29-malesand 17-females for the 5 km, 46-males and 33-femalesfor the 10 km, and 15-males and 12-females for the25 km) between 1.1% and 2% slower than winner’stime; Group 4 (G4; 112-males and 128-females forthe 5 km, 113-males and 113-females for the 10 km,and 62-males and 39-females for the 25 km) over 2%

Pacing profile in the main international open-water swimming competitions 3

ofwinner’s time.Thesepercentagewas selectedaccord-ing to the performance density of OWS competitionsobserved in previous studies (Baldassarre et al., 2017;Vogt et al., 2013; Zingg et al., 2014a, 2014b). More-over, the number or races that each athlete performedduring each championship was calculated.

Statistical analysis

Data are presented asmean ± standard deviation (SD).All statistical analysiswas performedusing the statisticalsoftware PASW statistics 22 (SPSS Inc, Chicago, Illi-noisUnited States). All data were tested for normal dis-tribution using a Kolmogorov–Smirnov test.The sample was split by gender and for all race dis-

tances and a separated mixed ANOVA for repeatedmeasures was performed to assess differences in thebetween splits for each group, considering the splits (i.e.5km, 1st split vs 2nd split) as within factor and group(G1vsG2vsG3vsG4) as between factors.When signifi-cant interactionwas observed, the samplewas divided by

group and an ANOVA for repeated measures was per-formed to assess the differences between the splitsmean speed. When a significant F-value was achieved, apost hoc procedure with Bonferroni correction was per-formed to locate the differences. In addition, one-wayANOVA with Bonferroni post hoc test was performed tocompare the mean speeds of last split between thegroups. The level of significance was set at p≤ .05.

Results

Number of races performed by each athlete during eachevent

During the WC and EC analysed between 2012 and2017, 471 athletes competed in only 1 race per cham-pionship (178 athletes only in the 5 km, 195 only inthe 10 km and 98 only in the 25 km). A total of 332athletes competed in 2 races, (220 athletes competedin both the 5- and 10 km, 23 competed in both the 5-and 25 km and 89 in both the10- and 25 km). Only19 athletes competed in all three distances in asingle championship.

Table I. Percentage change (Δ) in mean speed compared to the previous race splits of female and male swimmers during 5-, 10- and 25-kmrace (mean ± SD).

Percentage change in mean speed (%)

Δ 1st–2nd Δ 2nd–3rd Δ 3rd–4th Δ 4th–5th Δ 5th–6th Δ 6th–7th Δ 7th–8th Δ 8th–9th Δ 9th–10th

5-km WomenG1 7.5 ± 2.4G2 7.1 ± 1.4G3 5.3 ± 2.6G4 2.6 ± 3.0

MenG1 7.4 ± 2.0G2 6.0 ± 2.9G3 5.3 ± 3.0G4 2.6 ± 3.3

10-km WomenG1 1.8 ± 2.7 0.0 ± 1.9 5.6 ± 3.3G2 2.2 ± 2.2 1.8 ± 3.0 4.0 ± 1.7G3 2.1 ± 2.4 0.4 ± 1.9 2.3 ± 1.8G4 −0.4 ± 3.8 –2.7 ± 3.3 0.0 ± 3.8

MenG1 3.5 ± 2.7 0.5 ± 2.1 5.8 ± 3.8G2 3.1 ± 1.8 0.1 ± 1.4 4.4 ± 2.5G3 2.7 ± 2.7 −1.1 ± 1.6 3.1 ± 3.6G4 0.1 ± 4.3 −2.9 ± 3.1 −1.4 ± 4.7

25-km WomenG1 2.7 ± 2.6 −1.5 ± 2.5 0.5 ± 1.9 2.9 ± 7.8 2.0 ± 4.0 1.8 ± 4.2 0.3 ± 1.9 −4.4 ± 2.2 4.0 ± 2.3G2 1.8 ± 1.2 1.2 ± 2.2 3.3 ± 0.7 −7.7 ± 3.7 5.0 ± 0.4 9.6 ± 1.6 −1.7 ± 2.1 −5.6 ± 2.3 3.8 ± 1.6G3 1.1 ± 3.1 −1.0 ± 2.1 0.8 ± 2.0 0.1 ± 6.9 3.4 ± 3.4 3.3 ± 4.9 −4.4 ± 3.6 −3.6 ± 2.7 2.5 ± 4.1G4 2.2 ± 3.0 −0.2 ± 2.1 0.7 ± 2.4 −0.7 ± 6.2 0.5 ± 4.7 −0.2 ± 4.6 −4.0 ± 3.6 −2.3 ± 2.6 2.7 ± 3.8

MenG1 4.7 ± 3.2 2.4 ± 2.8 0.7 ± 3.0 −0.2 ± 4.8 0.4 ± 2.3 −0.9 ± 2.0 0.2 ± 3.1 0.6 ± 1.3 4.0 ± 2.5G2 3.4 ± 2.4 1.6 ± 2.8 1.7 ± 2.8 −1.4 ± 2.9 0.8 ± 2.7 −0.3 ± 0.8 −0.4 ± 2.1 0.4 ± 1.1 −0.6 ± 2.2G3 4.5 ± 2.5 1.4 ± 2.6 2.0 ± 2.8 −1.6 ± 3.1 1.2 ± 2.3 −0.7 ± 2.8 0.6 ± 3.0 −1.8 ± 2.4 −2.4 ± 3.2G4 3.1 ± 3.0 1.1 ± 1.9 −0.3 ± 3.6 1.1 ± 5.8 −1.2 ± 3.9 −3.1 ± 3.7 −3.1 ± 4.7 −2.5 ± 4.0 2.0 ± 4.5

4 R. Baldassarre et al.

Pacing

The percentage change in mean speed compared tothe previous race splits of female and male swimmersduring 5-, 10- and 25-km race is reported in Table I.

5 km

The mean speed for each group of female swimmersis shown in Figure 1(A). A significant main effect wasobserved within splits (F1, 192=560.73, p< .001).A significant interaction effect between splits andgroups was noted (F3, 192= 73.42, p< .001), indicat-ing that the pacing differed within the groups.All groups adopted a negative pacing. A significant

increase in speed was observed between the 1st and2nd split in G1 (F1, 38=559.09, p < .001), G2 (F1, 25-

=632,44, p< .001) G3 (F1, 12

=39.55, p < .001), andG4 (F1, 117=76,88, p < .001). In the last split therewas significant difference in the mean speedbetween groups (F3, 192=62.70, p< .001): G4 wassignificantly slower compared to G1 (p< .001), G2(p< .001) and G3 (p < .001).The mean speed for each group of male swimmers

is shown in Figure 1(B). A significant main effect wasobserved within splits (F1, 238=675.82, p < .001). Asignificant interaction effect between splits and

groups was noted (F3, 238=46.73, p< .001), indicat-ing that the pacing differed within the groups.All groups adopted a negative pacing. A

significant increase in speed was observed betweenthe 1st and 2nd split in G1 (F1, 60=981.94,p< .001), G2 (F1, 39=182,13, p < .001) G3(F1, 28=95.35, p< .001), and G4 (F1, 111=67,23,p< .001). In the last split, there was significantdifference in the mean speed between groups(F3, 238=96.17, p< .001): G4 was significantlyslower compared to G1 (p < .001), G2 (p < .001)and G3 (p < .001).

10 km

The mean speed for each group of female swimmersis shown in Figure 2(A). A significant main effect wasobserved within splits (F3, 753=85.72, p< .001). A sig-nificant interaction effect between splits and groupswas noted (F9, 753=61.35, p< .001), indicating thatthe pacing differed within the groups. Follow-upanalysis showed a significant increase of speed inthe 2nd and 4th split compared to the 1st and 3rdsplit, respectively, in G1 (F3, 252=143.65, p < .001)and G2 (F3, 69=68.48, p< .001). G3 showed a signifi-cant increase of speed in the 2nd and 4th split

Figure 1. Mean speeds (m s−1) for each group of female (A) andmale (B) swimmers during the 5-km race. Differences between pre-vious splits are shown as: ∗p< .05. Differences from the othergroups are shown as; c p< .05.

Figure 2. Mean speeds (m s−1) for each group of female (A) andmale (B) swimmers during the 10-km race. Differences betweenprevious splits are shown as: ∗ p< .05. Differences from the 1stsplit are shown as: § p< .05. Differences with G3 and G4 areshown as: a p< .05. Differences with G4 are shown as: b p< .05.

Pacing profile in the main international open-water swimming competitions 5

compared to the 1st and 3rd split, respectively, and asignificant decrease of speed in the 3rd split com-pared to the 2nd split (F3, 96=38.81, p < .001). A sig-nificant increase in speed was observed between the1st and 4th split in G1 (p< .001), G2 (p < .001)and G3 (p < .001). G4 showed a significant decreaseof speed in the 3rd split compared to the 2nd split andbetween the 1st and 4th split (F3, 336=35.76, p< .001). In the last split, there was a significant differ-ence in the mean speed between groups (F3,

251=263.55, p< .001): G1 and G2 were significantlyfaster compared to G3 (G1, p= 0.05; G2, p=0.015) and G4 (p< 0.001).The mean speed for each group of male swimmers

is shown in Figure 2(B). A significant main effect was

observed within splits (F3, 954=147.93, p < .001). Asignificant interaction effect between splits andgroups was noted (F9, 954=11.01, p< .001), indicat-ing that the pacing differed within the groups.Follow-up analysis showed a significant increase inspeed in the 2nd and 4th splits compared to the 1stand 3rd splits, respectively, in G1 (F3, 333=389.51,p< .001), G2 (F3, 150=210.51, p < .001) and G3(F3, 135=43.45, p< .001). A significant increase inspeed was observed between the 1st and 4th split inG1 (p< .001), G2 (p < .001) and G3 (p< .001). G4showed a significant decrease of speed after the 2ndsplit in time course and between the 1st and 4thsplit (F3, 336=44.3192, p < .001). In the last split,there was significant difference in the mean speed

Figure 3. Mean speeds (m s–1) for each group of female (A) and male (B) swimmers during the 25-km race. Differences between previoussplits are shown as: ∗ p< .05. Differences from the 1st split are show as: § p< .05. Differences with G3 and G4 are shown as: a p< .05. Differ-ences with G4 are shown as: b p< .05. Differences from the other groups are shown as; c p< .05.

6 R. Baldassarre et al.

between groups (F3, 318=263.55, p < .001): G1 andG2 were significantly faster compared to G3 (G1, p< .001; G2, p= .031) and G4 (p < .001).

25 km

The mean speed for each group of female swimmersis shown in Figure 3(A). A significant main effect wasobserved within splits (F9, 684=15.36, p < .001). Asignificant interaction effect between splits andgroups was noted (F27, 684=9.51, p< .001), indicat-ing that the pacing differed within the groups.Follow-up analysis showed a significant decrease of

speed in the 9th split compared to the8th split and a sig-nificant increase in speed in the 10th split compared tothe 9th split inG1 (F9, 216=18.90, p< .001) andG2 (F9,

27=29.50, p< .001). A significant increase in speedwasobserved between the 1st and the 10th split (p< .001)in G1 (p< .001) and G2 (p< .001). G3 showed a sig-nificant decrease of speed in the 9th split compared tothe 8th split (F9, 99=6.61, p= .047). G4 showed a sig-nificant decrease of speed between the 7th and 9thsplits (F9, 342=14.82, p< .001), and a significantlyincreasing speed in the 10th split compared to the 9thsplit (p= .002). G3 (p= 1.00) and G4 (p= 1.00) didnot show a significant difference in speed between the1st and the 10th split. In the last split, there was signifi-cant difference in the mean speed between groups (F3,

76=24.00, p< .001): G1 and G2 were significantlyfaster compared to G3 (G1, p= .005; G2, p= .05)and G4 (G1, p< .001; G2, p= .005).The mean speed for each group of male swimmers

is shown in Figure 3(B). A significant main effect wasobserved within splits (F9, 990=24.39, p < .001). Asignificant interaction effect between splits andgroups were noted (F27, 990=17.95, p< .001), indi-cating that the pacing differed within the groups.Follow-up analysis showed a significant increase ofspeed in the 2nd (p< .001), 3rd (p= .014) and 10th(p< .001) split compared to the 1st, 2nd and 9thsplit, respectively, in G1 (F9, 225=35.740). A signifi-cant increase in speed was observed between the 1stand the 10th split (p< .001).G2 showed a significant increase in speed in the 2nd

split compared to the 1st split (F9, 90=8.895, p= .041).No significant differences were observed in speedbetween the 3rd and the 10th split (p= 1.00). A sig-nificant increase in speed was observed between the1st and 10th split (p = .009). G3 showed a significantincrease in speed in the 2nd split compared to the1st split (F9, 126=15.859, p < .001), while no signifi-cant differences were observed in speed between the1st and the 10th split (p= .35). G4 showed a signifi-cant decrease of speed between the 6th and 9th split(F9, 549=52.560, p < .001) and between the 1st and10th split (p= .003).

In the last split, there was significant difference in themean speed (F3, 110=63,24, p< .001): G1 was signifi-cantly faster in the last split compared to G2(p= .002), G3 and G4 (p< .001). G2 was significantlyfaster in the last split compared to G4 (p< .001).

Discussion

The aim of the present study was to describe andcompare the pacing profile of male and female swim-mers during the 5-, 10- and 25-km OWS races in themain international competitions.The main finding of the present study supports our

hypothesis, that both male and female fastest swim-mers adopted a negative pacing compared with theslower swimmers. In the last split, the speed signifi-cantly increased by ∼7%, ∼6%, ∼4% in bothfemales and males of G1 during the 5-, 10- and 25-km races, respectively. Moreover, the last split inboth female and male fastest groups (G1 and G2)was significantly faster compared to all other groupsin all three distances.We can compare our data with only 1 previous

study that analysed pacing during the 10-km race ofthe 2015 World Championship in Kazan andreported that male medallists adopted a negativepacing and were located very far from the first pos-itions (40th, 43rd and 57th places) in the first lap,then slowly progressed towards the lead pack in thefollowing laps, and finally increased their speedduring the final lap (Rodriguez & Veiga, 2017). Wetook into consideration the overall 579 performanceson the 10-km distance between 2012 and 2017 andfor males found similar results. For females instead,Rodriguez and Veiga (2017) reported that the bestfemale athletes adopted an even pacing strategy andthe athletes from the other groups decreased speedand dropped off; while we showed that the bestfemales, similarly to the males race, adopted a nega-tive pacing significantly increasing speed in the lastlap of the race. Interesting in fact is that both malesand females adopt a similar strategy in the 10-kmrace (a race that lasts approximately 1 h 50 min formales and 2 h for females), a result that seems in con-tradiction with the gender differences observed inpacing during international marathon competitions.Male medallists maintain an even pace throughoutthe race, whereas the female medallists adopt amore conservative initial pace increasing the speedin the final 2.2 km (Hanley, 2016).The pacing profile observed during the 25-km race

showed that both females and males of G1 adopted anegative pacing, while the slowest swimmers adopteda positive or an even pacing. An interesting finding isthe decrease in speed during the 5th or 9th split offemale swimmers during the 25-km race. This drop

Pacing profile in the main international open-water swimming competitions 7

of speed, may be mainly explained by tactical reasonsdue also by the organization of the race (FINA,2017). During the race, females start 10–15 minafter the males, with the probability that female swim-mers will be overtaken by a lap. This delay has animportant influence on the female races, becauseduring the middle or the end of the race (approxi-mately between the 6th and 8th split) male swimmersstart increasing the speed while simultaneously thefemale swimmers decrease the speed waiting to beovertaken (personal communication of an OWScoach). The unification of the female group withthe male group allows the fastest female swimmersto benefit from drafting behind male swimmers andallows them to increase speed and leave behind theother swimmers. This means that the potentialwinners need to be correctly positioned in the frontof the pack at this stage of the race.Although a negative pacing profile is thought to

improve performance by reducing the rate of carbo-hydrate depletion (Abbiss & Laursen, 2005), lower-ing excessive oxygen consumption (Sandals, Wood,Draper, & James, 2006) and lowering blood lactateconcentration (Sandals et al., 2006), during ultra-endurance events (>4 h), athletes tend to adopt apositive pacing strategy (Abbiss and Laursen(2008). Angehrn, Rüst, Nikolaidis, Rosemann, andKnechtle (2016) reported that elite Ironman athletesadopted a positive pacing, decreasing their speed incycling and running, in most races of 2014.However, Ironman races are no draft during

cycling. Therefore, the reasons for a positive pacingstrategy in these distances may be related to thedecline in neuromuscular activity and continuousreduction in skeletal muscle glycogen contentduring competition (Rauch, 2005; Vleck, Bentley,Millet, & Burgi, 2008; Wu, Abbiss, Peiffer, Brisswal-ter, & Nosaka, 2014).Therefore, the negative pacing profile adopted by

OW-swimmers also in the 25 km may be explainedby the beneficial effect of drafting through the race.Swimming behind another swimmer at a distance of0 to 50 cm reduces by 11–38% the metabolicresponse of the draftee (Chatard & Wilson, 2003).A sheltered position in fact allows the swimmer tocontrol the race and to save energy for the criticalmoments (start/end) and thus decreases significantlythe metabolic cost (Chatard & Wilson, 2003). Theenergy cost of swimming (CS) represents the energyexpended to cover one unit of distance at a givenspeed and with a given stroke rate (SR) (Zamparoet al., 2011). Zamparo et al. (2005) reported anincrease in CS and stroke rate (SR) due to develop-ment of fatigue in OWS and a consequent decreasein stroke length (SL). Since the distance per strokeis an index of propelling efficiency, the deterioration

of stroke mechanics in fatigued subjects could beexpected to lead to a progressive increase in CS(Zamparo et al., 2005). Therefore, the data of thepresent study showed that the fastest OW-swimmersin all three distances adopted a conservative pacing byremaining in the lead pack for the majority of the race,increasing speed only in the last split. This pacingstrategy allows to optimize the benefits of drafting,to reduce the energy cost of swimming and toincrease the possibility to win a medal especially forthe fastest swimmers (Rodriguez & Veiga, 2017).Considering that energy cost of locomotion in swim-ming is higher compared to other types of humanlocomotion (di Prampero, 1986; Holmer, 1979),minimizing the energy expenditure (e.g. drafting) isimportant during OWS competitions more thanduring running. Swimming alone most of the raceis equivalent to an increase in the energy expenditurewithout any advantage. In fact, during the male 10-km the Olympic Games that took place in Rio in2016 , the athlete Jarrod Poort swam alone in thefirst position for 7.5 km but he was caught by thelead pack during the last split and finished the racein 21st position (www.olympic.org).This type of strategy is similar to what reported by

sprinters road cyclists, that remain in the group forthe majority of the race and within the last 10 minprior to the sprint are required to find the best positionswithin the pack before the final kilometres, when thespeed dramatically increases for the final end spurt(Menaspà, Quod, Martin, Peiffer, & Abbiss, 2015).Therefore, both OW-swimmers and road cyclistsneed to remain in a sheltered position in the packduring in the first phases of the race and thereafter pos-ition themselves accordingly to be able to increasespeed without remaining blocked by other athletes.Because swimmers remain packed for the major

part of the race and sprint for the final 400–500metres the ability of OW-swimmers is to correctlydecide on the timing of increasing speed in the lastpart of the race for the end spurt. The end spurt istypically observed also in 800- and 1500-m swim-ming events (Lipinska et al., 2016a, 2016b; McGib-bon et al., 2018), marathon running (Hanley, 2016)and 3000-m cycling time trials (Foster et al., 2004).The athletes appear to distribute their energeticresources over the duration of the event in amanner that preserves the ability to contribute tomuscular power output from anaerobic sourceseven during the closing stages of the event (Fosteret al., 2004).It must be pointed out that many of these athletes

compete in 2 or 3 races at major championships.Approximately 47% of the athletes performed morethan 1 race during major championships and only53% competed only in 1 race. Considering that the

8 R. Baldassarre et al.

EC andWC last one week, the pacing strategy may beaffected by the fatigue residual of a previous race,considering that only two days of rest are usuallygranted between races.A limitation of the present study was that we took

into consideration only the official splits while infor-mation on the last 500 metres would have beenmuch more significant. Most of the races aredecided in the last metres of the competition,however the FINA provides split data every 2.5 km.A second limitation of the study was the impossibilityto estimate the impact of environmental conditionson the pacing strategy during each race.

Conclusions

The main finding of the present study indicated thatboth male and female OW-swimmers adopted anegative pacing strategy during the 5-, 10- and 25-km competitions, increasing significantly the speedduring the last split of the races. On the contrary,the slowest swimmers adopted a positive or an evenpacing. OW-swimmers that use a conservativeapproach increase the possibility to win a medalduring the main international competitions.The data of the present study can help coaches and

athletes during training to develop an optimal pacing.The training schedule should include several exer-cises on negative pacing to in order to significantlyincrease speed in a fatigued condition.

Disclosure statement

No potential conflict of interest was reported by the authors.

References

Abbiss, C. R., & Laursen, P. B. (2005). Models to explain fatigueduring prolonged endurance cycling. Sports Medicine, 35(10),865–898. doi:10.2165/00007256-200535100-00004

Abbiss, C. R., & Laursen, P. B. (2008). Describing and under-standing pacing strategies during athletic competition. SportsMedicine, 38(3), 239–252. doi:10.2165/00007256-200838030-00004

Angehrn, N., Rüst, C. A., Nikolaidis, P. T., Rosemann, T., &Knechtle, B. (2016). Positive pacing in elite ironman triathletes.Chinese Journal of Physiology, 59(6), 305–314. doi:10.4077/CJP.2016.BAE418

Baldassarre, R., Bonifazi, M., Zamparo, P., & Piacentini, M. F.(2017). Characteristics and challenges of open-water swimmingperformance: A review. International Journal of Sports Physiologyand Performance, 12(10), 1275–1284. doi:10.1123/ijspp.2017-0230

Capelli, C., Pendergast, D. R., & Termin, B. (1998). Energetics ofswimming at maximal speeds in humans. European Journal ofApplied Physiology and Occupational Physiology, 78(5), 385–393. doi:10.1007/s004210050435

Chatard, J.-C., & Wilson, B. (2003). Drafting distance in swim-ming. Medicine & Science in Sports & Exercise, 35(7), 1176–1181. doi:10.1249/01.MSS.0000074564.06106.1F

de Koning, J. J., Foster, C., Lucia, A., Bobbert, M. F., Hettinga, F.J., & Porcari, J. P. (2011). Using modeling to understand howathletes in different disciplines solve the same problem:Swimming versus running versus speed skating. InternationalJournal of Sports Physiology and Performance, 6(2), 276–280.doi:10.1371/journal.pone.0001552;

di Prampero, P. E. (1986). The energy cost of human locomotionon land and in water. International Journal of Sports Medicine, 7(2), 55–72. doi:10.1055/s-2008-1025736

FINA. (2017). Open Water Swimming Rules 2017–2021. Retrievedfrom https://www.fina.org/content/open-water-swimming-rules

Foster, C., Dekoning, J. J., Hettinga, F., Lampen, J., Dodge, C.,Bobbert, M., & Porcari, J. P. (2004). Effect of competitive dis-tance on energy expenditure during simulated competition.International Journal of Sports Medicine, 25(3), 198–204.doi:10.1055/s-2003-45260

Foster, C., Malterer, K., de Koning, J. J., Bischel, S., Casalino, E.,Malterer, K. … Van Tunen, J. (2012). Pacing strategies forendurance performance. In I. Mujika (Ed.), Endurance training:Science and practice (pp. 89). Vitoria-Gasteriz Basque: InigoMujika.

Foster, C., Snyder, a. C., Thompson, N. N., Green, M. a., Foley,M., & Schrager, M. (1993). Effect of pacing strategy on cycletime trial performance. Medicine and Science in Sports andExercise, 25(3), 383–388. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/8455455

Gibson, A. S. C., Lambert, E. V., Rauch, L. H. G., Tucker, R.,Baden, D. A., Foster, C., & Noakes, T. D. (2006). The roleof information processing between the brain and peripheralphysiological systems in pacing and perception of effort. SportsMedicine (Auckland, N.Z.), 36(8), 705–722. doi:10.2165/00007256-200636080-00006

Hanley, B. (2014a). Senior men’s pacing profiles at theIAAF world cross country championships. Journal of SportsSciences, 32(11), 1060–1065. doi:10.1080/02640414.2013.878807

Hanley, B. (2015). Pacing profiles and pack running at the IAAFworld half marathon championships. Journal of Sports Sciences,33(11), 1189–1195. doi:10.1080/02640414.2014.988742

Hanley, B. (2016). Pacing, packing and sex-based differences inolympic and IAAF world championship marathons. Journal ofSports Sciences, 34(17), 1675–1681. doi:10.1080/02640414.2015.1132841

Holmer, I. (1979). Physiology of swimming man. Exercise and SportSciences Reviews, 7, 87–123.

Lipinska, P., Allen, S. V., & Hopkins, W. G. (2016a). Modelingparameters that characterize pacing of elite female 800-m free-style swimmers. European Journal of Sport Science, 16(3), 287–292. doi:10.1080/17461391.2015.1013996

Lipinska, P., Allen, S. V., & Hopkins, W. G. (2016b). Modelingparameters that characterize pacing of elite female 800-m free-style swimmers. European Journal of Sport Science, 16(3), 287–292. doi:10.1080/17461391.2015.1013996

McGibbon, K. E., Pyne, D. B., Shephard, M. E., & Thompson, K.G. (2018). Pacing in swimming: A systematic review. SportsMedicine (Auckland, N.Z.). doi:10.1007/s40279-018-0901-9

Menaspà, P., Quod, M., Martin, D. T., Peiffer, J. J., & Abbiss, C.R. (2015). Physical demands of sprinting in professional roadcycling. International Journal of Sports Medicine, 36(13), 1058–1062. doi:10.1055/s-0035-1554697

Munatones, S. (2011). Open water swimming: Improve your time,improve your performance. (Human Kinetics, Ed.). Champaign,IL: Human Kinetics.

Pacing profile in the main international open-water swimming competitions 9

Mytton, G. J., Archer, D. T., Turner, L., Skorski, S., Renfree, A.,Thompson, K. G., & St Clair Gibson, A. (2015). Increasedvariability of lap speeds: Differentiating medalists and nonmed-alists in middle-distance running and swimming events.International Journal of Sports Physiology and Performance, 10(3), 369–373. doi:10.1123/ijspp.2014-0207

Pyne, D. B., & Sharp, R. L. (2014). Physical and energy require-ments of competitive swimming events. International Journal ofSport Nutrition and Exercise Metabolism, 24(4), 351–359.doi:10.1123/ijsnem.2014-0047

Rauch, H. G. L. (2005). A signalling role for muscle glycogen inthe regulation of pace during prolonged exercise. BritishJournal of Sports Medicine, 39(1), 34–38. doi:10.1136/bjsm.2003.010645

Renfree, A., & Gibson, A. S. C. (2013). Influence of different per-formance levels on pacing strategy during the women’s worldchampionship marathon race. International Journal of SportsPhysiology and Performance, 8(3), 279–285.

Rodriguez, L., & Veiga, S. (2017). Effect of the pacing strategies onthe open water 10km world swimming championships perform-ances. International Journal of Sports Physiology and Performance,1–19. doi:10.1123/ijspp.2017-0274

Sandals, L. E., Wood, D. M., Draper, S. B., & James, D. V. B.(2006). Influence of pacing strategy on oxygen uptake duringtreadmill middle-distance running. International Journal ofSports Medicine, 27(1), 37–42. doi:10.1055/s-2005-837468

Shaw, G., Koivisto, A., Gerrard, D., & Burke, L. M. (2014).Nutrition considerations for open-water swimming.International Journal of Sport Nutrition and Exercise Metabolism,24, 373–381. doi:10.1123/ijsnem.2014-0018

Skorski, S., Faude, O., Caviezel, S., & Meyer, T. (2014).Reproducibility of pacing profiles in elite swimmers.International Journal of Sports Physiology and Performance, 9,217–225. doi:10.1123/IJSPP.2012-0258

Vleck, V. E., Bentley, D. J., Millet, G. P., & Burgi, A. (2008). Pacingduring an elite olympic distance triathlon: Comparison betweenmale and female competitors. Journal of Science and Medicine inSport, 11(4), 424–432. doi:10.1016/j.jsams.2007.01.006

Vogt, P., Rüst, C. A., Rosemann, T., Lepers, R., & Knechtle, B.(2013). Analysis of 10 km swimming performance of elitemale and female open-water swimmers. SpringerPlus, 2(2004),603. doi:10.1186/2193-1801-2-603

Wu, S. S. X., Abbiss, C., Peiffer, J., Brisswalter, J., & Nosaka, K.(2014). Factors influencing pacing in triathlon. Open AccessJournal of SportsMedicine,5(1), 223.doi:10.2147/OAJSM.S44392

Zamparo, P., & Bonifazi, M. (2013). Bioenergetics of cyclic sportsactivities on land. Walking, running and cycling. In B. Debasis,N. Sreejayan, & K. Sen Chandan (Eds.), Nutrition andenhanced sports performance: Muscle building, endurance, andstrength, 1st edition (pp. 143–150). San Diego, CA: AmericanPress.

Zamparo, P., Bonifazi, M., Faina, M., Milan, A., Sardella, F.,Schena, F., & Capelli, C. (2005). Energy cost of swimming ofelite long-distance swimmers. European Journal of AppliedPhysiology, 94(5–6), 697–704. doi:10.1007/s00421-005-1337-0

Zamparo, P., Capelli, C., & Pendergast, D. (2011). Energetics ofswimming: A historical perspective. European Journal ofApplied Physiology, 111(3), 367–378. doi:10.1007/s00421-010-1433-7

Zingg,M.A.,Rüst,C.A.,Rosemann,T.,Lepers,R.,&Knechtle,B.(2014a). Analysis of sex differences in open-water ultra-distanceswimming performances in the FINA world Cup races in 5 km,10 km and 25 km from 2000 to 2012. BMC Sports Science,Medicine & Rehabilitation, 6(1), 7. doi:10.1186/2052-1847-6-7

Zingg, M. A., Rüst, C. A., Rosemann, T., Lepers, R., & Knechtle,B. (2014b). Analysis of swimming performance in FINA worldCup long-distance open water races. Extreme Physiology &Medicine, 3(1), 2. doi:10.1186/2046-7648-3-2

10 R. Baldassarre et al.

View publication statsView publication stats