RESEARCH ARTICLE

Position specific player load during match-

play in a professional football club
Ivan Baptista1*, Dag Johansen2, André Seabra3, Svein Arne Pettersen1
1 School of Sport Sciences, University of Tromsø, the Arctic University of Norway, Tromsø, Norway,
2 Computer Science Department, University of Tromsø, the Arctic University of Norway, Tromsø, Norway,
3 Research Centre in Physical Activity, Health and Leisure (CIAFEL), Faculty of Sport, University of Porto,
Porto, Portugal

* [email protected]

Abstract
a1111111111 There is a rapid growing body of knowledge regarding physical aspects of a football match
a1111111111
due to studies using computer-assisted motion analysis. The present study used time-
a1111111111
a1111111111 motion analysis and triaxial-accelerometers to obtain new insights about differences in
a1111111111 physical profiles of elite football players across playing-positions. Player performance data
in 23 official home matches from a professional football club, during two seasons were col-
lected for analysis. Eighteen players from five different playing positions (central backs: n =
3; full-backs: n = 5; central midfielders: n = 6; wide midfielders: n = 3; and central forwards:
OPEN ACCESS n = 4), performing a total of 138 observations. A novel finding was that central backs and
central midfielders had significantly lower work-rate in sprints, decelerations and accelera-
Citation: Baptista I, Johansen D, Seabra A,
Pettersen SA (2018) Position specific player load tions than full-backs, wide midfielders and central forwards (p<0.001). Furthermore, wide
during match-play in a professional football club. midfielders and full-backs performed significantly more turns (>90˚) than central backs. The
PLoS ONE 13(5): e0198115. https://doi.org/ most common distance covered in high-intensity runs (19.8 km
h−1) for central backs, cen-
10.1371/journal.pone.0198115
tral midfielders, wide midfielders and central forwards was 1–5 m, but for full-backs was
Editor: Riccardo Di Giminiani, University of 6–10 m. This may help coaches in developing individualized training programs to meet the
L’Aquila, ITALY
demands of each position in match-play.
Received: October 4, 2017

Accepted: May 14, 2018

Published: May 24, 2018

Copyright: © 2018 Baptista et al. This is an open

access article distributed under the terms of the
Introduction
Creative Commons Attribution License, which
permits unrestricted use, distribution, and To understand physical demands of match-play in football objective data is essential andsuch
reproduction in any medium, provided the original data could be important for practitioners in designing training programs [1]. Of particular
author and source are credited.
importance is the potential value objective data provide for personalized prescription of train-
Data Availability Statement: All relevant data are ing load in a cohort of players following the same overall training regime.
within the paper and its Supporting Information Time motion analysis is commonly used in elite football to analyse player and team perfor-
files.
mance in training and match as it allows quantification of player running activities and indi-
Funding: The authors received no specific funding rect verification of the energetics of match-play [2], creating a rapid growing body of
for this work. knowledge regarding the physical aspects of football training and match-play [3].
Competing interests: The authors have declared Football has a high-intensity intermittent nature [4], characterised by prolonged intermit-
that no competing interests exist. tent exercise interspersed by periods of maximal or close to maximal effort [5]. Players may be

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 1 / 10

 Load in professional football

required to repeat sprints, accelerations and turns of short duration interspersed by brief
recovery periods over an extended period of time, and these activities have been reported as
crucial factors for team performance [6–9].
Previous research has focused on the influence of different factors in the players’ match
running profiles, such as the tactical systems [10], possession status [11, 12], competitive stan-
dard [13], seasonal fluctuations [14], environment [15], opponent [16] and playing positions
[17, 18].
Based on robust findings within the research literature, it is evident that specific playing
positions have an influence on total match-load. Midfielders appear to cover the greatest over-
all distances (~11.5 km) while defenders and forwards cover lower distance (10–10.5 km) [4,
19–21]. Regarding high-intensity runs (HIR), the literature shows that, typically, wide mid-
fielders (WM) and full-backs (FB) display superior HIR profiles [20, 22, 23] and central backs
(CB) perform a significantly less amount of time sprinting and running with high intensity
compared with other positions [1, 17].
The use of only distance and speed may underestimate the calculation of external player
workload since this type of time-motion analysis has neglected some essential and specific
movements of football (turns, accelerations, decelerations, etc.) that together appear numerous
times during every match and may cause significant physical stress on the players [19, 24].
A previous study, with a Norwegian elite football team [24], combined data from triaxial
accelerometer and time-motion analysis and experienced that player load was accumulated in
a variety of ways across the different playing positions with accelerations and decelerations
contributing 7–10% and 5–7%, respectively. Previous research has shown that players in lateral
positions (FB and WM) accelerate more often, whereas CB and central midfielders (CM)
decelerate less compared to other positions [24–26].
Therefore, the aims of the present study were to establish and compare the physical
demands during official match-play in five different playing positions (CB, FB, CM, WM and
central forwards [CF]) in a Norwegian elite football team using time-motion and triaxial-
accelerometers.

Methods
Subjects and match analysis
With approval from UiT The Arctic University of Norway Institutional Review Board, written
informed consent from players and approval from Norwegian Centre for Research Data, data
on performance in 23 official home matches from the first team (highest level) in a Norwegian
elite football club, during two seasons (2016 and 2017), were collected for analysis. The
matches were all played on artificial grass surface (Alfheim Stadium, Tromsø, length = 110m;
width = 68m). The sample included 18 players (25.2 ± 4.4 years; 76.2 ± 6.4 kg; 181.6 ± 5.6 cm;
in age, body mass and height, respectively) across five different playing positions: CB (n = 3,
observations[obs] = 35), FB (n = 5, obs = 34), CM (n = 6, obs = 38), WM (n = 3, obs = 18) and
CF (n = 4, obs = 13), making a total of 138 observations. These positions were chosen accord-
ing to team’s main tactic formation and previous research [8, 18, 20, 24, 26, 27].
Data was analysed only if: (1) players completed the entire match, (2) the player played in
the same position during all the match and (3) the team used 4-5-1 or 4-3-3 tactic formations.
To ensure players confidentiality, all data was anonymized before analyses.

Procedures
A stationary radio-based tracking system (ZXY Sport Tracking System, Trondheim Norway)
was used to characterize match activity profiles in the team. Each player wore a specially

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 2 / 10

 Load in professional football

designed belt, wrapped tightly around the waist, with an electronic sensor system at the play-
er’s lumbar spine [28]. The accuracy and reliability of the system in measuring player move-
ments in elite soccer competitions have been described in more detail in previous studies [26,
28, 29].

Physical performance variables

Physical parameters analysed included: number of accelerations (acccounts), acceleration dis-
tance per minute—work-rate—(accwr), number of decelerations (deccounts), deceleration
work-rate (decwr), HIR work-rate (HIRwr), HIR distance (HIRdist), sprint work-rate (sprintwr),
sprint distance (sprintdist) and turns.
The following locomotor categories were selected: HIR (19.8 km
h−1) and sprinting
(25.2 km
h−1). The speed thresholds applied for each locomotor categories are similar to
those reported in previous research [16, 20, 24, 26].
According to the ZXY Sport Tracking system, accelerations are defined by four event mark-
ers: (1) the start of the acceleration event is marked by the acceleration reaching the minimum
limit of 1 m
s −2, (2) the acceleration reaches the acceleration limit of 2 m
s −2, (3) the accelera-
tion remains above the 2 m
s −2 for at least 0.5 seconds and (4) the duration of the acceleration
ends when it decreases below the minimum acceleration limit (1 m
s −2).
A turn was defined as a continuous and significant rotation of the body in one direction
(derived from gyroscope and compass data). When a rotation in the opposite direction is mea-
sured, that will be the end of the previous turn and the start of the next turn. Due to the angle
threshold used by ZXY Sport Tracking system only turns 90 degrees were analysed.

Statistical analysis
Descriptive statistics (means and standard deviations) were calculated for the total sample and
playing position.
Differences in match performance measures by field position were tested with a one-way
analysis of variance (ANOVA). When significance was found, a Bonferroni post-hoc test was
performed.
Effect sizes (ES), using Cohen‘s d, was calculated and interpreted as trivial (0.2), small
(>0.2–0.6), moderate (>0.6–1.2) and large (>1.2). Significance level was set at 0.05 [30]. Sta-
tistical analyses were conducted using SPSS version 24.0.

Results
Acceleration and deceleration profiles
There were similar patterns in accwr and decwr with CB and CM performing less than FB, WM
and CF, with the most significant difference being between CB (3.5 ± 0.7) and CF (5.3 ± 1.0) in
decwr (p<0.001).
In relation to acccounts and deccounts WM presented higher values (76.7 ± 12.1; 86.1 ± 14.7)
than CB (64.9 ± 9.7; 61.5 ± 10.8) and CM (65.8 ± 15.6; 71.5 ± 20.6) (p<0.001), respectively.
Furthermore, all positions, except CB, performed less acccounts than deccounts during the
entire match (Table 1).

HIR and sprint profiles

Differences were observed in HIRWR and Sprintwr between CB and the other positions. CB
had the lowest values of all positions in both variables but especially pronounced in Sprintwr
(0.9 ± 0.5 m/min) when compared with CF (2.5 ± 1.0 m/min) (p<0.001).

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 3 / 10

 Load in professional football

Table 1. Descriptive statistic (mean and standard deviation) and ANOVA analysis (p-value) of different acceleration parameters analysed according to field posi-
tion and respective Effect Size (ES) of differences observed.
Variables Central Full- Central Wide Central p-value Post-hoc multiple comparisons
Backs backs Midfielders Midfielders Forwards (p<0.05) | Effect Size
AccWR (m/ 3.7 (0.7) 4.4 (0.6) 3.7 (1.2) 4.8 (0.9) 5.1 (1.3) <0.001 CB<FB (0.25); CB<WM (0.33); CB<CF (0.39); FB>CM
min) (0.26); CM<WM (0.34); CM<CF (0.40)
AccCOUNTS 64.9 (9.7) 71.2 65.8 (15.6) 76.7 (12.1) 71.7 (12.0) 0.008 CB<WM (0.28); CM<WM (0.26)
(total) (11.6)
DecWR (m/ 3.5 (0.7) 4.6 (0.7) 4.1 (1.4) 5.2 (0.9) 5.3 (1.0) <0.001 CB<FB (0.39); CB<WM (0.50); CB<CF (0.48); CM<WM
min) (0.31); CM<CF (0.31)
DecCOUNTS 61.5 (10.8) 73.7 71.5 (20.6) 86.1 (14.7) 80.3 (14.6) <0.001 CB<FB (0.29); CB<WM (0.47); CB<CF (0.33); CM<WM
(total) (14.0) (0.28)
https://doi.org/10.1371/journal.pone.0198115.t001

Regarding HIRdist, CF presented higher values in 26–30 m than all the other positions,
while distances of 36–40 and 46–50 m were covered more times by FB (1.7 ± 1.4; 0.9 ± 1.0). CB
(0.8 ± 0.9; 0.2 ± 0.6) were the players with lowest values in these longer distances (36–40 and
46–50). Furthermore, distances of 1–5 m were the distances covered more often by CB, CM,
WM and CF, whereas FB had higher values in distances of 6–10 m (Table 2).
In relation to sprintdist CB, FB, CM and WM performed higher number of 1–5 m, while CF
covered higher number of 6–10 m sprints. (Table 3).
Furthermore, there was a pattern of covariance in the work-rates analysed (acc, dec, HIR
and sprint) across playing positions (Fig 1).

Table 2. Descriptive statistics statistic (mean and standard deviation) and ANOVA analysis (p-value) of different HIR distances and work-rate parameters analysed
according to field position and respective Effect Size (ES) of differences observed.
Variables Central Full- Central Wide Central p-value Post-hoc multiple comparisons
Backs backs Midfielders Midfielders Forwards (p<0.05) | Effect Size
HIRWR (m/min) 5.2 (1.6) 8.1 (1.7) 8.0 (3.5) 9.2 (1.8) 9.4 (1.6) <0.001 CB<FB (0.46); CB<CM (0.46); CB<WM (0.54);
CB<CF (0.51)
HIRDIST 1–5 m 8.2 (2.7) 7.5 (2.5) 9.2 (3.1) 10.3 (2.6) 9.3 (4.2) 0.009 FB<WM (0.27)
(counts)
HIRDIST 6–10 m 7.6 (2.2) 8.3 (3.0) 8.2 (3.1) 8.9 (2.4) 8.2 (1.9) 0.591 No sig. differences
(counts)
HIRDIST 11–15 m 5.0 (2.7) 6.6 (3.0) 6.3 (3.0) 8.1 (3.0) 6.4 (1.4) 0.008 CB<WM (0.33)
(counts)
HIRDIST 16–20 m 4.8 (2.1) 5.0 (2.1) 5.2 (2.6) 5.8 (1.7) 6.0 (2.2) 0.301 No sig. differences
(counts)
HIRDIST 21–25 m 2.6 (1.5) 3.7 (1.5) 3.7 (2.1) 4.2 (1.9) 5.2 (1.5) <0.001 CB<WM (0.28); CB<CF (0.40)
(counts)
HIRDIST 26–30 m 1.7 (1.1) 2.7 (1.4) 2.7 (1.8) 2.3 (1.0) 4.3 (1.2) <0.001 CB<FB (0.26); CB<CF (0.50); FB<CF (0.31); CM<CF
(counts) (0.33); WM<CF (0.35)
HIRDIST 31–35 m 1.1 (0.8) 1.7 (1.2) 2.2 (1.6) 3.4 (1.9) 2.8 (2.1) <0.001 CB<CM (0.24); CB<WM (0.41); CB<CF (0.26);
(counts) FB<WM (0.30)
HIRDIST 36–40 m 0.8 (0.9) 1.7 (1.4) 1.2 (1.1) 2.0 (0.8) 1.5 (1.1) 0.001 CB<FB (0.31); CB<WM (0.33)
(counts)
HIRDIST 41–45 m 0.6 (0.9) 1.0 (1.1) 1.4 (1.3) 1.1 (1.0) 1.5 (1.0) 0.009 CB<CM (0.29)
(counts)
HIRDIST 46–50 m 0.2 (0.6) 0.9 (1.0) 0.8 (0.9) 0.8 (1.1) 1.2 (1.4) 0.007 CB<FB (0.23); CB<CF (0.26)
(counts)
https://doi.org/10.1371/journal.pone.0198115.t002

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 4 / 10

 Load in professional football

Table 3. Descriptive statistics statistic (mean and standard deviation) and ANOVA analysis (p-value) of different sprint distances and work-rate parameters ana-
lysed according field position and respective Effect Size (ES) of differences observed.
Variables Central Full- Central Wide Central p-value Post-hoc multiple comparisons
Backs backs Midfielders Midfielders Forwards (p<0.05) | Effect Size
SprintWR (m/min) 0.9 (0.5) 2.0 (0.6) 1.4 (1.0) 1.7 (0.7) 2.5 (1.0) <0.001 CB<FB (0.49); CB<CM (0.26); CB<WM (0.32); CB<CF
(0.55); FB>CM (0.24); CM<CF (0.37)
SprintDIST 1–5 m 1.9 (1.2) 3.0 (1.7) 2.3 (1.9) 4.2 (1.7) 3.4 (1.6) <0.001 CB<WM (0.42); CM<WM (0.35)
(counts)
SprintDIST 6–10 m 1.9 (1.5) 2.9 (1.2) 2.2 (1.8) 3.1 (1.6) 3.6 (2.5) 0.007 CB<CF (0.23)
(counts)
SprintDIST 11–15 m 1.0 (1.0) 2.2 (1.5) 1.5 (1.5) 1.8 (1.1) 2.3 (1.7) 0.008 CB<FB (0.29)
(counts)
SprintDIST 16–20 m 0.7 (0.7) 1.6 (1.2) 1.3 (1.5) 1.3 (0.9) 2.6 (1.6) <0.001 CB<FB (0.25); CB<CF (0.40); CM<CF (0.28); WM<CF
(counts) (0.26)
SprintDIST 21–25 m 0.4 (0.6) 1.1 (0.9) 0.9 (1.1) 1.0 (0.9) 1.5 (1.1) 0.001 CB<FB (0.29); CB<CF (0.33)
(counts)
SprintDIST 26–30 m 0.3 (0.5) 0.7 (0.7) 0.5 (0.6) 0.5 (0.6) 0.4 (0.5) 0.087 No sig. differences
(counts)
SprintDIST 31–35 m 0.1 (0.3) 0.5 (0.6) 0.3 (0.6) 0.4 (0.7) 1.1 (0.8) <0.001 CB<CF (0.42); FB<CF (0.26); CM<CF (0.34); WM<CF
(counts) (0.25)
SprintDIST 36–40 m 0.0 (1.7) 0.3 (0.5) 0.2 (0.4) 0.0 (0.0) 0.5 (0.7) 0.001 CB<FB (0.24); CB<CF (0.27); WM<CF (0.26)
(counts)
SprintDIST >41 m 0.3 (0.5) 0.4 (0.8) 0.2 (0.5) 0.4 (0.6) 0.2 (0.4) 0.436 No sig. differences
(counts)
https://doi.org/10.1371/journal.pone.0198115.t003

Turns
The main outcome was that CB performed less turns per match (32.7 ± 10.1) than FB
(41.0 ± 12.1) and WM (42.9 ± 12.3) (p = 0.009).
Moreover, turn angles, 90˚-180˚ were the angles performed more often by all positions,
while the turns with the highest angles (271˚-360˚) were the least common (Table 4).

Discussion
The present study shows that the physical demands in official match-play, in elite football, vary
greatly across playing positions. As previously mentioned, a novel finding from this study was
that the work-rates in HIR, sprints, accelerations and decelerations change in the same pattern
across playing positions. Although further research is needed to verify the correlation between
these variables, our results demonstrate that CB and CM had significantly lower work-rate in
sprints, accelerations and decelerations than FB, WM and CF with CB also having lower
HIRwr than these three playing positions (p<0.001). These findings are in line with the
research literature regarding FB covering greater high-intensity and sprinting distances during
matches compared to CB. [13, 18, 20, 31].
Previous studies have reported greater distances in HIR and sprint covered by wide players
(FB and WM) compared with more central positions (CB, CM and CF) [13, 20, 24, 31], how-
ever the present study shows significant higher work-rate for wide positions only in acc, dec
and sprints but not in HIR, even though the values for wide positions are slightly, though insig-
nificantly, higher than for central positions (excluding CF). No significant differences were
observed between CF and WM in HIRwr which is in line with previous research [18], but in
opposition to others [11, 20, 31]. Furthermore, our data show that CF is the most physical
demanding position with longer distances covered in HIR, sprints, accelerating and

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 5 / 10

 Load in professional football

Fig 1. Work-rate profiles across playing position. Mean work-rate in sprints, HIR, acc and dec.
https://doi.org/10.1371/journal.pone.0198115.g001

decelerating than the other positions. It has been speculated within the research literature that
these differences between wide and more central positions are due to a lack of space for reach-
ing sprinting velocity and the playing style (different roles for different positions) [24, 25, 32].
Taking into consideration the specific context of the club where our data was collected, it
seems evident that the style of play (playing many times with low defence and in counter-
attacking) had a crucial influence on position’s specific physical demands.
Table 2 illustrates that player position had a significant influence on the different distances
covered in HIR. To the best of our knowledge, no previous research has characterized players’
HIR profiles regarding specific distances covered per HIR in official match-play across differ-
ent playing positions. Our data show that while the most common distance covered in HIR for
CB, CM, WM and CF was 1–5 m, for FB it was 6–10 m. An aspect to consider is that we also
observed some HIR longer than the ones presented in Table 2 but with no significant differ-
ences between positions.

Table 4. Descriptive statistics statistic (mean and standard deviation) and ANOVA analysis (p-value) of different parameters of turns analysed according to field
position and respective Effect Size (ES) of differences observed.
Variables Central Backs Full-backs Central Midfielders Wide Midfielders Central Forwards p-value Post-hoc multiple comparisons
(p<0.05) | Effect Size
Turns 32.7 (10.1) 41.0 (12.1) 37.0 (12.4) 42.9 (12.3) 41.6 (12.9) 0.009 CB<FB (0.25); CB<WM (0.25)
Turns (90˚-180˚) 20.3 (6.3) 21.8 (7.2) 20.2 (7.4) 24.2 (6.9) 20.9 (5.7) 0.277 No sig. differences
Turns (181˚-270˚) 9.8 (5.3) 16.4 (6.1) 13.7 (5.0) 14.9 (6.4) 15.9 (7.8) <0.001 CB<FB; CB<WM; CB<CF
Turns (271˚-360˚) 2.3 (1.9) 2.8 (2.1) 3.2 (2.1) 3.6 (1.9) 5.0 (1.9) 0.001 CB<CF; FB<CF
https://doi.org/10.1371/journal.pone.0198115.t004

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 6 / 10

 Load in professional football

Different patterns appear in sprintdist with CB, FB, CM and WM covering more often
shorter distances (1–5 m) in sprint while CF had higher values in longer distances (6–10 m).
Another important finding is that CF and WM accelerated more often compared with play-
ers in the other positions, which differs from a previous study with another Norwegian profes-
sional football club [24]. However, some similar trends were observed between these studies,
with CB being the players who decelerated the least times compared with other playing posi-
tions. Furthermore, when comparing our data with results from previous research [4, 24, 25]
we observed slightly lower values of acccounts in almost all the positions (CB, FB, CM and
WM). The inverse trend was observed in deccounts with all positions presenting higher values
in our study, probably due to style of play.
A main finding of the present study refers to the number of turns observed across playing
positions. In fact, even though our study has taken into consideration only turns >90˚ (angle
threshold defined by ZXY Sport Tracking), total different values were obtained compared with
previous research [17]. One difference is related to the total number of turns per match with
our study presenting a mean of ~42 ± 13 to attackers (CF), ~39 ± 13 to midfielders (CM and
WM) and ~37 ± 12 to defenders (CB and FB), while previous research [17] presented mean
values significantly higher for each position: attackers (~101), midfielders (~107) and defend-
ers (~97) in turns >90˚. They observed that midfielders performed significantly fewer turns
during a match than defenders and strikers. Our data show that CM did not perform signifi-
cantly different compared to the other positions while WM performed more turns than CB.
These differences may be caused by the different sampling technology used.
Both turns, acceleration and deceleration activities add substantial load in addition to high-
intensity running and must be taken into consideration when analysing physical demands of
match-play.
It should be noted that different measurement technologies could cause the discrepancy in
results between the present study and previous research [5]. Also, different playing styles, cul-
tural and competitive contexts may account for differences observed.
In summary, our data show that speed and distance measures only to some extent predict
the physical demands of a football player and that these demands vary greatly across playing
positions. Taking into consideration the law of training specificity [33] and the idea that the
physical loading of the training session should be individually designed to improve perfor-
mance and avoid excess of fatigue and overtraining [34] the coaches need a clear view how dif-
ferent playing positions achieve load.

Practical application
The present results may provide useful and novel insight regarding positional differences in
physical profiles of elite football players during match-play. The positional differences in work-
load and work pattern need to be taken into consideration when designing and implementing
training program cycles, according to the team’s style of play. As for the team explored in the
present study, lateral players should perform some longer sprints  30 m in normal training
weeks to be prepared for these actions that appear during match. Performing sprints in addi-
tion to small sided games must be taken into consideration when planning the trainings since
small and medium sided games do not provide enough space to elicit these actions.
Apart from providing valuable information to coaches about the activity profiles of differ-
ent positions, the results may also provide the foundation for a real-time personalization com-
puterized coach toolkit based on our whole-field video analysis system [35] that integrates
with positional data in real-time. We are currently developing such a mobile system to custom-
ize individual training load to player positions while the practice is unfolding.

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 7 / 10

 Load in professional football

Supporting information
S1 File. Data review.
(SAV)

Author Contributions
Conceptualization: Ivan Baptista.
Data curation: André Seabra.
Formal analysis: André Seabra.
Investigation: Ivan Baptista.
Methodology: Ivan Baptista.
Project administration: Svein Arne Pettersen.
Supervision: Dag Johansen, Svein Arne Pettersen.
Writing – original draft: Ivan Baptista.
Writing – review & editing: Dag Johansen, Svein Arne Pettersen.

References
1. Ade J, Fitzpatrick J, Bradley PS. High-intensity efforts in elite soccer matches and associated move-
ment patterns, technical skills and tactical actions. Information for position-specific training drills. J
Sports Sci. 2016; 34(24):2205–14. Epub 2016/08/19. https://doi.org/10.1080/02640414.2016.1217343
PMID: 27537056.
2. Carling C. Interpreting physical performance in professional soccer match-play: should we be more
pragmatic in our approach? Sports medicine (Auckland, NZ). 2013; 43(8):655–63. Epub 2013/05/11.
https://doi.org/10.1007/s40279-013-0055-8 PMID: 23661303.
3. Paul DJ, Bradley PS, Nassis GP. Factors affecting match running performance of elite soccer players:
shedding some light on the complexity. International journal of sports physiology and performance.
2015; 10(4):516–9. Epub 2015/05/01. https://doi.org/10.1123/IJSPP.2015-0029 PMID: 25928752.
4. Bradley PS, Di Mascio M, Peart D, Olsen P, Sheldon B. High-intensity activity profiles of elite soccer
players at different performance levels. J Strength Cond Res. 2010; 24(9):2343–51. Epub 2009/11/18.
https://doi.org/10.1519/JSC.0b013e3181aeb1b3 PMID: 19918194.
5. Randers MB, Mujika I, Hewitt A, Santisteban J, Bischoff R, Solano R, et al. Application of four different
football match analysis systems: a comparative study. J Sports Sci. 2010; 28(2):171–82. Epub 2010/
04/15. https://doi.org/10.1080/02640410903428525 PMID: 20391091.
6. Bishop D, Spencer M, Duffield R, Lawrence S. The validity of a repeated sprint ability test. J Sci Med
Sport. 2001; 4(1):19–29. Epub 2001/05/08. PMID: 11339490.
7. Carling C, Dupont G. Are declines in physical performance associated with a reduction in skill-related
performance during professional soccer match-play? J Sports Sci. 2011; 29(1):63–71. Epub 2010/11/
16. https://doi.org/10.1080/02640414.2010.521945 PMID: 21077004.
8. Carling C, Le Gall F, Dupont G. Analysis of repeated high-intensity running performance in professional
soccer. J Sports Sci. 2012; 30(4):325–36. Epub 2012/01/18. https://doi.org/10.1080/02640414.2011.
652655 PMID: 22248291.
9. Girard O, Mendez-Villanueva A, Bishop D. Repeated-sprint ability—part I: factors contributing to
fatigue. Sports medicine (Auckland, NZ). 2011; 41(8):673–94. Epub 2011/07/26. https://doi.org/10.
2165/11590550-000000000-00000 PMID: 21780851.
10. Bradley P, Carling C, Archer D, Roberts J, Dodds A, Di Mascio M, et al. The effect of playing formation
on high-intensity running and technical profiles in English FA Premier League soccer matches. J Sports
Sci. 2011; 29(8):821–30. Epub 2011/04/23. https://doi.org/10.1080/02640414.2011.561868 PMID:
21512949.
11. Bradley P, Lago-Penas C, Rey E, Gomez Diaz A. The effect of high and low percentage ball possession
on physical and technical profiles in English FA Premier League soccer matches. J Sports Sci. 2013; 31
(12):1261–70. Epub 2013/05/24. https://doi.org/10.1080/02640414.2013.786185 PMID: 23697463.

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 8 / 10

 Load in professional football

12. Di Salvo V, Pigozzi F, Gonzalez-Haro C, Laughlin MS, De Witt JK. Match performance comparison in
top English soccer leagues. Int J Sports Med. 2013; 34(6):526–32. Epub 2012/11/28. https://doi.org/10.
1055/s-0032-1327660 PMID: 23184481.
13. Bradley P, Carling C, Gomez Diaz A, Hood P, Barnes C, Ade J, et al. Match performance and physical
capacity of players in the top three competitive standards of English professional soccer. Human move-
ment science. 2013; 32(4):808–21. Epub 2013/08/28. https://doi.org/10.1016/j.humov.2013.06.002
PMID: 23978417.
14. Mohr M, Krustrup P, Bangsbo J. Match performance of high-standard soccer players with special refer-
ence to development of fatigue. J Sports Sci. 2003; 21(7):519–28. Epub 2003/07/10. https://doi.org/10.
1080/0264041031000071182 PMID: 12848386.
15. Mohr M, Nybo L, Grantham J, Racinais S. Physiological Responses and Physical Performance during
Football in the Heat. PLOS ONE. 2012; 7(6):e39202. https://doi.org/10.1371/journal.pone.0039202
PMID: 22723963
16. Rampinini E, Coutts AJ, Castagna C, Sassi R, Impellizzeri FM. Variation in top level soccer match per-
formance. Int J Sports Med. 2007; 28(12):1018–24. Epub 2007/05/15. https://doi.org/10.1055/s-2007-
965158 PMID: 17497575.
17. Bloomfield J, Polman R, O’Donoghue P. Physical Demands of Different Positions in FA Premier League
Soccer. Journal of sports science & medicine. 2007; 6(1):63–70. Epub 2007/01/01. PMID: 24149226;
PubMed Central PMCID: PMCPMC3778701.
18. Schuth G, Carr G, Barnes C, Carling C, Bradley PS. Positional interchanges influence the physical and
technical match performance variables of elite soccer players. J Sports Sci. 2016; 34(6):501–8. Epub
2015/12/25. https://doi.org/10.1080/02640414.2015.1127402 PMID: 26700131.
19. Bangsbo J. The physiology of soccer—with special reference to intense intermittent exercise. Acta phy-
siologica Scandinavica Supplementum. 1994; 619:1–155. Epub 1994/01/01. PMID: 8059610.
20. Bradley P, Sheldon W, Wooster B, Olsen P, Boanas P, Krustrup P. High-intensity running in English FA
Premier League soccer matches. J Sports Sci. 2009; 27(2):159–68. Epub 2009/01/21. https://doi.org/
10.1080/02640410802512775 PMID: 19153866.
21. Rienzi E, Drust B, Reilly T, Carter JE, Martin A. Investigation of anthropometric and work-rate profiles of
elite South American international soccer players. The Journal of sports medicine and physical fitness.
2000; 40(2):162–9. Epub 2000/10/18. PMID: 11034438.
22. Bush M, Barnes C, Archer DT, Hogg B, Bradley PS. Evolution of match performance parameters for
various playing positions in the English Premier League. Human movement science. 2015; 39:1–11.
Epub 2014/12/03. https://doi.org/10.1016/j.humov.2014.10.003 PMID: 25461429.
23. Rampinini E, Impellizzeri FM, Castagna C, Abt G, Chamari K, Sassi A, et al. Factors influencing physio-
logical responses to small-sided soccer games. J Sports Sci. 2007; 25(6):659–66. Epub 2007/04/25.
https://doi.org/10.1080/02640410600811858 PMID: 17454533.
24. Dalen T, Ingebrigtsen J, Ettema G, Hjelde GH, Wisloff U. Player Load, Acceleration, and Deceleration
During Forty-Five Competitive Matches of Elite Soccer. J Strength Cond Res. 2016; 30(2):351–9. Epub
2015/06/10. https://doi.org/10.1519/JSC.0000000000001063 PMID: 26057190.
25. Aughey RJ, Varley MC. Acceleration profiles in elite Australian soccer. Int J Sports Med. 2013; 34
(3):282. Epub 2013/02/20. https://doi.org/10.1055/s-0032-1331776 PMID: 23420194.
26. Ingebrigtsen J, Dalen T, Hjelde GH, Drust B, Wisloff U. Acceleration and sprint profiles of a professional
elite football team in match play. Eur J Sport Sci. 2015; 15(2):101–10. Epub 2014/07/10. https://doi.org/
10.1080/17461391.2014.933879 PMID: 25005777.
27. 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):519–26. Epub 2015/11/19. https://doi.org/10.1080/02640414.2015.1112023
PMID: 26580089.
28. Bendiksen M, Pettersen SA, Ingebrigtsen J, Randers MB, Brito J, Mohr M, et al. Application of the
Copenhagen Soccer Test in high-level women players—locomotor activities, physiological response
and sprint performance. Human movement science. 2013; 32(6):1430–42. Epub 2013/09/11. https://
doi.org/10.1016/j.humov.2013.07.011 PMID: 24016711.
29. Andersen TB, Krustrup P, Bendiksen M, Orntoft CO, Randers MB, Pettersen SA. Kicking Velocity and
Effect on Match Performance When using a Smaller, Lighter Ball in Women’s Football. Int J Sports
Med. 2016; 37(12):966–72. Epub 2016/11/05. https://doi.org/10.1055/s-0042-109542 PMID:
27551936.
30. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale NJ: Lawrence Ear-
lbaum Associate; 1988.

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 9 / 10

 Load in professional football

31. Di Mascio M, Bradley PS. Evaluation of the most intense high-intensity running period in English FA pre-
mier league soccer matches. J Strength Cond Res. 2013; 27(4):909–15. Epub 2012/06/02. https://doi.
org/10.1519/JSC.0b013e31825ff099 PMID: 22652921.
32. Di Salvo V, Baron R, Gonzalez-Haro C, Gormasz C, Pigozzi F, Bachl N. Sprinting analysis of elite soc-
cer players during European Champions League and UEFA Cup matches. J Sports Sci. 2010; 28
(14):1489–94. Epub 2010/11/05. https://doi.org/10.1080/02640414.2010.521166 PMID: 21049314.
33. Small K, McNaughton L, Greig M, Lovell R. Effect of timing of eccentric hamstring strengthening exer-
cises during soccer training: implications for muscle fatigability. J Strength Cond Res. 2009; 23
(4):1077–83. Epub 2009/06/17. https://doi.org/10.1519/JSC.0b013e318194df5c PMID: 19528859.
34. Rebelo AN, Silva P, Rago V, Barreira D, Krustrup P. Differences in strength and speed demands
between 4v4 and 8v8 small-sided football games. J Sports Sci. 2016; 34(24):2246–54. Epub 2016/06/
10. https://doi.org/10.1080/02640414.2016.1194527 PMID: 27278256.
35. Stensland HK, Gaddam VR, Tennøe M, Helgedagsrud E, Næss M, Alstad HK, et al. Bagadus: An inte-
grated real-time system for soccer analytics. ACM Trans Multimedia Comput Commun Appl. 2014; 10
(1s):1–21. https://doi.org/10.1145/2541011

PLOS ONE | https://doi.org/10.1371/journal.pone.0198115 May 24, 2018 10 / 10