Content uploaded by Tim Kambič
Author content
All content in this area was uploaded by Tim Kambič on Oct 22, 2024
Content may be subject to copyright.
Four Weeks of Power Optimized Sprint Training Improves Sprint
Performance in Adolescent Soccer Players
Mikael Derakhti, Domen Bremec, Tim Kambicˇ, Lasse Ten Siethoff, and Niklas Psilander
Purpose: This study compared the effects of heavy resisted sprint training (RST) versus unresisted sprint training (UST) on sprint
performance among adolescent soccer players. Methods: Twenty-four male soccer players (age: 15.7 [0.5] y; body height: 175.7
[9.4] cm; body mass: 62.5 [9.2] kg) were randomly assigned to the RST group (n = 8), the UST group (n = 10), or the control
group (n = 6). The UST group performed 8 ×20 m unresisted sprints twice weekly for 4 weeks, whereas the RST group performed
5×20-m heavy resisted sprints with a resistance set to maximize the horizontal power output. The control group performed only
ordinary soccer training and match play. Magnitude-based decision and linear regression were used to analyze the data. Results:
The RST group improved sprint performances with moderate to large effect sizes (0.76–1.41) across all distances, both within and
between groups (>92% beneficial effect likelihood). Conversely, there were no clear improvements in the UST and control
groups. The RST evoked the largest improvements over short distances (6%–8%) and was strongly associated with increased
maximum horizontal force capacities (r= .9). Players with a preintervention deficit in force capacity appeared to benefit the most
from RST. Conclusions: Four weeks of heavy RST led to superior improvements in short-sprint performance compared with
UST among adolescent soccer players. Heavy RST, using a load individually selected to maximize horizontal power, is therefore
highly recommended as a method to improve sprint acceleration in youth athletes.
Keywords:team sport, resistance training, force–velocity profiling, youth athletes, 50%v
dec
The ability to accelerate over short distances is essential in
field-based team sports such as soccer.
1,2
Since 90% of sprints
performed during a soccer match are shorter than 20 m, maximal
sprinting speed is likely less important than acceleration.
3
Short-
sprint performance mirrors actual game situations and is an impor-
tant determinant of match-winning actions. For example, straight-
line sprinting is the most frequent action during goal situations in
professional soccer.
4,5
Top-level players perform numerous intense
actions every match and are significantly faster in the first 10 to
15 m than are low-level players.
6–8
Thus, the fastest players will be
approximately 1 m ahead of slower players after only 10 m of
sprinting, which could be a decisive advantage in a match.
8
Short-sprint performance is primarily determined by 2 abili-
ties: the generation of large ground reaction forces and the technical
ability to apply a proportion in the direction of the sprint
(ie, horizontally).
9,10
Researchers have investigated various meth-
ods of training these physical and technical qualities with the aim of
targeting the development of sprinting performance parameters.
Notably, resisted sprint training (RST) is a popular method of
providing resistance in a specific“horizontal”manner. In this
method, where the resistance usually is created by towing a
sled, the user can target the development of various sprint phases
by increasing or decreasing the load.
11
This loading represents a
continuum, with heavy loads roughly corresponding to horizontal
force at low speeds, early sprint phases, and short distances and
lighter loads corresponding to horizontal force at high speeds, late
acceleration (or perhaps maximum velocity), and long distances.
12
The RST with light loads (∼10% body mass [BM] or ∼10%
reduction of maximal speed [v
dec
]) has traditionally been studied
and recommended for improving sprint acceleration.
13,14
However,
RST using low resistance has been criticized because it only targets
one part of the loading continuum (high speed and late accelera-
tion) and often results in performance outcomes similar to non-
resisted sprinting, particularly in well-trained individuals.
15
This
likely occurs because RST with light loads does not acutely deviate
much from unresisted sprinting and therefore results in a training
outcome that is not substantially different. Accordingly, recent
studies show that heavier loads (>30% BM and >30% v
dec
) during
RST are necessary to improve short-distance sprint performance
among team-sport athletes.
16–18
RST based on %BM reduces the maximal velocity to
different degrees, depending on the athlete’s level of develop-
ment and the actual resistance determined via friction.
19
The
amount of velocity reduction, and not the absolute load, deter-
mines the training-induced stress and the type and magnitude of
the adaptations.
20,21
Individual–force velocity (F–v)profiling
and load–velocity (L–v)profiling can be used to identify whether
an athlete is proficient or deficient and to prescribe the suitable
load for a specific velocity reduction.
12,22
Values extracted from
the latter profile characterize the neuromuscular limits of the
system for force production, such as the maximal horizontal
force (F
0
), maximal running velocity (v
0
), and maximal hori-
zontal power (P
max
). Additionally, individual profiling also
enables analysis of the ratio of force produced in the horizontal
direction (RF%) and a theoretical maximal value of RF%
(RF
max
), which is a measure of the maximal mechanical
effectiveness of force application in the forward direction at
the sprint start.
23
Currently, individual F–vand L–vprofiling can
be more easily achieved using a robotic system because the
actual resistance is programmable and standardized across
environments.
24
Derakhti, Bremec, Siethoff, and Psilander are with The Swedish School of Sport and
Health Sciences, Stockholm, Sweden. Bremec is also with the SuperTrening Sport
Performance Centre, Celje, Slovenia. Kambičis with the Faculty of Sport, Univer-
sity of Ljubljana, Ljubljana, Slovenia; and the General Hospital Murska Sobota,
Murska Sobota, Slovenia. Psilander (niklas.psilander@gih.se) is corresponding
author.
1
International Journal of Sports Physiology and Performance, (Ahead of Print)
https://doi.org/10.1123/ijspp.2020-0959
© 2021 Human Kinetics, Inc. ORIGINAL INVESTIGATION
First Published Online: Oct. 27, 2021
Recent studies examining RST in elite athletes show a
relationship between pretraining F–vprofiles and how these
profiles are affected by training.
25,26
For example, athletes with
low initial F
0
values show larger improvements in F
0
and short-
sprint performance compared with F
0
-proficient athletes.
26
Thus,
individualized resistance based on F–vprofiling is recommended
in elite athletes.
27
However, F–vprofiling is time-consuming,
particularly in a team-sport setting where many athletes must be
tested and trained in quick succession. Moreover, adolescent
athletes often have imbalanced F–vprofiles, displaying a force
deficiency.
20,21
Therefore, a more generalized method that spe-
cifically targets this deficiency, such as power-optimized RST,
may be helpful for youth athletes. This method characterizes the
“optimal load”(L
opt
) as that which allows P
max
to be reached
during the maximum resisted velocity plateau (ie, 50%v
dec
)and
thus maintained for longer than a single instant during train-
ing.
11,19,24
Consequently, L
opt
represents the loading at which the
athletes can maximize the time spent in conditions close to
maximum horizontal power.
Although a growing number of studies show positive
effects of RST at or close to L
opt
on short-sprint perfor-
mance,
18,20,21,24,26,28
limited research among youth athletes is
available.
20,21
Furthermore, there are few short-duration studies
(<8 wk), with most examining the effects of RST during periods
of 8 to 12 weeks. The multifactorial physical, tactical, and
technical demands of team sports, together with the tight com-
petitive schedule, reduce the applicability of such long training
periods. Therefore, the present study aimed to compare the
effects of a 4-week heavy RST program using a robotic resis-
tance system applying power-optimized loading, with an un-
resisted sprint training (UST) program on sprint performance
among adolescent soccer players. We hypothesized that RST
would lead to the largest improvement in sprint performance and
that changes in F
0
would be associated with pretraining F
0
values and changes in sprint performance.
Methods
Subjects
Twenty-seven adolescent male soccer players volunteered to par-
ticipate in this study, with a mean (SD) age of 15.7 (0.5) years, a
mean height of 175.7 (9.4) cm, and a mean weight of 62.5 (9.2) kg.
All participants competed at the highest national level (per their age
group) in Sweden. They were familiar with strength training, but
not on a regular basis, and they had no previous RST experience
(except for the 3 familiarization sessions). The study inclusion
criteria were as follows: the absence of lower-limb injury and the
ability to perform maximal sprints and jumps. After baseline
testing, the participants were assigned to either the RST group
(n = 9), the UST group (n = 10), or the control (CON) group (n = 8).
There were no statistically significant differences in age or anthro-
pometrics between the groups at the baseline or after the training
intervention. All 3 groups followed the same soccer training
routine. At the end of the intervention, 3 participants (2 from
the CON group and 1 from the RST group) could not undergo
posttesting because of personal reasons. Hence, the final analysis
ultimately included test data from 24 participants. All subjects were
informed about the risks and benefits of the study via an institu-
tionally approved document, and they or their guardians signed
written consent forms. The Stockholm Regional Board of Ethics
approved this study (reference number DNR 2018/746-31/1).
Design
A 4-week randomized controlled trial was conducted to compare
the effects of 2 different sprint training programs among adolescent
soccer players: heavy RST versus UST. The effects of these 2
training programs were evaluated against the CON group. To
ensure as much similarity between experimental groups as possi-
ble, both groups were matched based on the participants’F–v
profiles obtained from the baseline testing. Specifically, the parti-
cipants with similar results were paired, followed by a random
division into the RST or UST group. In addition to the 4 weeks of
training, the participants underwent full familiarization, baseline
testing, and posttesting. Both experimental groups performed the
training twice a week, with each session consisting of 5 heavy
resisted sprints for the RST group and 8 unresisted sprints for the
UST group. The accumulated duration of the sprints was balanced
between the groups. The CON group performed standard soccer
training without any additional activities. The sample size was
based on previous studies in this field.
18,19
The primary outcome
variables were 5-, 10-, 20-, and 30-m sprint performances, and the
secondary outcomes were counter movement jump, standing long
jump, F
0
,P
max
,RF
max
, and maximal velocity during 30-m sprint-
ing (v
max
). The variables were measured at the baseline and after the
training intervention in a fully rested state. We used magnitude-
based decision with compatibility intervals and probabilities based
on the tdistribution to provide estimates of the uncertainty in the
mean effect size within and between groups.
29
Magnitude-based
inference has been criticized because of its increased risk of type I
errors.
30
However, this criticism has been addressed,
31
and given
the small sample size, the study was likely underpowered for null
hypothesis significance testing.
30
Equipment
A computerized sprint resistance system (1080 Motion AB, Li-
dingö, Sweden) was used for training, testing, and data collection.
The unit provided isotonic horizontal resistance in increment loads
of 1 kg to target the appropriate resistance. Instantaneous velocity
time data were then collected from the manufacturer software, at a
rate of 333 Hz. The manufacturer has previously reported the
repeatability (±0.7%) and accuracy of velocity (±0.5%) and force
(±4.8 N) for the 1080 Sprint system (www.1080 motion.com/
science). Baseline and posttraining measurements were performed
with the same protocol and equipment. Vertical jump trials were
performed using OptoJump hardware and software (MicroGate,
Bolzano, Italy).
General Testing Procedures
Testing was completed during the participants’late preseason and
early in-season period. The same researchers who supervised all
training performed the pretesting and posttesting. To minimize
possible learning effects, the participants underwent full familiari-
zation (ie, 3 familiarization sessions >48 h prior to baseline testing).
Testing was performed on 2 separate days. Day 1 began with 2
unloaded (1-kg resistance, as the practical minimum provided by
the machine) maximal effort 30-m sprints (T
30
m) to measure 5-,
10-, 20-, and 30-m split times and to create F–vprofiles. The raw
velocity–time data and Samozino’s inverse dynamics method were
used to compile the F–vrelationships.
23
Next, 4 progressively
loaded 20-m sprints were performed to compute the participants’
L–vprofile. The F–vand L–vprofiles were used to calculate and
measure the individual training load, F
0
,P
max
,RF
max
, and v
max
.
2Derakhti et al
(Ahead of Print)
Day 2 began with 3 CMJs followed by 3 SLJs. Prior to any testing,
the participants performed a standardized 20-minute warm-up
(SWU) consisting of jogging, dynamic stretching, technical drills,
and 4 submaximal 30-m stride outs. All sprint tests were conducted
outdoors on the same soccer pitch with an artificial “astro”turf
surface; the jumps were performed on a hard, flat asphalt surface.
All groups performed the tests on the same days during similar
weather conditions, and the participants wore the same clothes and
footwear during the pretesting and posttesting.
Sprint Testing and F–vand L–vProfiling
Post-SWU, the participants rested for 5 minutes before being
attached to the 1080 Sprint device via a hip belt. The participants
positioned themselves in a standing-split stance at the starting line,
after which they initiated the first T
30
m. They were instructed to
“lean”into their first step (removing the slack from the line), push
off with their front leg, and start sprinting when they felt like they
were about to fall. The selected starting leg remained constant
throughout all training and testing sessions. Although the sprint
data were gathered over a distance of 30 m, an actual sprinting
distance of 35 m was utilized to ensure that the participants did not
slow down before reaching 30 m. The second T
30
m was performed
after a 3-minute rest. The faster of the 2 trials was used to compute
split times and F–vprofiles.
After the T
30
-m testing, individual horizontal L–vprofiles were
assessed by utilizing a testing battery of resisted sprints based on
the procedures outlined by Cross et al.
16,19,24
The testing battery
consisted of four 20-m sprints performed with increasing loads.
The loads were adjusted to approximately equal-weighted sleds
loaded with 25%, 50%, 75%, and 100% of the participants’BM.
This loading span was selected to facilitate proper plotting of the
participants’L–vprofiles and to determine each athlete’sL
opt
.
Specifically, L
opt
represents the load that allows maximum power
(ie, is aligned with the apex of their individual horizontal power–
velocity relationship) to be developed at the maximum resisted
velocity plateau and, subsequently, a larger proportion of training
to be performed around P
max
by attaining and maintaining maxi-
mum resisted velocity.
19
Table 1displays an overview of the sprint
testing procedure.
Vertical and Horizontal Jump Testing
Three CMJs, interspersed by 1 minute of rest, were performed
5 minutes after an SWU, with the highest jump recorded as the test
result. The participants were instructed to place their hands slightly
above their hips and keep them there throughout the entire jump.
Moreover, they were instructed to go “fast down fast up,”land on
their toes, and not bend their knees duringthe flight phase or landing.
The standing long jump was performed 2 minutes after the
counter movement jump testing, using an extended measuring tape
placed on the ground that marked distances of 1, 2, and 3 m with red
tape for the participants to see. Three attempts were allowed, with
1 minute of rest between each jump. The participants were
instructed to stand erect with their feet parallel behind the given
zero line, to use their arms as a pendulum, and to jump as far as
possible, landing on both feet. The jump was measured as the
distance from the starting point to the heel of the foot that was
furthest back, with the furthest jump recorded as the test result.
Training Regimen
Training was performed prior to the regular soccer training and
consisted of linear maximal effort sprints. The RST group performed
an SWU prior to 5 ×20-m maximal-effort resisted sprints. The 1080
Sprint system was used during the resisted sprints, with the load
optimized such that the participant’s maximal velocity was reduced
by half (50%v
dec
). The training was performed in this manner for
Table 1 Sprint Testing Procedure
∼20 min of SWU followed by 5 min of passive rest prior to sprint testing procedure
Sprint nr Load T
30
mF–vprofiling L–vprofiling Rest period, min
1 Unloaded ✓✓ –3
2 Unloaded ✓✓ –3
3 25% BM –– ✓3
4 50% BM –– ✓3
5 75% BM –– ✓3
6 100% BM –– ✓3
Abbreviations: SWU, standardized warm-up; BM, body mass; F–vprofiling, force–velocity profiling; L–vprofiling, load–velocity profiling.
Table 2 Intervention Design
Session Repetitions ×distance, m Sprint type, load Total sprint duration, s Warm-up
RST
13×20 Maximal effort, individual L
opt
20–21 SWU
2–95×20 Maximal effort, individual L
opt
32–35 SWU
UST
1–98×20 Maximal effort, unloaded 32–35 SWU
Abbreviations: L
opt
, optimal load for the subjects to work at their P
max
; RST, resisted sprint training group; SWU, standardized warm-up; UST, unresisted sprint training
group.
Resisted Sprint Training in Youth Soccer 3
(Ahead of Print)
every session apart from the first training session, during which the
participants only performed 3 resisted sprints. The UST group
performed an SWU prior to 8 ×20-m maximal-effort unresisted
sprints. Each sprint was interspersed by a 3-minute rest, and the
training was conducted twice per week for 4 weeks on nonconsecutive
days. The remaining team-specific training included 4 soccer sessions
per week, ranging from 45 to 90 minutes. After the fifth intervention
session, the participants’regular game season began, and 1 competi-
tive soccer match per week was added to the weekly load. The CON
group performed only regular soccer-specific team training and
matches. The participants of all 3 groups were instructed not to
expose themselves to any other training stimuli during the interven-
tion. Table 2provides further details on the intervention design.
Statistical Analysis
All data were imported and processed in Microsoft Excel 2016
(Microsoft Corporation, Redmond, WA). Linear regression models
were analyzed in IBM SPSS statistics (version 26.0.01, Armonk,
NY), and Figures 1and 2were processed in GraphPad Prism
version 8.34 for Windows (GraphPad Software, La Jolla, CA). The
data in the text and figures are presented as the mean (SD) or ±90%
compatibility intervals. The practical relevance of the outcome
variables was assessed using magnitude-based decisions.
29
The
effects of the training (RST, UST, or CON) differences over time
(pre to post) and the differences between groups were calculated.
The smallest worthwhile change was set to 0.2 ×SD. Since the
sample sizes were small (n = 8, 10, and 6), we used a tdistribution
to calculate the 90% compatibility intervals and the chances of
beneficial, harmful, or trivial changes. Qualitative statements were
assessed as follows: 25% to 75%, possibly; 75% to 95%, likely;
95% to 99.5%, very likely; >99.5%, most likely. If the chances of
having beneficial/higher or harmful/lower performances were both
>5%, the true difference was considered unclear. The effect sizes
were qualitatively described as trivial, small, moderate, large, very
large, and extremely large for standardized thresholds of <0.2,
≥0.2, ≥0.6, ≥1.2, ≥2, and ≥4 and for regression coefficients of <0.1,
≥0.1, ≥0.3, ≥0.5, ≥0.7, and ≥0.9, respectively.
29
To provide
information about individual responses, we presented the number
of individuals who displayed changes that were better, worse, or
within the smallest worthwhile change.
To test our hypothesis that pretraining F
0
values were associ-
ated with improvements in F
0
, we performed a linear regression
with group and initial F
0
values as the independent variables and
changes in F
0
as the dependent variables. The same method was
used to test whether the association between changes in F
0
(independent) were associated with changes in 20-m sprint perfor-
mance (dependent) and training group (independent). Changes in
20-m sprint time (T
20
) were chosen as the primary outcome
variable for sprint performance, since 20-m sprinting was the basis
of the training interventions.
Results
The RST group likely improved sprinting performance across all
time points and comparisons. The improvement increased progres-
sively with decreasing distance from ∼4% at 30 m to ∼8% at 5 m.
Conversely, there were no clear changes in sprinting performance
for the UST or CON groups at any time point or for any comparison
(Table 3). Thus, there was a likely to very likely beneficial effect for
the RST group compared with the UST and CON groups for all
sprinting distances. No clear between-group differences were
observed when comparing the UST and CON groups (Figure 1).
The counter movement jump height possibly increased for the
UST group (∼6%), but not for the other 2 groups. The standing long
jump length likely increased for both the RST and UST groups
(∼6%–7%), but not for the CON group (Table 4). Thus, there was a
very likely to most likely better effect in jumping performance for
the 2 intervention groups compared with the CON group
(Figure 2).
Figure 1 —Pairwise comparison of the changes in sprint performance
between (A) the RST and UST groups, (B) the RST and CON groups, and
(C) the UST and CON groups. T
5
–T
30
denotes the standardized change in
sprint performance at 5, 10, 20, and 30 m, respectively. Error bars represent
the 90% CIs of the mean change in performance. The gray area marks the
limits of a trivial change. Numbers separated by a slash are the probability
(in percentage) of an improvement that is better for the left group/within
the trivial range/better for the right group. CI indicates compatibility
interval; CON, control group only performing ordinary soccer training;
RST, resisted sprint training group; UST, unresisted sprint training group.
4Derakhti et al
(Ahead of Print)
The performance-related variables F
0
,P
max
, and RF
max
likely
improved in the RST group (approximately 9%–18%). Conversely,
there were no improvements in the UST group, but a possible
decrease for F
0
(∼4%) (Table 4). This resulted in a likely better
effect for the RST group compared with the other 2 groups for F
0
,
P
max
, and RF
max
(Figure 2). The v
max
was unaffected in the RST
and UST groups, whereas the CON group displayed a very likely
decrease (∼3%) (Table 4).
Change in F
0
was a very large predictor for changes in 20-m
sprinting time (P<.001, r=−.84), and the preintervention F
0
was a
large predictor for improvements in F
0
(P<.001, r=−.70)
(Table 5). The intervention group was a significant predictor for
changes in F
0
(P<.006, r=−.48), but not for changes in sprint
performance (P<.079, r= .18). The overall models explained 80%
of the improvement in sprint performance and 48% of the change
in F
0
.
Discussion
This is the first study to compare the effects of a short, power-
optimized, heavy RST program with a UST program among youth
athletes. Our primary finding was that 4 weeks of heavy RST
improved sprint performance among late pubertal adolescent soc-
cer players, while the UST group displayed no improvement. The
improved sprint performance was primarily due to increases in
maximal horizontal force production and improvements in early
sprint acceleration, which was in line with our hypothesis. The
CON group, which only received regular soccer training, showed
no performance improvements in any of the measured outcome
variables.
The RST group displayed similar or greater improvement in
sprint performance than that reported in previous studies,
17,18,24,32
despite a shorter training period (4 vs 8–12 wk), supporting the
effectiveness of the training protocol. This can be explained by
several factors. First, we used a resistance that reduces velocity by
50% to maximally stimulate the ability to produce horizontal
power. Additionally, the resistance was applied by a robotic system
via a hip harness, which is ideal for providing the right amount of
horizontal resistance while maintaining an optimal sprinting posi-
tion. Together, this may be a highly efficient way for stimulating
power adaptations. Second, the participants in the present study
were younger than those in most previous studies. Adolescent
athletes may be more sensitive to training-induced adaptations that
improve sprint performance.
33
In line with this, the participants had
relatively low starting values of F
0
, which have been shown to
correlate with improvements in horizontal force production and
sprint performance.
26
Furthermore, the individual changes in F
0
,
together with the assigned training group of the participants (RST,
UST, or CON), explained 80% of the improvement in the 20-m
sprint performance. Therefore, the comparatively large effect of
this short-training intervention is likely a consequence of a good
match between an effective training method and population.
The short-training intervention utilized in the present study
was primarily chosen to mimic real-world team-sport periodiza-
tion. Short periods of specific training, called block mesocycles,
have been frequently used among team-sport athletes, with
studies recommending that these periods last between 2 and 4
weeks.
34
The present intervention spanned both the precompeti-
tive and competitive seasons (2 wk in each); therefore, our results
indicate that heavy RST is a suitable training form during this
important transition period, particularly because the CON group
experienced a performance decrease in some of the measured
variables (P
max
and v
max
). This decrease may have occurred
because the CON group was less prepared for intense match
play and became more fatigued compared with the other 2 groups.
Therefore, the CON group might not have been fully recovered
when performing the posttests. Multiple postintervention assess-
ment points would have enabled these factors to be more clearly
Figure 2 —Pairwise comparison of the changes in secondary outcome
variables between (A) the RST and UST groups, (B) the RST and CON
groups, and (C) the UST and CON groups. Error bars represent the 90%
CIs of the mean change in performance. The gray area marks the limits of a
trivial change. Numbers separated by a slash are the probability (in
percentage) of an improvement that is better for the left group/within
the trivial range/better for the right group. CI indicates compatibility
interval; CMJ, counter movement jump; CON, control group only
performing ordinary soccer training; F
0
, maximal horizontal force;
P
max
, maximal horizontal power; RF
max
, maximum ratio of force
produced in the forward direction at sprint start; RST, resisted sprint
training group; SLJ, standing long jump; UST, unresisted sprint training
group; v
max
, maximal velocity during 30-m sprinting.
Resisted Sprint Training in Youth Soccer 5
(Ahead of Print)
elucidated,
28
but the competitive schedules of the athletes did not
permit such a design.
Previous studies have demonstrated the importance of devel-
oping maximal horizontal force production to improve the early
acceleration phase of sprinting.
35,36
Since team-sport athletes
primarily perform short sprints, there is a compelling argument
to target the development of this ability during training. Interest-
ingly, the present study indicated that the improvement in sprint
performance in the RST group increased gradually from the 30-m
to the 5-m sprint (T
30
=3.8%, T
20
=4.2%, T
10
=5.6%, and
T
5
= 7.9%). This finding, combined with the large increases in
F
0
,P
max
,andRF
max
, but not in v
max
,confirms that RST mainly
improved sprint acceleration, which agrees with results from both
adult and adolescent populations.
17,18,20,21,26,32,35
The fact that
these changes were observed after only 4 weeks of training
indicates that the changes were primarily driven by neural and
technical improvements. This, in combination with our finding
that horizontal jump length, but not vertical jump height,
increased, demonstrates that the athletes appear to have dispro-
portionately developed technical capacities, rather than gross
physical ones.
Importantly, we did not detect a decrease in v
max
in the RST
group. This agrees with recent findings showing a decrease in v
max
when applying very heavy loads (75%v
dec
), but not for loads close
to L
opt
(50%v
dec
).
21
A possible explanation is that, in theory, RST at
L
opt
would improve both maximum velocity and maximum force
since it targets the development of the middle portion of the F–v
relationship. However, in light of this, and recent studies, the most
efficient training to improve maximal velocity seems to be sprint-
ing at velocities close to or above v
max
(ie, assisted sprint train-
ing).
25
Nevertheless, our results indicate that power-optimized RST
does not longitudinally impede the v
max
of young soccer players
following a short intervention.
The present study has some limitations. First, we used
magnitude-based decision to provide usable results from this
otherwise small data set. Therefore, the results should be inter-
preted with the relatively underpowered nature of this data set in
mind. Second, although sprint performance increased by ∼4% to
8%, which is an unusually large increase for athletes, one should
be careful when translating this improvement directly to sport-
specific performance. The improved acceleration and horizontal
forces could be partly due to improved body orientation during
early sprinting phases.
26
Although this improves sprint accelera-
tion, it might not be practically relevant for a soccer player who
requires an upright posture throughout the game. Additionally,
many accelerations in soccer are performed from a flying start and
not from a dead start (eg, from jogging to sprinting). Future
studies should therefore include sprint testing with a flying start
and examine the effect of training on the starting angle. Finally,
completely unloaded sprinting is not possible when using the
1080 system (see “Methods”section for details). This might have
affected v
max
, and the results should, therefore, be interpreted with
this in mind.
Practical Applications
1. A loading prescription, applied to stimulate the development
of maximal horizontal power, appears to be more beneficial to
short-sprint performance than unresisted sprinting. This
method could be integrated into the training of youth athletes
to enhance sport-specific performance.
2. L–vprofiling and loading of the desired resistance are greatly
simplified by a robotic system. However, this can also be done
by more cost-efficient devices such as timing gates, smart
phone applications, and a weight-adjustable sled.
23
Table 3 Descriptive Statistics (Mean [SD]) and Changes (With 90% CI) in Sprint Performance in the RST, UST,
and CON Groups
Group
Pre,
mean (SD)
Post,
mean (SD) % Change (90% CI) ES (90% CI)
Chances
(B/T/H)
Ind. Resp.
(B/T/H) Outcome
T
30
,s
RST 5.34 (0.24) 5.15 (0.20) −3.67 (−6.43 to −0.92) −0.89 (−1.56 to −0.22) 95/4/1 6/1/1 Very likely beneficial
UST 5.42 (0.47) 5.45 (0.38) 0.54 (−3.98 to 5.06) 0.07 (−0.51 to + 0.65) 21/45/35 3/4/3 Unclear
CON 5.39 (0.10) 5.48 (0.20) 1.70 (−0.56 to 3.97) 0.62 (−0.20 to 1.44) 5/13/82 1/1/4 Unclear
T
20
,s
RST 3.96 (0.22) 3.79 (0.17) −4.20 (−7.46 to −0.94) −0.86 (−1.53 to −0.19) 95/4/1 6/1/1 Likely beneficial
UST 4.0 (0.36) 4.06 (0.28) 1.52 (−3.15 to 6.19) 0.19 (−0.39 to 0.77) 12/39/49 4/1/5 Unclear
CON 4.06 (0.1) 4.11 (0.18) 1.27 (−1.60 to 4.14) 0.36 (−0.46 to 1.19) 11/24/65 1/3/2 Unclear
T
10
,s
RST 2.50 (0.2) 2.36 (0.15) −5.69 (−10.41 to −0.97) −0.81 (−1.48 to −0.14) 94/5/1 6/1/1 Likely beneficial
UST 2.55 (0.30) 2.61 (0.21) 2.23 (−3.52 to 7.99) 0.23 (−0.35 to 0.80) 11/36/53 3/2/5 Unclear
CON 2.65 (0.11) 2.68 (0.17) 1.01 (−3.30 to 5.31) 0.19 (−0.63 to 1.02) 19/32/49 2/2/2 Unclear
T
5
,s
RST 1.67 (0.20) 1.54 (0.15) −7.87 (−14.85 to −0.89) −0.76 (−1.43 to −0.09) 92/6/2 7/0/1 Likely beneficial
UST 1.71 (0.26) 1.77 (0.18) 3.45 (−3.93 to 10.82) 0.27 (−0.31 to 0.85) 9/33/59 3/3/4 Unclear
CON 1.84 (0.11) 1.84 (0.16) 0.45 (−5.49 to 6.40) 0.06 (−0.76 to 0.89) 27/35/38 2/2/2 Unclear
Abbreviations: B, beneficial; CI, compatibility interval; CON, control group only performing ordinary soccer training; Ind. Resp., individual response; H, harmful;
RST, resisted sprint training group; T, trivial; T
30
, time to sprint 30 m; T
20
, time to sprint 20 m; T
10
,timetosprint10m;T
5
, time to sprint 5m; UST, unresisted sprint training
group.
6Derakhti et al
(Ahead of Print)
3. Only 4 weeks of power-optimized RST, performed 2 times per
week, was sufficient to improve performance. This greatly
improves the applicability of the described method, that is, it
can be used as a block mesocycle during the late preseason or
in-season periods.
Conclusions
We showed that 4 weeks of power-optimized RST was more
beneficial than UST at improving short-sprint performance in ado-
lescent soccer players. The improvement in sprint times increased
gradually with decreasing distance. Additionally, maximal horizontal
power, maximal horizontal force application, and maximal effective-
ness of force application improved, indicating that the training
primarily affected sprint acceleration performance. Maximal velocity
remained unchanged. Finally, the effect was more pronounced in
athletes with low horizontal force capabilities at baseline. Overall,
these results show that power-optimized RST is a very efficient and
easy method for improving sprint performance in youth athletes.
Acknowledgments
The authors would like to thank the participants and coaches from Älvsjö
AIK for their cooperation and maximal effort during the training interven-
tion. M.D. and D.B. contributed equally to this work.
Table 5 Results From Linear Regression Models
With 2 Predictors (N = 24)
Predictors rr
2
Adjusted r
2
P
Dependent variable = changes
in 20-m sprint performance
Training group .180 .032 —.079
Changes in F
0
−.836 .699 —<.001
Overall model .903 .815 .798 <.001
Dependent variable = changes
in F
0
Training group −.478 .228 —.006
Initial F
0
−.698 .487 —<.001
Overall model .722 .522 .476 <.001
Abbreviations: F
0
, maximal horizontal force; r, correlation coefficient.
Table 4 Descriptive Statistics (Mean [SD]) and Changes (With 90% CI) in Physical Performance in the RST, UST,
and CON Groups
Group
Pre,
mean (SD)
Post,
mean (SD)
% Change
(90% CI) ES (90% CI)
Chances
(B/T/H)
Ind. Resp.
(B/T/H) Outcome
CMJ, cm
RST 31.1 (3.9) 32.6 (4.9) 4.66 (−4.73 to 14.05) 0.33 (−0.34 to 1.00) 64/27/9 3/5/0 Unclear
UST 29.2 (4.8) 31.01 (4.9) 6.34 (−3.33 to 16.02) 0.38 (−0.20 to 0.96) 71/24/5 8/1/1 Possibly beneficial
CON 32.0 (4.5) 31.35 (3.6) −1.98 (−12.48 to 8.52) −0.16 (−0.98 to 0.67) 21/33/46 1/4/1 Unclear
SLJ, cm
RST 211.3 (17.9) 226.4 (20.0) 7.16 (4.16 to 13.16) 0.80 (0.01 to 1.47) 93/5/1 7/1/0 Likely beneficial
UST 206.9 (14.2) 219.10 (17.7) 5.90 (1.43 to 10.37) 0.76 (0.01 to 1.34) 95/5/1 8/2/0 Likely beneficial
CON 226.3 (11.5) 224.0 (13.8) −0.99 (−5.59 to 3.60) −0.18 (−1.00 to 0.64) 20/32/48 0/4/2 Unclear
P
max
, W/kg
RST 11.7 (2.4) 13.4 (2.5) 14.38 (0.19 to 28.56) 0.68 (0.01 to 1.35) 89/9/2 6/1/1 Likely beneficial
UST 11.6 (4.0) 10.84 (2.1) −8.62 (−23.66 to 6.42) −0.33 (−0.91 to 0.17) 6/28/66 3/3/4 Unclear
CON 10.8 (0.5) 10.29 (0.8) −4.34 (−9.08 to −0.40) −0.75 (−1.58 to −0.07) 3/8/88 1/0/5 Likely harmful
RF
max
,%
RST 36.8 (5.1) 40 (3.8) 8.84 (0.78 to 16.90) 0.74 (0.07 to 1.40) 91/7/2 7/0/1 Likely beneficial
UST 35.6 (6.7) 34.2 (4.2) −3.93 (−12.87 to 5.00) −0.26 (−0.83 to 0.32) 9/34/57 3/3/4 Unclear
CON 32.67 (2.58) 32.50 (3.6) −0.51 (−8.32 to 7.30) −0.26 (−1.08 to 0.56) 28/35/37 3/1/2 Unclear
v
max
, m/s
RST 7.54 (0.42) 7.56 (0.51) 0.23 (−3.89 to 4.35) 0.04 (−0.63 to 0.71) 33/41/26 4/1/3 Unclear
UST 7.62 (0.38) 7.70 (0.61) 1.05 (−2.72 to 4.82) 0.16 (−0.42 to 0.74) 45/40/14 6/0/4 Unclear
CON 8.15 (0.24) 7.91 (0.18) −2.99 (−4.68 to −1.29) −1.45 (−2.27 to −0.63) 1/1/99 0/1/5 Very likely harmful
F
0
, N/kg
RST 5.50 (1.66) 6.50 (1.70) 18.17 (0.65 to 35.68) 0.70 (0.03 to 1.37) 90/8/2 7/0/1 Likely beneficial
UST 5.33 (2.17) 4.66 (1.04) −12.46 (−30.93 to 6.00) −0.39 (−0.97 to 0.19) 5/23/72 3/3/4 Possibly harmful
CON 4.13 (0.59) 4.18 (0.80) 1.33 (−8.68 to 11.34) 0.11 (−0.71 to 0.93) 42/34/24 4/0/2 Unclear
Abbreviations: B, beneficial; CI, compatibility interval; CMJ, counter movement jump; CON, control group only performing ordinary soccer training; F
0
, maximal
horizontal force; H, harmful; P
max
, maximal horizontal power; RF
max
, maximum ratio of force produced in the forward direction at sprint start; RST, resisted sprint training
group; SLJ, standing long jump; T, trivial; UST, unresisted sprint training group; v
max
, maximal velocity during 30-m sprinting.
Resisted Sprint Training in Youth Soccer 7
(Ahead of Print)
References
1. Schimpchen J, Skorski S, Nopp S, Meyer T. Are “classical”tests of
repeated-sprint ability in football externally valid? A new approach to
determine in-game sprinting behaviour in elite football players. J
Sports Sci. 2016;34(6):519–526. PubMed ID: 26580089 doi:10.
1080/02640414.2015.1112023
2. Spencer M, Bishop D, Dawson B, Goodman C. Physiological and
metabolic responses of repeated-sprint activities: specifictofield-
based team sports. Sports Med. 2005;35(12):1025–1044. PubMed ID:
16336007 doi:10.2165/00007256-200535120-00003
3. Vigne G, Gaudino C, Rogowski I, Alloatti G, Hautier C. Activity
profile in elite Italian soccer team. Int J Sports Med. 2010;31(5):304–
310. PubMed ID: 20301042 doi:10.1055/s-0030-1248320
4. Alcaraz PE, Carlos-Vivas J, Oponjuru BO, Martinez-Rodriguez A.
The effectiveness of resisted sled training (RST) for sprint perfor-
mance: a systematic review and meta-analysis. Sports Med. 2018;
48(9):2143–2165. PubMed ID: 29926369 doi:10.1007/s40279-018-
0947-8
5. Cometti G, Maffiuletti NA, Pousson M, Chatard JC, Maffulli N.
Isokinetic strength and anaerobic power of elite, subelite and amateur
French soccer players. Int J Sports Med. 2001;22(1):45–51. PubMed
ID: 11258641 doi:10.1055/s-2001-11331
6. Bangsbo J, Mohr M, Krustrup P. Physical and metabolic demands of
training and match-play in the elite football player. J Sports Sci.
2006;24(7):665–674. PubMed ID: 16766496 doi:10.1080/02640
410500482529
7. Haugen T, Tonnessen E, Hisdal J, Seiler S. The role and development
of sprinting speed in soccer. Int J Sports Physiol Perform. 2014;
9(3):432–441. PubMed ID: 23982902 doi:10.1123/ijspp.2013-0121
8. Stolen T, Chamari K, Castagna C, Wisloff U. Physiology of soccer:
an update. Sports Med. 2005;35(6):501–536. PubMed ID: 15974635
doi:10.2165/00007256-200535060-00004
9. Morin JB, Bourdin M, Edouard P, Peyrot N, Samozino P, Lacour JR.
Mechanical determinants of 100-m sprint running performance. Eur J
Appl Physiol. 2012;112(11):3921–3930. PubMed ID: 22422028
doi:10.1007/s00421-012-2379-8
10. Colyer SL, Nagahara R, Takai Y, Salo AIT. How sprinters accelerate
beyond the velocity plateau of soccer players: waveform analysis of
ground reaction forces. Scand J Med Sci Sports. 2018;28(12):2527–
2535. PubMed ID: 30230037 doi:10.1111/sms.13302
11. Cross MR, Tinwala F, Lenetsky S, et al. Assessing horizontal force
production in resisted sprinting: computation and practical interpre-
tation. Int J Sports Physiol Perform. 2019;14(5):689–693. PubMed
ID: 30975007 doi:10.1123/ijspp.2018-0578
12. Cross MR, Samozino P, Brown SR, Morin JB. A comparison between
the force-velocity relationships of unloaded and sled-resisted sprint-
ing: single vs. multiple trial methods. Eur J Appl Physiol.
2018;118(3):563–571. PubMed ID: 29302753 doi:10.1007/s00421-
017-3796-5
13. Alcaraz PE, Palao JM, Elvira JL. Determining the optimal load for
resisted sprint training with sled towing. J Strength Cond Res.
2009;23(2):480–485. PubMed ID: 19197200 doi:10.1519/JSC.
0b013e318198f92c
14. Lockie RG, Murphy AJ, Spinks CD. Effects of resisted sled towing on
sprint kinematics in field-sport athletes. J Strength Cond Res.
2003;17(4):760–767. PubMed ID: 14636109
15. Petrakos G, Morin JB, Egan B. Resisted sled sprint training to
improve sprint performance: a systematic review. Sports Med.
2016;46(3):381–400. PubMed ID: 26553497 doi:10.1007/s40279-
015-0422-8
16. Cross MR, Brughelli M, Samozino P, Morin JB. Methods of power-
force-velocity profiling during sprint running: a narrative review.
Sports Med. 2017;47(7):1255–1269. PubMed ID: 27896682 doi:10.
1007/s40279-016-0653-3
17. Kawamori N, Newton RU, Hori N, Nosaka K. Effects of weighted
sled towing with heavy versus light load on sprint acceleration ability.
J Strength Cond Res. 2014;28(10):2738–2745. PubMed ID:
23539079 doi:10.1519/JSC.0b013e3182915ed4
18. Morin JB, Petrakos G, Jimenez-Reyes P, Brown SR, Samozino P,
Cross MR. Very-heavy sled training for improving horizontal-force
output in soccer players. Int J Sports Physiol Perform. 2017;
12(6):840–844. doi:10.1123/ijspp.2016-0444
19. Cross MR, Brughelli M, Samozino P, Brown SR, Morin JB. Optimal
loading for maximizing power during sled-resisted sprinting. Int J
Sports Physiol Perform. 2017;12(8):1069–1077. PubMed ID:
28051333 doi:10.1123/ijspp.2016-0362
20. Cahill MJ, Oliver JL, Cronin JB, Clark KP, Cross MR, Lloyd RS.
Influence of resisted sled-push training on the sprint force-velocity
profile of male high school athletes. Scand J Med Sci Sports.
2020;30(3):442–449. PubMed ID: 31742795 doi:10.1111/sms.13600
21. Cahill MJ, Oliver JL, Cronin JB, et al. Influence of resisted sled-pull
training on the sprint force-velocity profile of male high-school
athletes. J Strength Cond Res. 2020;34(10):2751–2759. PubMed
ID: 32773545 doi:10.1519/JSC.0000000000003770
22. Morin JB, Samozino P. Interpreting power-force-velocity profiles for
individualized and specific training. Int J Sports Physiol Perform.
2016;11(2):267–272. PubMed ID: 26694658 doi:10.1123/ijspp.
2015-0638
23. Samozino P, Rabita G, Dorel S, et al. A simple method for measuring
power, force, velocity properties, and mechanical effectiveness in
sprint running. Scand J Med Sci Sports. 2016;26(6):648–658.
PubMed ID: 25996964 doi:10.1111/sms.12490
24. Cross MR, Lahti J, Brown SR, et al. Training at maximal power in
resisted sprinting: optimal load determination methodology and pilot
results in team sport athletes. PLoS One. 2018;13(4):e0195477.
PubMed ID: 29641589 doi:10.1371/journal.pone.0195477
25. Lahti J, Jimenez-Reyes P, Cross MR, et al. Individual sprint force-
velocity profile adaptations to in-season assisted and resisted veloc-
ity-based training in professional rugby. Sports. 2020;8(5):74. doi:10.
3390/sports8050074
26. Lahti J, Huuhka T, Romero V, Bezodis I, Morin JB, Hakkinen K.
Changes in sprint performance and sagittal plane kinematics after
heavy resisted sprint training in professional soccer players. Peer J.
2020;8:e10507. PubMed ID: 33362970 doi:10.7717/peerj.10507
27. Simpson A, Waldron M, Cushion E, Tallent J. Optimised force-
velocity training during pre-season enhances physical performance in
professional rugby league players. J Sports Sci. 2021;39(1):91–100.
PubMed ID: 32799729 doi:10.1080/02640414.2020.1805850
28. Morin JB, Capelo-Ramirez F, Rodriguez-Perez MA, Cross MR,
Jimenez-Reyes P. Individual adaptation kinetics following heavy
resisted sprint training. J Strength Cond Res. 2020. PubMed ID
32058358 doi:10.1519/jsc.0000000000003546
29. 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):3–12. PubMed ID: 19092709 doi:10.1249/
MSS.0b013e31818cb278
30. Sainani KL. The problem with “magnitude-based inference.”Med Sci
Sports Exerc. 2018;50(10):2166–2176. PubMed ID: 29683920
doi:10.1249/MSS.0000000000001645
31. Batterham AM, Hopkins WG. The problems with “the problem with
‘magnitude-based inference.’” Med Sci Sports Exerc. 2019;51(3):599.
PubMed ID: 30365421 doi:10.1249/MSS.0000000000001823
8Derakhti et al
(Ahead of Print)
32. Cahill MJ, Oliver JL, Cronin JB, Clark KP, Cross MR, Lloyd RS.
Sled-pull load-velocity profiling and implications for sprint training
prescription in young male athletes. Sports. 2019;7(5):119. doi:10.
3390/sports7050119
33. Meyers RW, Oliver JL, Hughes MG, Lloyd RS, Cronin JB. The
influence of maturation on sprint performance in boys over a 21-
month period. Med Sci Sports Exerc. 2016;48(12):2555–2562.
PubMed ID: 27434083 doi:10.1249/MSS.0000000000001049
34. Issurin VB. Benefits and limitations of block periodized training
approaches to athletes’preparation: a review. Sports Med. 2016;
46(3):329–338. PubMed ID: 26573916 doi:10.1007/s40279-015-
0425-5
35. Buchheit M, Samozino P, Glynn JA, et al. Mechanical determinants
of acceleration and maximal sprinting speed in highly trained young
soccer players. J Sports Sci. 2014;32(20):1906–1913. PubMed ID:
25356503 doi:10.1080/02640414.2014.965191
36. Rabita G, Dorel S, Slawinski J, et al. Sprint mechanics in world-class
athletes: a new insight into the limits of human locomotion. Scand J
Med Sci Sports. 2015;25(5):583–594. PubMed ID: 25640466 doi:10.
1111/sms.12389
Resisted Sprint Training in Youth Soccer 9
(Ahead of Print)