Match performance of high-standard soccer players with
special reference to development of fatigue
MAGNI MOHR, PETER KRUSTRUP and JENS BANGSBO*
Institute of Exercise and Sport Sciences, August Krogh Institute, Department of Human Physiology, University of
Copenhagen, Universitetsparken 13, DK-2100 Copenhagen Ø, Denmark
Accepted 18 Decembe r 2002
The aim of this study was to assess physical fitness, match performance and development of fatigue during
competitive matches at two high standards of professional soccer. Computerized time–motion analyses were
performed 2–7 times during the competitive season on 18 top-class and 24 moderate professional soccer players.
In addition, the players performed the Yo-Yo intermittent recovery test. The top-class players performed 28 and
58% more (P 5 0.05) high-intensity running and sprinting, respectively, than the moderate players (2.43 + 0.14
vs 1.90 + 0.12 km and 0.65 + 0.06 vs 0.41 + 0.03 km, respectively). The top-class players were better (11%;
P 5 0.05) on the Yo-Yo intermittent recovery test than the moderate players (2.26 + 0.08 vs 2.04 + 0.06 km,
respectively). The amount of high-intensity running, independent of competitive standard and playing position,
was lower (35–45%; P 5 0.05) in the last than in the first 15 min of the game. After the 5-min period during
which the amount of high-intensity running peaked, performance was reduced (P 5 0.05) by 12% in the
following 5 min compared with the game average. Substitute players (n = 13) covered 25% more (P 5 0.05)
ground during the final 15 min of high-intensity running than the other players. The coefficient of variation in
high-intensity running was 9.2% between successive matches, whereas it was 24.8% between different stages of
the season. Total distance covered and the distance covered in high-intensity running were higher (P 5 0.05) for
midfield players, full-backs and attackers than for defenders. Attackers and full-backs covered a greater
(P 5 0.05) distance in sprinting than midfield players and defenders. The midfield players and full-backs
covered a greater (P 5 0.05) distance than attackers and defenders in the Yo-Yo intermittent recovery test
(2.23 + 0.10 and 2.21 + 0.04 vs 1.99 + 0.11 and 1.91 + 0.12 km, respectively). The results show that: (1) top-
class soccer players performed more high-intensity running during a game and were better at the Yo-Yo test
than moderate professional players; (2) fatigue occurred towards the end of matches as well as temporarily
during the game, independently of competitive standard and of team position; (3) defenders covered a shorter
distance in high-intensity running than players in other playing positions; (4) defenders and attackers had a
poorer Yo-Yo intermittent recovery test performance than midfielders and full-backs; and (5) large seasonal
changes were observed in physical performance during matches.
Keywords: high-intensity intermittent exercise, playing positions, seasonal changes, standard of competition,
time–motion analysis, Yo-Yo intermittent recovery test.
Introduction
Time–motion analysis is useful for examining the
activity pattern and physical aspects of socce r (Reilly
and Thomas, 1976; Mayhew and Wenger, 1985;
Bangsbo et al ., 1991; Bangsbo, 1994). Several authors
have examined low-standard and amateur players
(Saltin, 1973; Whitehead, 1975; Van Gool et al.,
1988; Ohashi et al., 1988); others have obs erved players
at the top of their national league (Bangsbo et al., 1991;
Bangsbo, 1994). These studies have suggested that the
amount of high-intensity exercise is a valid measure of
physical performance in soccer. It has also been
observed that players at a higher standard of competi-
tion perform significantly more high-intensity running
than those at a lower standard (Ekblom, 1986; Bangsbo
et al., 1991). It is unclear whether elite international
players perform even more high-intensity exercise.
Furthermore, most previous studies were performed
more than 10 years ago and it is unclear how the
exercise intensity in professional soccer has developed
over the last decade. Information about the physical
performance of players in different playing positions
* Author to whom all correspondence should be addressed.
e-mail: jbangsbo@aki.ku.dk
Journal of Sports Sciences, 2003, 21, 519–528
Journal of Sports Sciences ISSN 0264-0414 print/ISSN 1466-447X online
#
2003 Taylor & Francis Ltd
DOI: 10.1080/0264041031000071182
and at various stages of a season is also limited. To
examine these aspects of soccer, it woul d be necessary
to have a large number of participants representing all
playing positions, since considerable variations in
activity patterns have been found from game to game
and between players within each position (Bangsbo,
1994; Reilly, 1997).
It is also unclear to what extent the players experience
fatigue during a high-standard soccer game. Several
researchers have observed a reduction in total distance
covered in the second half compared with the first
(Reilly and Thomas, 1976; Van Gool et al ., 1988 ;
Bangsbo et al., 1991). This reduction may indicate the
development of fatigue in the second half, although
total distance covered appears not to be a perfect
indicator of physical performance in a match (Bangsbo,
1994). Rebelo et al. (1998) reported that soccer players
had a reduced ability to perform repeated sprints after
than before a soccer match, indicating that fatigue had
developed before the end of a soccer game. The extent
to which fatigue occurs among soccer players during
elite competitive games could be assessed by analysing
high-intensity exercise throughout a match and by
comparing the performance of substitute players, who
only participate in parts of the second half, with the
performance of players taking part in the entire match.
An assumption would be that the demand and
opportunity to engage in high-intensity runs is constant
during a game.
The aims of the present study were: (1) to examine
the activity pattern of high-standard soccer players in
relation to positional role and performance level; (2) to
establish whether fatigue occurs during a high-standard
soccer match; and (3) to examine players’ match
performance at different stages of the season.
Materials and methods
Participants
Eighteen professional top-class soccer players (age
26.4 + 0.9 years, body mass 75.4 + 1.5 kg, height
1.80 + 0.01 m; mean + s
x
_
) participated in the study,
all of whom were playing for an elite European team,
competing in the Italian league and in the European
Champions League. Fourteen of these players were also
playing for their respective national teams (representing
five nations), which were among the strongest in the
world (ranked 1–10 on the official FIFA list). In
addition, 24 professional players (age 26.5 + 1.0 years,
body mass 75.4 + 1.7 kg, height 1.81 + 0.02 m) from
the top Danish league took part in the study, five of
whom were playing for their respective national teams,
which were of moderate ability (ranked higher than 20
on the official FIFA list).
The players were observed individually by video
filming in up to seven different competitive matches
throughout two consecutive seasons. A season was
divided into three phases – start, middle and end of the
season – to assess seasonal changes in match perfor-
mance. The sample included 11 central defenders, 9
full-backs, 13 midfield players and 9 attackers. To
evaluate any differences between players who played an
entire match and substitute players, 13 top-class
substitutes (age 25.0 + 1.0 years, body mass
75.8 + 1.6 kg, height 1.81 + 0.01 m) were also video-
filmed (five were also represented in the group that
played an entire match). The substitutes partici pated in
a part of the second half of the match and were observed
in up to four matches on different occasions over the
season. Eight of these players also played for their
respective national teams (ranked 1–10 on the official
FIFA list).
Match analysis
Each player was video-filmed close up during the entire
match. The VHS-format cameras (NV-M50, Panaso-
nic, Germany) were positioned at the side of the pitch,
at the level of the midfield line, at a height of about
15 m and at a distance of 30–40 m from the touchline.
The videotapes were later replayed on a monitor
for computerized coding of the activity pattern .
The following locomotor categories were used:
standing (0 km × h
71
), walking (6 km × h
71
), jogging
(8 km × h
71
), low-speed running (12 km × h
71
), mod-
erate-speed running (15 km × h
71
), high-speed running
(18 km × h
71
), sprinting (30 km × h
71
) and backward
running (10 km × h
71
). The locomotor categories were
chosen in accordance with Bangsbo et al. (1991),
whereas the mean speed for each category was
determined after detailed studies of the videotapes.
Thus, the time for the player to pass pre-markers in the
grass, the centre circle and other known distances was
used to calculate the speed for each activity of
locomotion. The above activities were later divided
into four locomotor categories: (1) standing; (2)
walking; (3) low-intensity running, encompassing jog-
ging, low-speed running and backward running; and (4)
high-intensity running, consisting of moderate-speed
running, high-speed running and sprinting. The fre-
quency and duration of each activit y were recorded and
the data are presented for 5-, 15-, 45- and 90-min
periods. The distance covered for each activity within
each interval was determined as the product of the total
time and mean speed for that activity. The total
distance covered during a match was calculated as the
sum of the distances covered during each type of
activity. The peak distance covered in high-intensity
running in a 5-min period will be presented. This
520 Mohr et al.
period represents that 5 min which contains the most
high-intensity running in a game and is specific for each
of the monitored players. In addition, the numbe rs of
headers and tackles were registered for each half.
Reproducibility of the results obtained by the time–
motion analysis has previously been determined. In a
study by Krustrup and Bangsbo (2001), no systematic
differences were observed in a test–retest analys is of a
match and the mean intra-individual difference in total
distance covered was less than 0.2 km (coefficient of
variation, CV = 1%). The intra-individual variations
(CV) in walking, low-intensity running, high-intensity
running and backward running were 2, 5, 3 and 3%,
respectively. These values are in the same range as those
reported by Bangsbo et al. (1991).
The 129 matches analysed in the present study were
all analysed by the same experienced observer, who has
analysed more than 250 soccer matches. The first and
the second hal f of each match were analysed in a
random order. Before analysis of each player was
initiated, the individual players’ styles of locomotion
were studied intensively and several validation tests
were performed for each player according to the pre-
determined categories of locomotion.
Yo-Yo intermittent recovery test
Both the top-class and the moderate players performed
the Yo-Yo intermittent recovery test (level 1) on two
occasions during the competitive season. The highest
Yo-Yo test score for each individual was used in the
comparison of test performance between standards of
competition as well as playing positions. After a 10-min
warm-up, the players performed the test, which consists
of repeated 2 6 20 m runs back and forth between the
start and finish line at a progressively increased speed
controlled by audio bleeps from a tape-recorder
(Bangsbo, 1995). Between running bouts, the partici-
pants had a 10-s active rest period cons isting of 2 6 5m
jogs. When the participants failed twice to reach the
finish line in time, the distance covered was recorded
and was used as the test result. Performance in this test
has been shown to be closely correlated with the
amount of high-intensi ty running performed by soccer
players (Krustrup et al., 2003) and referees (Krustrup
and Bangsbo, 2001) in high-standard soccer matches.
Statistical analysis
Results are presented as the mean + standard error of
the mean (s
x
_
). Differences between the first and second
half were determined using Student’s paired t-test.
Differences between the two standards of competition
were analysed using Student’s unpaired t-test. Differ-
ences between the four team positions were tested by a
one-way analysis of variance. Differences between 5- or
15-min periods within a match, as well as seasonal
changes in match activities, were determined using one-
way analyses of variance with repeated measurements.
In the event of a significant difference between
treatments, Tukey’s post-hoc tests were used to identify
specific differences between the means of the two
standards of play. Statistical significance was set at
P 5 0.05. The coefficient of variation was used as a
measure of intra-individual variation in high-intensity
running and total distance covered during matches and
was calculated as the standard deviation of the
difference between repeated measurements divided by
the mean and multiplied by 100 (Atkinson and Nevill,
1998).
Results
Match activities
The top-class players spent 19.5 + 0.7, 41.8 + 0.9 and
29.9 + 1.3% of the time during a game standing,
walking and in running of low intens ity, respectively,
which was similar to the moderate players (Fig. 1). The
top-class players performed more (P 5 0.05) high-
intensity running and sprinting than the moderate
players (8.7 + 0.5 vs 6.6 + 0.4% and 1.4 + 0.1 vs
0.9 + 0.1%, respectively; Fig. 1). In addition, the top-
class players spent more (P 5 0.05) time running
backwards than the moderate players (3.7 + 0.3 vs
2.9 + 0.2%). The numbers of low-intensity runs, high-
intensity runs and sprints were also higher (P 5 0.05)
for the top-class than the moderate players (587 + 23 vs
566 + 20, 217 + 13 vs 171 + 7 and 39 + 2 vs 26 + 1,
respectively; Table 1). There was no difference in the
mean duration of any of the activities between the two
groups (Table 1). The top-class players ran more
(P 5 0.05) at both a low and high intensity during the
first compared with the second half (31.1 + 1.3 vs
28.4 + 1.4% and 9.0 + 0.4 vs 7.7 + 0.5%, respectively;
Fig. 2). In addition, the top-class players sprinted more
(P 5 0.05) in the first than in the second half (1.6 + 0.1
vs 1.2 + 0.1%; Fig. 2). Similar differences (P 5 0.05)
were observed for the moderate players when low- and
high-intensity running and sprinting were compared in
the two halves (first half vs second half: 32.5 + 1.4 vs
29.6 + 1.2, 7.0 + 0.4 vs 6.3 + 0.3 and 0.9 + 0.1 vs
0.8 + 0.1%, respectively; Fig. 2). The top-class players
performed more (P 5 0.05) bouts of low- and high-
intensity running in the first than in the second half
(302 + 12 vs 285 + 12 and 114 + 6 vs 103 + 7,
respectively). The moderate players showed no sig-
nificant difference between the two halves (first half vs
second half: 281 + 11 vs 285 + 10 and 88 + 4 vs
82 + 4, respectively). Both the top-class and the
521Development of fatigue in high-standard soccer players
moderate players sprinted more frequently (P 5 0.05)
in the first than in the second half (22 + 2 vs 17 + 1 and
14 + 1 vs 12 + 1, respectively). For the top-class
players, the mean times of standing and walking
intervals were longer (P 5 0.05) in the second half than
in the first half (7.5 + 0.4 vs 6.5 + 0.3 s and 6.5 + 0.3
vs 6.2 + 0.3 s, respectively). There was no difference in
the mean duration of any of the other locomotor
activities betwe en the first and the second half for
players of either standard.
Distances covered
The total distance covered during a match was
10.86 + 0.18 km for the top-class players, which was
5% more (P 5 0.05) than for the moderate players
(10.33 + 0.26 km). The amount of high-intensity run-
ning during a match was 28% greater (P 5 0.05) for
top-class than for moderate players (2.43 + 0.14 vs
1.90 + 0.12 km). The distance covered when sprinting
was 0.65 + 0.06 km for to p-class players, which was
58% more (P 5 0.0 5) than for moderate players
(0.41 + 0.03 km). The distance covered during the
first half was longer (P 5 0.05 ) than in the second hal f
for top-clas s players (5.51 + 0.10 vs 5.35 + 0.09 km);
no difference was observed for moderate players
(5.20 + 0.14 vs 5.13 + 0.12 km). Both groups of
players covered a greater (P 5 0.05) distance in high-
intensity running in the first than in the second half of
the match (top-class: 1.27 + 0.07 vs 1.15 + 0.08 km;
moderate: 1.01 + 0.07 vs 0.90 + 0.06 km, respec-
tively). Furthermore, distance covered when sprinting
both for top-class and moderate players was also greater
(P 5 0.05) in the first than in the second half
(0.35 + 0.04 vs 0.30 + 0.03 and 0.21 + 0.03 vs
0.19 + 0.02 km, respectively).
Performance with in a match
The distance covered in high-intensity running by top-
class players in the last 15 min was 14–45% lower
(P 5 0.05) than in the first four 15-min periods
(0.32 + 0.03 vs 0.36 + 0.02 to 0.46 + 0.03 km; Fig.
3a). The moderate players covered 17–35% less
(P 5 0.05) distance in high-intensity running in the last
four 15-min periods than in the first 15 min
(0.27 + 0.03 to 0.31 + 0.03 vs 0.36 + 0.03 km; Fig.
3a). The distance covered by the to p-class players when
sprinting was shorter (P 5 0.05) in the last compared
with the first four 15-min periods (0.07 + 0.01 vs
0.09 + 0.01 to 0.13 + 0.02 km); there was no difference
for the moderate players (Fig . 3b). Thus, the distance
covered by top-class players when sprinting was 43% less
(P 5 0.05) in the last 15 min than in the first 15 min.
The peak distance covered in high-intensity running
in a 5-min period was 219 + 8 m for top-class players,
which was 27% more (P 5 0.05) than for the moderate
players (172 + 9 m) (Fig. 4). In the next 5-min period,
the amount of high-intensity running was 106 + 6m
for top-class players, which was 12% less (P 5 0.05)
than the average distance covered during all 5-min
intervals (except that of peak intensity; 121 + 4m).
The moderate players performed 94 + 8 m of high-
intensity runn ing in the 5 min after the most intense
period, which tended to be lower (P = 0.08) than the
average distance covered in high-intensity running
(105 + 5 m; Fig. 4). In addition, when the data for
the two groups of players were combined, the distance
covered in high-intensity running during the first 5 min
of the first half was greater (P 5 0.05) than during the
first 5 min of the second half (160 + 10 vs 130 + 7m,
respectively), whereas no significa nt difference was
apparent between the second 5-min period of the first
and the second half (130 + 9 and 109 + 7 m, respec-
tively).
The total number of headers and tackles performed
during a game was 15 + 2 and 20 + 2, respectively. No
difference was observed in the number of headers
between the first and second half (8 + 1 vs 7 + 1),
whereas the players performed more (P 5 0.05) tackles
during the first than the second half (12 + 1 vs 8 + 1).
When comparing the performance of players who
played the entire match with the performance of the
substitutes, the amount of high-intensity running in the
last 15 min of the game was 25% higher (P 5 0.05) for
the substitutes (0.40 + 0.03 vs 0.32 + 0.03 km). The
amount of sprinting in the last 15 min of the game was
St
W
J
LS
MS
HS
Sp
BR
45
40
35
30
25
20
15
10
5
0
Time (min)
*
*
*
*
*
Activit
y
cate
g
ories
Fig. 1. Locomotor categories for top-class players (&) and
moderate players (&) during a soccer game expressed as total
time (mean + s
x
_
) *Significant difference (P 5 0.05) between
groups. St = standing; W = walking; J = jogging; LS, MS and
HS = running at a low, moderate and high speed, respectively;
Sp = sprinting; BR = backwards running.
522 Mohr et al.
Table 1. Frequency, mean duration and percent of time spent on the locomotor categories for top-class and moderate players (mean + s
x
_
)
Low-speed Backwards Moderate-speed High-speed
Standing Walking Jogging running running running running Sprinting
Total
Frequency (n)
Top-class 163+6 379 + 10* 316 + 15 198 + 873+ 4* 109 + 7* 69 + 5* 39 + 2* 1346 + 34*
Moderate 163 + 10 398 + 12 321 + 13 185 + 860+ 496+ 549+ 326+ 1 1297 + 27
Mean duration (s)
Top-class 7.0 + 0.4 6.4 + 0.3 3.0 + 0.1 2.6 + 0.0 2.7 + 0.1 2.2 + 0.0 2.1 + 0.0 2.0 + 0.0 3.5 + 0.1
Moderate 7.1 + 0.4 6.4 + 0.3 3.1 + 0.1 2.7 + 0.1 2.7 + 0.1 2.4 + 0.0 2.2 + 0.0 1.9 + 0.0 3.6 + 0.1
% of total time
Top-class 19.5 + 0.7 41.8 + 0.9 16.7 + 0.9* 9.5 + 0.4 3.7 + 0.3* 4.5 + 0.3* 2.8 + 0.2* 1.4 + 0.1* 100.0
Moderate 18.4 + 1.5 43.6 + 0.8 19.1 + 0.9 9.4 + 0.4 2.9 + 0.2 3.8 + 0.3 1.9 + 0.1 0.9 + 0.1 100.0
*Significant difference (P 5 0.05) between top-class and moderate players.
