Content uploaded by Teun Van Erp
Author content
All content in this area was uploaded by Teun Van Erp on Mar 02, 2023
Content may be subject to copyright.
Sprint Tactics in the Tour de France: A Case Study
of a World-Class Sprinter (Part II)
Teun van Erp, Marcel Kittel, and Robert P. Lamberts
Purpose:To describe the performance and tactical sprint characteristics of a world-class sprinter competing in the Tour de
France. In addition, differences in the sprint tactics of 2 teams and won versus lost sprints are highlighted. Method:Power output
(PO) and video footage of 21 sprints were analyzed. Position in the peloton and number of teammates supporting the sprinter at
different times before the finish line together with PO for different time intervals were determined. Sprints were classified as team
Shimano (2013–2014) and team Quick-step (2016–2017), as well as won or lost.Results:The sprinter was highly successful,
winning 14 out of the 21 sprints. At time intervals 10 to 5, 3 to 2, and 1.5 to 1 minute, POs were significantly lower in team Quick-
step compared with team Shimano, but the sprinter was positioned further away from the front at 10, 2, 1.5, 1, and 0.5 minutes at
team Quick-step compared with team Shimano. The PO was higher at time interval 0.5 to 0.25 minutes before the finish line with
team Quick-step when compared with team Shimano. The position of the sprinter in the peloton in lost sprints was further away
from the front at 0.5 minutes before the finish compared with won sprints, while no differences were noted for PO and the number
of teammates between won and lost sprints. Conclusions:Differences in sprint tactics (Shimano vs Quick-step)influence the PO
and position in the peloton during the sprint preparation. In addition, the position at 0.5 minutes before the finish line influences
the outcome (won or lost) of the sprint.
Keywords:elite, cycling, power output, performance, Grand Tour
The Tour de France (TdF) is one of the 3 Grand Tours (GTs)
on the professional road cycling calendar. It consists of 21 stages
with only 2 or 3 rest days in between.
1
Although there are some
prestigious 1-day races on the World Tour circuit, winning a stage
in the TdF is considered one of the highest achievements possible
in the career of a professional road cyclist. The 21 stages of the
TdF can be categorized into 4 different stage types, namely: “flat
stages,”“semi-mountain stages,”“mountain stages,”and “time
trials.”
2,3
In general, the flat stages and some of the semi-mountain
stages are specifically designed for sprinters, and thus, typically 7
of the 21 TdF stages will end in a peloton sprint. The difference
between winning and losing a sprint often hinges on only a few
centimeters to a wheel length. Winning a peloton sprint in the TdF
is thus incredibly difficult. This is highlighted by the fact that of
the 31 TdF peloton sprints between 2013 and 2017, 28 were won
by only 3 world-class sprinters namely Mark Cavendish, André
Greipel, and Marcel Kittel.
4
Relatively little research has been conducted on flat
stages,
2,3,5,6
sprinting,
7–11
and the sprint preparation (ie, 10 km
before the finish line).
8,9
Menaspa et al
8
studied the physical
demands exerted on professional sprinters during different types
of races (ie, World Tour, Hors Category, and Category 1 races).
Their findings revealed a surprisingly low average power output
(PO) during the race (approximately 200 W). In addition, they
ascertained that intensities more than doubled in the last minute
(approximately 487 W) and peaked at 1020 to 1248 W during the
last 9 to 17 seconds of the sprint.
8
Although this study provided
unique and valuable insights into the physical demands exacted by
sprinting, the study was not performed in “world-class”sprinters,
and only 4 of the analyzed sprints were performed at World Tour
level. Therefore, the physical demands placed upon these world-
class sprinters, especially during events like the TdF, might be
different than those reported by Menaspa et al.
8
The high intensity found in sprint preparation is combined
with tactical decisions made by both the team as well as the
individual sprinter himself. During the final kilometers of the
sprint, a team’s principal aim is to maneuver their sprinter into
the best possible position to win the peloton sprint,
9
all done as
efficiently as possible. The support of teammates is extremely
important as drafting behind teammates could save up to 60% of
energy for the sprinter in the sprint preparation.
12
Therefore, teams
have designated domestiques to support the sprinter in the sprint
preparation and form the so-called “sprint train.”In 1-day races, an
entire team’s goal may be to assist their sprinter, although this
modus operandi is for most teams not possible during a GT. Due to
the high stakes and prestige associated with GTs, teams generally
select a range of different riders for these events. For example, 1
general classification rider, 2 or 3 climbing domestiques, as well as
1 sprinter, and 2 or 3 sprint domestiques. This mixture of riders as
well as their variable roles is believed to impact negatively upon the
size and efficiency of the team’s sprint train. Therefore, it could be
that tactics in the sprint preparations are different between a team
that is fully committed to support the sprinter and a team with a
mixture of riders and thus with only 2 or 3 domestiques designated
to support the sprinter.
This study, therefore, aims to describe tactics (ie, position in
the peloton and the number of supporting teammates) and perfor-
mance characteristics in the sprint (preparation). In addition, as the
sprinter rode for 2 different teams, variances between the ap-
proaches adopted by the teams could be investigated. Furthermore,
this study aims to highlight differences between sprints that were
van Erp and Lamberts are with the Div of Orthopedic Surgery and the Dept of
Sport Science, Faculty of Medicine and Health Sciences, Stellenbosch University,
Stellenbosch, South Africa. Kittel is a retired professional road cyclist. van Erp
(teunvanerp@hotmail.com) is corresponding author.
1371
International Journal of Sports Physiology and Performance, 2021, 16, 1371-1377
https://doi.org/10.1123/ijspp.2020-0701
© 2021 Human Kinetics, Inc. BRIEF REPORT
Brought to you by IJSPP Board Membership | Authenticated teunvanerp@hotmail.com | Downloaded 03/02/23 11:54 AM UTC
won and lost. Furthermore, changes in performance characteristics
in the sprint (preparation) during the TdF are described.
Method
Participant
The participant, a professional cyclist who specialized in sprinting,
competed on the UCI World Tour circuit from 2013 to 2019 and
was very successful. He won 89 cycling races, of which 32 were in
the World Tour and minimal one stage in the 3 GTs. He was most
successful in the seasons 2012, 2013, 2014, 2016, and 2017 and
was able to win 21%, 21%, 19%, 16%, and 22% of the mass start
races he started in, respectively. This cyclist can thus rightfully be
considered as one of the best sprinters of his generation. The
participant provided informed consent for an in-depth analysis of
his PO data, collected during 4 editions of the TdF, while ethical
approval for the study was granted by the Health Research Ethics
Committee of Stellenbosch University (C20/06/018). His best 20-
minutes PO and body weight were: 452 W with 90.0 kg, 465 W
with 90.0 kg, 454 W with 88.5 kg, and 461 W with 89 kg, for the
2013, 2014, 2016, and 2017 TdF editions, respectively, indicating
that the cyclist had a similar level for the analyzed seasons.
Research Design
In total, 21 sprint stages, in which the sprinter was riding for a
win, were included in the study. In 2013 and 2014, the sprinter
rode for team Argos-Shimano and team Giant-Shimano (the same
team and herein referred to as team Shimano), while in 2016 and
2017 he rode for team Etixx—Quick-step and Quick-step Floors
(the same team and herein referred to as team Quick-step).
Although bike brands differed between the teams, the bike setup
and crank length were similar for the different seasons. Due to
sponsor commitments, PO data were collected with different
power meters: SRM POWERMETER (SRM, Jülich, Welldorf,
Germany) at team Shimano (1 Hz) and 4iiii PRECISION PRO
dual-sided (4iiii; 4iiii Innovations Inc, Cochrane, Canada) at team
Quick-step. The PO data were sampled at a frequency of 1 Hz.
The sprinter was aware of the importance of the 0 offset but due to
the retrospective nature of the study, this aspect could not be
controlled. The manufacturer calibrated the SRM power meters
through a static calibration in the preseason, and the 4iiii innova-
tions were calibrated statically at the factory, and they are used for
a maximum period of one season.
Performance and Tactical Characteristics
of Sprints
Performance characteristics in the sprint were analyzed based on
maximal mean PO for multiple durations (ie, 5, 10, and 15 s)
obtained during the last 20 seconds of the sprint stage and the mean
PO for the duration of the sprint. The duration of the sprint was
calculated based on video footage as the time between the moment
that the sprinter started to sprint (ie, moved off the wheel in the
front and began sprinting out of the saddle) and the finish line. The
end of the sprint in the PO data was based on a visual inspection of
the PO and speed. Similar to Menaspa et al,
9
but extended, video
footage was used to analyze the sprinter’s position in the peloton as
well as the number of teammates, who supported him at 10, 5, 3, 2,
1½, 1 minute, 30, and 15 seconds before the finish. Only teammates
who supported the sprinter, thus allowing him to draft and save
energy, while maintaining a good position in the peloton, were
included in the analyses. Mean PO was calculated for 8 different
time frames before the finish, namely: 10 to 5, 5 to 3, 3 to 2, 2 to 1.5,
1.5 to 1, 1 to 0.5 minute, 30 to 15, and 15 to 0 seconds. In stage 15
of the 2014 edition (from 15 s to the finish line) and stage 4 of the
2017 edition (from 60 s to the finish line), the sprinter stopped
competing for the victory and therefore data from those 2 moments
onward were excluded from analyses for these 2 stages. In addition,
there was no helicopter footage for stage 2 of the 2017 edition, and
consequently, it was not possible to analyze the position in the
peloton and the number of teammates.
To investigate whether performance or tactical characteristics
influenced the outcome of the sprint, sprint stages were categorized
into won or lost sprints. In addition, sprints were further categorized
as riding for team Shimano or team Quick-step to investigate
whether team tactics played a role in the performance and/or tactical
decisions of the sprint. It is interesting to note that the sprinter had
a leadership role in team Shimano, while within team Quick-step
he had to share leadership with a general classification contender
who finished 9th and 6th in 2016 and 2017, respectively.
Statistical Analysis
The data were extracted with MATLAB (Release2019b; The
MathWorks, Inc, Natick, MA) and analyzed with SPSS (IBM
SPSS Statistics version 23; IBM Corp, Armonk, NY). The homo-
geneity of the data was tested with Kolmogorov–Smirnov and
Lilliefors tests. Data are expressed as mean (SD). Differences in
performance parameters between won and lost stages as well when
riding for team Shimano or team Quick-step were analyzed with an
independent ttest. A fitting mixed model was used to determine
differences in tactical and performance characteristics in the sprint
preparation between won and lost and between team Shimano and
team Quick-step at different times (intervals). An independent ttest
was used to identify differences when the fitting mixed model
indicated a significant main effect. Statistical significance was
accepted at P<.05. In addition, Cohen deffect sizes (d) were
calculated and interpreted as follows: 0 to 0.19 as trivial, 0.20 to
0.59 as small, 0.6 to 1.19 as moderate, 1.20 to 1.99 as large, and
≥2.00 as very large.
13
Results
In total, PO data of 21 sprints and video footage of 20 sprints, 14
of which resulted in a stage victory, were analyzed from 4 editions
of the TdF. Sprints ranged from 7 to 17 seconds, during which the
mean PO ranged from 1026 to 1576 W (Table 1). Mean speed
during the final bunch sprints varied from 52 to 73 km·h
−1
with a
general gear ratio of 53/11 and mean cadence of 103 to 121
revolutions per minute during the sprint. At team Shimano, 8 out
of 10 sprints were successful (80%), while in team Quick-step,
6 out of 11 sprints were successful (55%).
Team Tactics
Sprint characteristics of the rider did not significantly differ when
riding for team Shimano or team Quick-step (Table 1). The number
of teammates supporting the sprinter did not significantly differ
between team Shimano and team Quick-step, although a moderate
to large (d=0.65–1.29) lower number of teammates supporting the
sprinter was observed at 3, 2, 1.5 minute, and 15 seconds before the
finish line at team Quick-step. Team tactics between team Shimano
IJSPP Vol. 16, No. 9, 2021
1372 van Erp, Kittel, and Lamberts
Brought to you by IJSPP Board Membership | Authenticated teunvanerp@hotmail.com | Downloaded 03/02/23 11:54 AM UTC
and team Quick-step differed substantially. This approach was
reflected in a different position in the peloton at 10 (P=.003), 2
(P=.025), 1½ (P=.002), 1 minute (P=.046), and 30 seconds
(P=.002), while the sprinter was riding moderate to large
(d=0.83–1.73) further from the front at 10, 3, 2, 1.5, 1 minute,
30, and 15 seconds when riding for team Quick-step. In line with
this, mean PO was moderately to largely lower for the time
intervals 10 to 5 (P=.019, d=1.12), 3 to 2 (P=.006, d=1.35),
and 1½ to 1 minute (P=.028, d=1.05) when riding for team
Quick-step compared with when riding for team Shimano (Table 2).
In contrast, as illustrated in Figure 1A, mean PO between time
interval 30 to 15 seconds was largely higher when riding for team
Quick-step compared with team Shimano (P=.001, d=1.86).
Won Versus Lost
As per Table 1, sprint characteristics such as PO, cadence, speed,
and sprint duration were not significantly different between for
Table 1 Mean Maximal POs (1, 5, 10, and 15 s) of the Last 20 Seconds and PO, Sprint Duration, Speed,
and Cadence From the Whole Sprint
Won vs lost Teams
Variable
All sprints
(n =19
ab
)
Won
(n =14)
Lost
(n =5
ab
)
Shimano
(n =9
a
)
Quick-step
(n =10
b
)
1-s PO, W 1737 (94)
[1556–1878]
1736 (104)
[1556–1878]
1741 (65)
[1640–1814]
1722 (109)
[1556–1868]
1751 (82)
[1628–1878]
5-s PO, W 1610 (121)
[1283–1813]
1628 (96)
[1508–1813]
1606 (72)
[1524–1773]
1614 (165)
[1283–1813]
1606 (72)
[1524–1773]
10-s PO, W 1515 (123)
[1231–1701]
1525 (108)
[1296–1688]
1492 (99)
[1296–1701]
1541 (146)
[1231–1688]
1492 (99)
[1296–1701]
15-s PO, W 1383 (167)
[968–1602]
1402 (161)
[968–1602]
1329 (190)
[994–1449]
1408 (173)
[994–1602]
1360 (168)
[968–1538]
Whole sprint PO, W 1411 (117)
[1026–1576]
1441 (71)
[1311–1576]
1326 (181)*
[1026–1519]
1421 (163)
[1026–1576]
1402 (59)
[1311–1519]
Duration, s 13.1 (2.5)
[7–17]
13.2 (2.7)
[7–17]
12.6 (2.3)
[9–15]
13.2 (1.8)
[11–16]
12.9 (3.2)
[7–17]
Mean speed, km·h
−1
65.5 (6.1)
[52–73]
65.2 (6.2)
[52–71]
66.3 (6.4)
[59–73]
64.0 (5.5)
[55–71]
66.6 (6.5)
[52–73]
Mean cadence, rpm 112 (5)
[103–121]
112 (4)
[103–119]
113 (7)
[105–121]
110 (3)
[105–117]
114 (5)
[103–121]
Abbreviation: PO, power output.
No significant difference. *Effect size =0.9 moderate different to won
a
stage 15, edition 2014 and
b
stage 4, edition 2017 are excluded because the sprinter did not participate
in the sprint. Data were collected in the 2013, 2014, 2016, and 2017 editions of the Tour de France of a world-class sprinter. Sprints are categorized in won versus lost
and team Shimano (2013 and 2014) versus team Quick-step (2016 and 2017). Values are presented as mean (SD) [min–max].
Table 2 Number of Teammates and Position in the Peloton at Different Times to the Finish Line Determined
From Video Analysis, Together With PO for Different Time Intervals (10–5 min, 5–3 min, 2–1.5 min, 1.5–1 min,
1 min–30 s, 30–15 s, and the last 15 s) From Sprints at Team Shimano (2013 and 2014) and Team Quick-step
(2016 and 2017) From a World-Class Sprinter in the 2013, 2014, 2016, and 2017 Editions of the Tour de France
Teammates Position Mean power (time intervals)
Time to
finish
Team
Shimano
(n =10,
won =8)
Team
Quick-step
(n =10
a
,
won =5)
Cohen
d
Team
Shimano
(n =10,
won =8)
Team
Quick-step
(n =10
a
,
won =5
Cohen
d
Team
Shimano
(n =10,
won =8)
Team
Quick-step
(n =11,
won =6)
Cohen
d
10 min 5.1 (1.5) 4.8 (1.7) 0.19 (T) 28.9 (14.9) 54.3 (18.9)*1.52 (L) 332 (53) 274 (49)*1.12 (M)
5 min 4.6 (1.2) 3.7 (2.0) 0.59 (S) 28.5 (12.7) 34.4 (16.6) 0.40 (S) 380 (53) 349 (50) 0.60 (M)
3 min 3.9 (1.0) 3.0 (1.5) 0.72 (M) 11.6 (9.2) 21.6 (10.1) 0.83(M) 392 (48) 321 (56)*1.35 (L)
2 min 3.3 (0.7) 2.2 (1.0) 1.29 (L) 7.4 (6.3) 16.6 (8.9)*1.12 (M) 469 (115) 418 (91) 0.49 (S)
1.5 min 2.8 (0.9) 2.2 (0.9) 0.65 (M) 4.4 (3.4) 15.6 (6.1)*1.81 (L) 513 (117) 412 (76)*1.05 (M)
1 min 1.7 (0.9) 1.9 (1.2) 0.18 (T) 5.0 (3.9) 10.0 (3.7)*1.00 (M) 525 (101) 475 (87) 0.54 (S)
30 s 1.2 (0.8) 0.9 (0.6) 0.45 (S) 3.2 (2.4) 8.6 (3.8)*1.73 (L) 654 (119) 893 (139)*1.86 (L)
15 s 0.4 (0.5) 0.1 (0.3) 0.68 (M) 3.2 (2.1) 5.8 (0.52) 0.87 (M) 1403 (172) 1370 (77) 0.27 (S)
Abbreviations: L, large; M, medium; PO, power output; S, small; T, trivial.
*Significantly different (P<.05) from team Shimano.
a
Edition 2017, stage 2 no helicopter footage; therefore, position and number of teammates were not possible to determine.
IJSPP Vol. 16, No. 9, 2021
How to Win Sprints in the Tour de France 1373
Brought to you by IJSPP Board Membership | Authenticated teunvanerp@hotmail.com | Downloaded 03/02/23 11:54 AM UTC
won and lost sprints. In addition, no significant differences were
noted as to the number of teammates, which supported the sprinter
during the sprint preparation in won or lost sprints. The sprinter’s
position in the peloton was largely (P=.042, d=1.16) further from
the front at 30 seconds in the lost sprints compared with won
sprints. As per Table 3, PO at any moment during the sprint
preparation did not significantly differ between won and lost. The
mean PO during the last 3 minute of won versus lost stages is
illustrated in Figure 1B.
Sprint Characteristics and Days in TdF
The maximal mean PO (ie, 5, 10, and 15 s) and mean POs during the
sprint (preparation) for the timeframes (10–5, 3–2min,and
30–15 s before the finish) over the 3 weeks of the TdF are presented
in Figure 2. No relationships were found between any of the maximal
mean POs or mean POs during the sprint (preparation) over the
3-week period. This indicates that similar POs were achieved at the
beginning and at the end of a TdF in the sprint and sprint preparation.
Discussion
To our knowledge, this is the first study to describe the performance
characteristics of a world-class sprinter competing in the TdF. In
addition, the effect of different team tactics on the performance
characteristics during the final preparation phase of the sprint as
well as differences in performance characteristics between success-
ful (win) and less successful (lost) sprints were studied. Although a
study by Menaspa et al
9
has also researched sprint tactics and mean
speed during the final kilometer in a world-class sprinter, these data
were collected based on video analyses rather than direct power
files, as was the case in this study. Furthermore, this study describes
the performance demands in the sprint (preparation) for the 3-week
duration of the TdF.
In line with the world-class status of our cyclist, mean PO
during the entire sprint was substantially higher (approximately
209 W or approximately 1.4 W·kg
−1
) when compared with that
previously reported in professional sprinters competing at a lower
race level.
8
In line with this, peak PO during the sprint was also
substantially higher (489 W or 1.9 W·kg
−1
) in our world-class
sprinter than that reported by Menaspa et al
8
in other professional
sprinters. Interestingly, in both studies, the difference between peak
PO and the mean PO of the total sprint was the same at 23%.
8
This
suggests that the difference between world-class and nonworld-
class is mainly a higher PO in the sprint and not the ability to hold
the mean PO during the sprint. To the best of the author’s
knowledge, the presented absolute and relative PO’s in the sprint
are the highest to have ever been reported in studies analyzing
Figure 1 —Average power output at different time points (in seconds) from the last 3 minutes of the Tour de France. (A) Sprints are categorized in team
Shimano (n =10) and team Quick-step (n =11). (B) Sprints are categorized in won (n =14) and lost (n =7). *Significantly different (P<.05) from team
Quick-step. #Moderate effect size (d>0.60).
IJSPP Vol. 16, No. 9, 2021
1374 van Erp, Kittel, and Lamberts
Brought to you by IJSPP Board Membership | Authenticated teunvanerp@hotmail.com | Downloaded 03/02/23 11:54 AM UTC
performance demands in races
2,3,5,6,8,14,15
and are similar to re-
ported peak POs in track cycling.
16
Team Tactics
This is the first study to investigate the differences between the
sprint tactics of 2 different teams. Team Shimano’stactic was to
ride the last 3 or 2 km in front of the peloton, with the help of 5 or 6
teammates, thus a full sprint train. This tactic aimed to maneuver
the sprinter into the best possible position from which he could then
launch his sprint. In contrast, team Quick-step’s tactic was that the
sprinter, supported by 2 or 3 teammates, would ride between
positions 20 to 10 in the peloton and only move to the front of
the peloton during the last 30 seconds of the sprint (Figure 1and
Table 2). The chosen tactic is highly dependent on the number of
available support riders in the preparation phase of the sprint as
riding in front of the peloton with only 2 to 3 support riders is
impossible. Although a sprinter’s train, as used by team Shimano,
might afford the sprinter an advantage from a positioning point of
view, it might be less effective from a drafting point of view
(Table 2). For example, the fourth position in the sprint train will
result in a PO reduction of 53%
12
while riding in the fourth position
in the peloton results in a PO reduction of 83%.
17
Although drafting
in the peloton might, therefore, be beneficial, the sprinter is
positioned further away from the front as in the case of team
Quick-step (Table 2). Therefore, the sprinter needs to move up in
the last 30 seconds. This results in an incredible high PO for the last
30 seconds, with a mean PO of 893 W between 30 and 15 seconds
before the finish, followed by 1370 W during the final 15 seconds.
It could be that this sprint tactic is only suitable for those sprinters
who can maintain this high PO for 30 seconds. Additional risks
to team Quick-step’s tactics included becoming boxed in or being
trapped behind a crash. It is somewhat speculative, but these
disadvantages could explain team Quick-step’s lower success ratio
when compared with team Shimano (55% vs 80%). A lower
success ratio, however, could also be ascribed to the team’s
experience and/or the number of top sprinters competing in a
specific TdF. In the case of team Quick-step, however, as fewer
teammates were needed to support the sprinter, they could also
successfully support their general classification contender who
finished 9th and 6th in 2016 and 2017, respectively.
Won Versus Lost
The reported number of teammates and positioning in the peloton
during the last 60 seconds of the race were similar to previously
reported values.
9
However, in contrast with the results obtained
by Menaspa et al,
9
this study noted no difference in the number of
teammates supporting the sprinter in won versus lost sprints.
Furthermore, Menaspa et al
9
noted differences of the sprinter’s
position in the peloton at 60, 30, and 15 seconds, while this study
only found differences at 30 seconds between won and lost
sprints. It could be that the results of this study were influenced
by the limited number of lost sprints. In addition, this study shows
that team tactics influence the position of the sprinter in the
peloton in the last 60 seconds. Thus, differing team tactics
adopted by 2 world-class sprinters could very well yield con-
trasting results.
The Sprint Preparation
Similar to Menaspa et al,
8
this study analyzed PO in the sprint
preparation. Higher PO values in the TdF compared with PO values
in the sprint preparation of lower level races were expected.
However, mean PO during the last 5 and 10 minutes before the
finish was substantially higher in professional riders (5 min:
4.1 W·kg
−1
, 10 min: 4.7 W·kg
−1
) when compared with values
observed in our world-class sprinter (5 min: 3.3 W·kg
−1
, 10 min:
3.9 W·kg
−1
). In contrast to the 5- and 10-minute PO, the last 60-
second PO of our world-class sprinter (5.3–9.9 W·kg
−1
) was
substantially higher than that of the professional sprinters
(6.3 W·kg
−1
), as reported by Menaspa.
8
The difference in PO,
as noted in the 2 studies, can be ascribed to the use of different
methods. The time frames used in the studies differed, for example,
this study determined PO from 10 to 5 minutes before the finish
line, while Menaspa et al determined PO from 10 minutes to the
finish line. Therefore, it is somewhat difficult to compare the
reported values in this study to those reported by Menaspa et al.
8
Table 3 Number of Teammates and Position in the Peloton at Different Times to the Finish Line Determined
From Video Analysis, Together With PO for Different Time Intervals (10–5 min, 5–3 min, 2–1.5 min, 1.5–1 min,
1 min–30 s, 30–15 s, and the last 15 s) From Lost and Won Sprints From a World-Class Sprinter in the 2013,
2014, 2016, and 2017 Editions of the Tour de France
Teammates Position Mean power (time interval)
Time to
finish
Won
(n =13
a
)
Lost
(n =7)
Cohen
d
Won
(n =13
a
)
Lost
(n =7)
Cohen
d
Won
(n =14)
Lost
(n =7)
Cohen
d
10 min 4.8 (1.2) 5.3 (2.2) 0.31 (S) 3.9 (16.9) 52.1 (24.9) 0.78 (M) 310 (56) 286 (64) 0.40 (S)
5 min 4.3 (1.4) 3.8 (2.1) 0.27 (S) 32.8 (11.7) 27.7 (19.5) 0.33 (S) 352 (78) 355 (48) 0.05 (S)
3 min 3.4 (1.3) 3.6 (1.4) 0.14 (T) 14 (10.3) 21.4 (17.2) 0.54 (S) 363 (70) 338 (44) 0.44 (S)
2 min 2.8 (1.2) 2.7 (0.8) 0.06 (T) 10.4 (7.3) 15.0 (12.6) 0.46 (S) 427 (116) 473 (71) 0.50 (S)
1.5 min 2.5 (1.1) 2.6 (0.8) 0.12 (T) 9.3 (7.0) 10.4 (11.4) 0.13 (S) 465 (127) 450 (64) 0.16 (T)
60 s 1.8 (1.1) 1.7 (1.0) 0.17 (T) 6.7 (5.7) 8.8 (5.3) 0.39 (S) 492 (94) 518 (105) 0.26 (S)
30 s 1.2 (0.8) 0.8 (0.4) 0.53 (S) 4.5 (4.0)*8.5 (3.0) 1.16 (M) 784 (176) 750 (188) 0.19 (S)
15 s 0.3 (0.5) 0.2 (0.4) 0.32 (S) 3.9 (2.6) 5.5 (4.3) 0.45 (S) 1412 (102) 1321 (185) 0.63 (M)
Abbreviations: M, medium; PO, power output; S, small; T, trivial; TdF, Tour de France.
*Significantly different (P<.05) from lost sprints.
a
Edition 2017, stage 2 no helicopter footage; therefore, position and number of teammates were not possible to determine.
IJSPP Vol. 16, No. 9, 2021
How to Win Sprints in the Tour de France 1375
Brought to you by IJSPP Board Membership | Authenticated teunvanerp@hotmail.com | Downloaded 03/02/23 11:54 AM UTC
Sprint Characteristics and Days in TdF
Similar to results attained in a case study describing a general
classification contender’s performances on key mountains,
15
this
study did not note any PO decline in the sprint and sprint
preparation throughout the TdF. This contrasts with results attained
by Rodriguez-Marroyo et al
18
who noted a significant decrement
(approximately 10%) in both maximal and submaximal endurance
performance during a laboratory exercise test after a GT. One of
the reasons for these conflicting results could be that riders are
somewhat mentally fatigued at the end of a GT and thus competing
for a stage victory, or taking part in a laboratory test, could yield
performance declines. However, in this study, only 2 of the
analyzed sprints took place in the third week and both in stage
21, which is also referred to as the champagne stage. In this stage,
cyclists only race the last hour, compared with a normal race of 4 to
6 hours. Thus, it may be that the rider was somewhat fresher when
the sprint preparation started. The limited number of sprints in
week 3, in combination with the lower intensities in stage 21, could
have blunted the effects of fatigue. Therefore, based on the
presented data, it is somewhat difficult to assess the influence of
fatigue.
Figure 2 —Maximal mean PO for 5 seconds (A), 10 seconds (B), 15 seconds (C), and average PO for the time frames of last 10 to 5 minutes (D), last 3
to 2 minutes (E), and last 30 to 15 seconds (F) of 21 sprints in relation with days in the TdF—together with the regression line. Numbers (13, 14, 16, and
17) indicate which edition of the TdF. No significant different slopes. PO indicates power ouput; TdF, Tour de France.
IJSPP Vol. 16, No. 9, 2021
1376 van Erp, Kittel, and Lamberts
Brought to you by IJSPP Board Membership | Authenticated teunvanerp@hotmail.com | Downloaded 03/02/23 11:54 AM UTC
Limitations
The cyclist used 2 different brands of power meters, which both
claim to have an accuracy of ±2%, although this is only confirmed
for the SRM power meter.
19
During his participation in the TdF,
the sprinter used multiple bikes fitted with different power meters.
Consequently, within one TdF PO data could have been yielded
by multiple power meters. Further helicopter video footage is
necessary to analyze position in the peloton as well as the number
of teammates during specific timeframes. It was therefore not
always possible to perform analyses at exactly the specific time
point required. Analyses were performed at 568 (57), 303 (12), 181
(14), 121 (8), 91 (3), 58 (5), 30 (2), and 15 (1) seconds for 10, 5, 3,
2, 1.5, 1 minute, 30, and 15 seconds, respectively. Furthermore, as
this was a case study, only a limited number of sprints were
analyzed, which could be considered a limitation. Based on the
video footage, 2 different sprint tactics were identified. However,
sprints are a hectic and dynamic process, it could be that teams
applied different sprint tactics for certain parts of the sprint. Hence,
one should be cautious about generalizing these results to a wider
population of sprints.
Practical Applications
This is the first study to describe performance characteristics in
combination with tactical decisions in the (preparation) sprint of
a world-class sprinter. Together with PART I, this case study
provides valuable insights into the intensity, load, and perfor-
mance demands of a world-class sprinter at the highest level of
racing, namely the TdF. Coaches and practitioners could use
these insights to improve their training process and identify
future world-class sprinters. Decision makers should be aware of
the advantages and disadvantages of different sprint tactics and
could use these results to help guide the selection of their GT
squats.
Conclusions
High explosive POs are necessary when competing as a sprinter at
the highest level. It seems as if the PO in the sprint (preparation) is
not influenced by the duration of the TdF. Different sprint tactics
adopted by different teams do influence the PO and the position in
the peloton during the sprint preparation. In addition, the position at
30 seconds before the sprint determines the outcome (ie, won vs
lost) of the sprint.
Acknowledgments
The authors would like to thank Albert Timmer for his help in collecting
video footage.
References
1. Lucia A, Hoyos J, Chicharro JL. Physiology of professional road
cycling. Sports Med. 2001;31(5):325–337. PubMed ID: 11347684
doi:10.2165/00007256-200131050-00004
2. Sanders D, Heijboer M. Physical demands and power profile of
different stage types within a cycling grand tour. Eur J Sport Sci.
2019;19(6):736–744. PubMed ID: 30589390 doi:10.1080/17461391.
2018.1554706
3. Padilla S, Mujika I, Orbananos J, Santisteban J, Angulo F, Jose
Goiriena J. Exercise intensity and load during mass-start stage races
in professional road cycling. Med Sci Sports Exerc. 2001;33(5):
796–802. PubMed ID: 11323551 doi:10.1097/00005768-200105000-
00019
4. www.procyclingstats.com. Published 2020. Accessed January 1,
2020.
5. Vogt S, Schumacher YO, Roecker K, et al. Power output during the
Tour de France. Int J Sports Med. 2007;28(9):756–761. PubMed ID:
17497569 doi:10.1055/s-2007-964982
6. Vogt S, Schumacher YO, Blum A, et al. Cycling power output
produced during flat and mountain stages in the Giro d’Italia: a case
study. J Sports Sci. 2007;25(12):1299–1305. PubMed ID: 17786683
doi:10.1080/02640410601001632
7. Peiffer JJ, Abbiss CR, Haakonssen EC, Menaspa P. Sprinting for the
win: distribution of power output in women’s professional cycling.
Int J Sports Physiol Perform. 2018;13(9):1237–1242. PubMed ID:
29688105 doi:10.1123/ijspp.2017-0757
8. Menaspa P, Quod M, Martin DT, Peiffer JJ, Abbiss CR. Physical
demands of sprinting in professional road cycling. Int J Sports Med.
2015;36(13):1058–1062. PubMed ID: 26252551
9. Menaspa P, Abbiss CR, Martin DT. Performance analysis of a
world-class sprinter during cycling grand tours. Int J Sports Physiol
Perform. 2013;8(3):336–340. PubMed ID: 23038704
10. Merkes PFJ, Menaspa P, Abbiss CR. Reducing aerodynamic drag by
adopting a novel road-cycling sprint position. Int J Sports Physiol
Perform. 2019;14(6):733–738. PubMed ID: 30427244 doi:10.1123/
ijspp.2018-0560
11. Merkes PFJ, Menaspa P, Abbiss CR. Power output, cadence, and
torque are similar between the forward standing and traditional sprint
cycling positions. Scand J Med Sci Sports. 2020;30(1):64–73.
PubMed ID: 31544261 doi:10.1111/sms.13555
12. Blocken B, Toparlar Y, van Druenen T, Andrianne T. Aerodynamic
drag in cycling team time trials. J Wind Eng Ind Aerodyn.
2018;182:128–145. doi:10.1016/j.jweia.2018.09.015
13. Hopkins WG, Marshall SW, Batterham AM, Hanin J. Progressive
statistics for studies in sports medicine and exercise science. Med Sci
Sports Exerc. 2009;41(1):3–13. PubMed ID: 19092709 doi:10.1249/
MSS.0b013e31818cb278
14. Sanders D, van Erp T, de Koning JJ. Intensity and load characteristics
of professional road cycling: differences between men’s and women’s
races. Int J Sports Physiol Perform. 2019;14(3):296–302. PubMed
ID: 30080422 doi:10.1123/ijspp.2018-0190
15. van Erp T, Hoozemans M, Foster C, de Koning J. Case report: load,
intensity, and performance characteristics in multiple grand tours.
Med Sci Sports Exerc. 2020;52(4):868–875. doi:10.1249/MSS.
0000000000002210
16. Gardner AS, Martin JC, Martin DT, Barras M, Jenkins DG. Maximal
torque- and power-pedaling rate relationships for elite sprint cyclists
in laboratory and field tests. Eur J Appl Physiol. 2007;101(3):
287–292. PubMed ID: 17562069 doi:10.1007/s00421-007-0498-4
17. Blocken B, van Druenen T, Toparlar Y, et al. Aerodynamic drag in
cycling pelotons: new insights by CFD simulation and wind tunnel
testing. J Wind Eng Ind Aerodyn. 2018;179:319–337. doi:10.1016/
j.jweia.2018.06.011
18. Rodriguez-Marroyo JA, Villa JG, Pernia R, Foster C. Decrement in
professional cyclists’performance after a grand tour. Int J Sports
Physiol Perform. 2017;12(10):1348–1355.
19. Gardner AS, Stephens S, Martin DT, Lawton E, Lee H, Jenkins D.
Accuracy of SRM and power tap power monitoring systems for
bicycling. Med Sci Sports Exerc. 2004;36(7):1252–1258. PubMed
ID: 15235334 doi:10.1249/01.MSS.0000132380.21785.03
IJSPP Vol. 16, No. 9, 2021
How to Win Sprints in the Tour de France 1377
Brought to you by IJSPP Board Membership | Authenticated teunvanerp@hotmail.com | Downloaded 03/02/23 11:54 AM UTC
