Journal of Sports Sciences

ISSN: 0264-0414 (Print) 1466-447X (Online) Journal homepage: http://www.tandfonline.com/loi/rjsp20

Activity profile and physiological demands of top-

class soccer assistant refereeing in relation to
training status

Peter Krustrup , Magni Mohr & Jens Bangsbo

To cite this article: Peter Krustrup , Magni Mohr & Jens Bangsbo (2002) Activity profile and
physiological demands of top-class soccer assistant refereeing in relation to training status,
Journal of Sports Sciences, 20:11, 861-871, DOI: 10.1080/026404102320761778

To link to this article: http://dx.doi.org/10.1080/026404102320761778

Published online: 09 Dec 2010.

Submit your article to this journal

Article views: 296

View related articles

Citing articles: 41 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rjsp20

Download by: [University of California Santa Barbara] Date: 06 November 2015, At: 03:05
 Journal of Sports Sciences, 2002, 20, 861± 871

Activity pro® le and physiological demands of top-class

soccer assistant refereeing in relation to training status
PETER KRUSTRUP, MAGNI MOHR and JENS BANGSBO*
Institute of Exercise and Sport Sciences, August Krogh Institute, University of Copenhagen, Universitetsparken 13,
DK-2100 Copenhagen é , Denmark

Accepted 22 March 2002

Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015

To determine the movement patterns and physiological demands of top-class soccer assistant referees, we per-
formed computerized time± motion analysis and measured heart rate and blood lactate concentration in 15
assistant referees during 22 competitive matches in the top Danish league. To relate match performance to the
physical capability of the assistant referees, they performed a 3 ´ 30 m sprint protocol before and after matches
and a laboratory treadmill test within 3 weeks of the games. The mean total distance covered by the top-class
assistant referees was 7.28 (range 5.78± 8.16) km, of which 1.15 (0.86± 1.44) km was high-intensity running and
1.16 (0.12± 2.34) km was sideways running. The amount of high-intensity running during a game was correlated
with performance of repeated sprints (r = 0.80, P < 0.05). Mean heart rate was 137 (117± 159) beats ´min-1,
corresponding to 73% (60± 88%) of maximal heart rate and 65% (53± 80%) of maximal oxygen uptake. Blood
lactate concentration was 4.7 (1.6± 11.0) and 4.8 (1.1± 13.7) mmol ´l-1 after the ® rst and second half, respec-
tively. Sprinting performance was poorer (P < 0.05) after than before the games. The peak distance to the
oþ side line was greater (P < 0.05) in the second than the ® rst half (7 ± 1 vs 5 ± 0 m). Our results show that: (1)
top-class assistant soccer refereeing is characterized by brief intense bouts of forward and sideways running
interspersed with long periods of low activity; (2) top-class soccer assistant referees have moderate aerobic
energy production during games with episodes of high aerobic and anaerobic energy turnover; (3) assistant
referees’ performance of repeated sprints correlates with the amount of high-intensity running performed in a
game; and (4) sprint performance decreases towards the end of a game, which appears to aþ ect assistant referees’
ability to keep up with play.

Keywords: blood lactate, heart rate, high-intensity running, muscle temperature, sprint performance,
time± motion analysis.

Introduction (Asami et al., 1988; Catterall et al., 1993; Krustrup

Detailed information is available about the activity
pro® le and physiological demands of soccer players
(Bangsbo, 1994a,b), whereas little is known about
soccer referees. To date, only three studies have
examined the movement patterns and physiological
strain of soccer referees (Asami et al., 1988; Catterall
et al., 1993; Krustrup and Bangsbo, 2001) and no stud-
ies have focused on assistant referees during match-play.
The total distance covered by players and referees
during a soccer match has been reported to be 7± 13 km
(Reilly and Thomas, 1976; Mayhew and Wenger, 1985;
Bangsbo et al., 1991; Bangsbo, 1994a,b) and 8± 12 km
* Author to whom all correspondence should be addressed. e-mail:
jbangsbo@aki.ku.dk
and Bangsbo, 2001), respectively. However, the total
distance covered during a match is a poor measure of
physical stress, since the exercise performed is of an
intermittent nature, where low-intensity activities
account for more than three-quarters of total match-
play. Thus, time± motion analyses that record the
number of activity changes during a match, as well as
distances covered at low and high speed, can provide
comprehensive information about the movement pat-
terns of assistant referees. It has previously been shown
that the amount of high-intensity running and sprint
performance decrease at the end of games for elite
soccer players (Rebelo et al., 1998; Mohr and Bangsbo,
2001) and top-class referees (Krustrup and Bangsbo,
2001). To determine whether this is also the case for
top-class assistant referees, comparisons could be made

Journal of Sports Sciences ISSN 0264-0414 print/ISSN 1466-447X online Ó 2002 Taylor & Francis Ltd
http://www.tandf.co.uk/journals
 862
of high-intensity running during diþ erent periods of a
game and of sprint performance immediately before and
after a game.
The physiological demands of assistant refereeing can
be assessed by the continuous recording of heart rate
and by measuring blood lactate concentration after each
half of a game. Determination of match heart rates, as
well as the. relationship between heart rate and oxygen
uptake (V O2), has demonstrated that aerobic loading
is high both for soccer players (Reilly and Thomas,
1976; Rohde and Espersen, 1988; Bangsbo, 1994a,b)
and referees (Catterall et al., 1993; Krustrup and
Bangsbo, 2001). The anaerobic energy contribution
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015
can also be examined during matches. Measurement of
blood lactate concentration during matches has shown
that anaerobic energy turnover is high in certain periods
of a game both for players (Ekblom, 1986; Bangsbo
et al., 1991) and referees (Krustrup and Bangsbo,
2001), re¯ ecting the frequent high-intensity exercise
that is performed. However, neither heart rate nor blood
lactate concentration have been recorded for assistant
referees during match-play.
Another important issue is the extent to which the
activity pro® le and physiological strain of assistant
referees is related to their physical capacities. Recent
studies have demonstrated that the amount of high-
intensity running performed by soccer players and
referees varies widely between individuals but only little
between successive matches for the same individual
(Bangsbo et al., 1991; Krustrup and Bangsbo, 2001),
indicating that performance at the highest standard is
mainly dictated by the physical capacity of players and
referees rather than diþ erences in the tempo of games.
However, this could be diþ erent for assistant referees,
as they are con® ned to operate within a zone that is half
the length of the touchline. To investigate this, analysis
of the match activities for the same assistant referee in
successive matches and for two assistant referees in the
same game could be made, in addition to physiological
tests and time-matched determinations of match
performance.
The aims of the present study were to examine: (1)
the movement patterns of top-class assistant referees;
(2) the physiological strain imposed on assistant referees
to meet these demands; (3) the performance of assistant
referees throughout a game; and (4) the relationship
between the physical capacity and match performance
of top-class assistant referees.
Materials and methods
Participants
Fifteen assistant referees, 13 of whom held FIFA
licenses, participated in the study. Their mean (range)
Krustrup et al.
age, height and body mass were 40 (32± 47) years,
1.81 (1.67± 1.91) m and 80.5 (69.1± 98.4) kg, respec-
tively. They all had more than 5 years of experience
in the top Danish league, the `Superliga’ . The assistant
referees were informed of the experimental pro-
cedures and possible discomfort associated with the
study before providing their informed consent to
participate.
Match analysis
Fifteen assistant referees were ® lmed during 22 matches
in the 1998± 99 and 1999± 2000 seasons to determine
their locomotor activity and their ability to keep up with
play. We positioned two VHS movie cameras (NV-M50,
Panasonic, Germany) at the side of the pitch, at the level
of the penalty arch, at a height of about 15 m and at a
distance of about 30 m from the touchline.
The ® rst camera ® lmed at a broad angle to evaluate
the assistant referees’ ability to keep up with the second
last defender (the oþ side line). The videotapes were
later replayed to register the distance to the oþ side line
in the mid-zone (the ® rst 20 m on each side of the half-
way line) and the attacking zones (the rest of the pitch),
using pre-markers in the grass as a ruler. The second
camera ® lmed the assistant referees close up during the
game for determination of locomotor activity. To assess
intra-individual variations in match activities, seven
assistant referees were ® lmed again during a competitive
match within 1 month of the ® rst ® lming. These video-
tapes were later replayed on a monitor for computerized
recording of their movement patterns.
The following locomotor categories were used: stand-
ing (0 km ´h-1), walking (6 km ´ h-1), jogging (8 km ´ h-1),
low-speed running (12 km ´h-1), moderate-speed
running (15 km ´ h-1), high-speed running (18 km ´ h-1),
sprinting (25 km ´ h-1), sideways running (12 km ´h-1)
and backwards running (10 km ´h-1). The locomotor
categories were chosen in accordance with Bangsbo
et al. (1991), whereas the mean speed for each loco-
motor activity was determined after detailed study of
the videotapes. Thus, the time for an assistant referee
to cover known distances between pre-markers in the
grass was used to calculate the speed for each locomotor
activity. The match activities were later divided into
four locomotor categories: (1) standing; (2) walking; (3)
low-intensity running, de® ned as jogging, low-speed
running, sideways running and backwards running; and
(4) high-intensity running, de® ned as moderate-speed
running, high-speed running and sprinting.
The frequency and duration of each activity were
recorded and the data presented for 5, 15 and 45 min
intervals. The distance covered by each locomotor
activity within each interval was determined as the
product of total time and mean speed for that activity.
 Physical demands of top-class soccer assistant refereeing
The total distance covered during a match was then
calculated as the sum of the distances covered during
each type of activity. Variations in results obtained by
time± motion analysis have previously been shown to be
in the range of 1± 5% in each of the individual activity
categories (Bangsbo et al., 1991; Krustrup and Bangsbo,
2001). The 22 match analyses were all performed by the
same experienced observer.
Heart rate and blood samples
Heart rate (HR) was recorded at 5 s intervals through-
out each game using a Polar Vantage NV heart rate
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015
monitor (Polar, Kempele, Finland). The assistant
referees attached the heart rate monitor approximately
45 min before kick-oþ . A blood sample was collected
from a ® nger prick 1± 2 min after the end of the ® rst
and second halves. Within 10 s of sampling, 100 ml
of blood was haemolysed in 100 ml of ice-cold triton
X-100 buþ er solution and later analysed for lactate and
glucose using a YSI 2300 lactate analyser (Yellow
Spring Instruments, Yellow Springs, OH, USA; Foxdal
et al., 1992).
Quadriceps muscle temperature
The temperature of the quadriceps was measured in 9 of
the 22 games using a needle thermistor (MKA08050-A,
Ellab A/S, Rù dovre, Denmark) with a precision of
0.1°C. Muscle temperature was measured eight times ±
before and after the warm-up and <1, 10 and 15 min
after each half. All measurements were performed in
the vastus lateralis of the right leg at a muscle depth of
~ 3 cm, adjusting for the thickness of the skinfold using
a Harpenden skinfold caliper (British Indicators Ltd,
UK). After each game, the thermistors were calibrated
against a mercury thermometer.
Fluid loss and intake
To determine sweat loss during a game, the assistant
referees were weighed immediately before, at half
time and immediately after the match using a platform
scale (model 1-10, OHAUS, New Jersey, USA). The
assistant referees wore shorts when being weighed. They
were allowed to drink water ad libitum during the game
and their intake was recorded.
Testing
Within 3 weeks of ® lming, 12 assistant referees under-
took an incremental treadmill run to determine maximal
heart .rate and oxygen uptake, as well as the individual
HR± V O2 relationship. In the laboratory, running at
863
speeds of 10, 12 and 14 km ´ h-1 was performed in 6 min
bouts separated by 2 min rest, followed by an incre-
mental maximal test. The maximal test began at a speed
of 14 km ´ h-1 for 2 min and continued at 16 km ´ h-1 for
30 s with a stepwise increase of 1 km ´ h-1 every 30 s until
exhaustion. Time to exhaustion was recorded.
Heart rate was recorded at 5 s intervals throughout
the protocol using a Polar Vantage NV heart rate moni-
tor (Polar, Kempele, Finland); pulmonary oxygen
uptake was measured during the last 2 min of each
submaximal running speed and during the maximal
test by a MedGraphics CPX/D online system (St. Paul,
Minneapolis,
. MN, USA). Individual maximal oxygen
uptake (V O2max) and maximal heart rate (HRmax) were
determined as the peak values reached in 15 s and 5 s,
respectively, at the end of the incremental maximal test.
In addition, the assistant referees (n = 15) performed
an oý cial FIFA test on a 400 m track consisting of a
12 min run followed by two 50 m sprints. The time
between trials was ~ 20 min. The distance covered in the
12 min run and the times for each of the two sprints
were recorded. In an additional study, performance of
repeated sprints was determined before and after the
game. In total, 12 assistant referees performed three
30 m sprints, separated by 25 s of active recovery,
approximately 5 min before, and within 2 min of the
end of, the game. The sprint times were recorded by
infrared light sensors, with a precision of 0.01 s (Time
It, Eleiko Sport, Halmstad, Sweden).
Statistical analysis
Diþ erences in match activities and physiological
responses to match-play during the ® rst and second
halves of the game were determined by a Student’ s
paired t-test. Changes in match activities and heart rates
within each 15 min period of the game were evaluated
by one-way analysis of variance (ANOVA) with repeated
measures. Performance on the repeated sprint test
immediately before and after the game was evaluated by
a two-way ANOVA with repeated measures. When a
signi® cant interaction was detected, the data were
subsequently analysed using a Newman-Keuls post-hoc
test. Correlation coeý cients were determined and
tested for signi® cance using Pearson’ s product± moment
coeý cient. The coeý cient of variation was used as a
measure of intra-individual variation in match per-
formance and calculated as the standard deviation of
the diþ erences between repeated measures divided by
the mean and multiplied by 100 (Atkinson and Nevill,
1998). Signi® cance was set at P < 0.05. The data are
presented as the mean ± standard error of the mean (sxÅ ).
The standard error of the mean is only given in the text
where this value cannot be obtained from a table or a
® gure.
 864
Results
Match performance
Match activities. The top-class assistant referees
(n = 15) were standing and walking for 43.9% (36.4±
55.4%) and 32.8% (26.2± 41.8%) of the total time,
respectively. Low-intensity running accounted for
19.3% (13.3± 25.7%) of the total time, whereas high-
intensity running accounted for the remaining 4.1%
(3.2± 5.2%) of the total time (Fig. 1). The time standing
increased (P < 0.05) in the second half compared with
the ® rst half, whereas the amount of walking and low-
intensity running was reduced (P < 0.05) in the second
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015
half (Fig. 1). No diþ erences were observed in high-
intensity running between the two halves. The mean
frequency of low-intensity running activities was 375
(288± 501) with a mean duration of 2.9 ± 0.1 s, whereas
110 (89± 147) high-intensity running activities were
performed with a mean duration of 2.1 ± 0.0 s. The
total number of match activities recorded was 1053
(832± 1459), corresponding to a change in activity every
5 s (Table 1).
Match distances. The distance covered when walking
was 3.08 (2.33± 3.83) km, whereas jogging, low-speed
running, backwards running and sideways running
accounted for 1.03 (0.73± 1.59), 0.80 (0.57± 1.04), 0.06
(0.01± 0.16) and 1.16 (0.12± 2.13) km, respectively
(Table 1). The total distance covered during a match
was 7.28 (5.78± 8.16) km, with a decrease (P < 0.05)
from 3.73 ± 0.07 km in the ® rst half to 3.55 ± 0.12 km
in the second half (Fig. 2a). The distance covered by
Fig. 1. Match activities during the ® rst (j ) and second ( )
halves expressed as a percentage of total time (mean ± sxÅ ).
* Signi® cant diþ erence (P < 0.05) between the ® rst and second
halves. S = standing, W = walking, J = jogging, LS = low-speed
running, MS = moderate-speed running, HS = high-speed
running and sprinting, SB = sideways and backwards running.
Krustrup et al.
low-intensity running decreased (P < 0.05) in the
second compared with the ® rst half, because less
(P < 0.05) sideways running was performed in the ® rst
30 min of the second than the ® rst half (Fig. 2b). The
distance covered by high-intensity running was 1.15
(0.86± 1.44) km, with no diþ erence between the two
halves (Fig. 2c), whereas it decreased by 25 ± 8% from
the ® rst to the last 15 min of the game (Fig. 2c). The
mean distance covered in high-intensity running bouts
was 14 (12± 15) m; 90% (81± 97%) of these were shorter
than 25 m (Fig. 3). The mean intra-individual variation
in total distance covered and high-intensity running
of the seven assistant referees who were tested twice
was 0.49 (0.09± 1.39) km and 0.18 (0.05± 0.32) km,
respectively, with coeý cients of variation of 9 and 13%,
respectively. In comparison, the coeý cients of variation
for total distance covered and high-intensity running
between the two assistant referees within the same
game (n = 8) were 17 and 22%, respectively. The intra-
individual variation in walking, low-intensity running
and sideways running for repeated analysis of the same
referee was 0.52 (0.19± 0.82), 0.79 (0.20± 1.78) and
0.47 (0.08± 0.93) km, respectively, with coeý cients of
variation of 13, 17 and 25%, respectively.
Distance from oþ side line. The mean distance from the
oþ side line during the game was 2 (1± 3) m, with no
signi® cant diþ erences between the two halves. Peak
distance from the oþ side line was 7 (4± 12) m, with an
increase (P < 0.05) from 5 ± 0 to 7 ± 1 m from the ® rst
to the second half. Mean and peak distances to the
oþ side line were higher (P < 0.05) in the attacking
zone than the middle zone (3 ± 0 vs 2 ± 0 m and 7 ± 1 vs
3 ± 1 m, respectively).
Sprinting performance during match. The time to com-
plete the 30 m sprints performed before the game
increased (P < 0.05) from 4.74 ± 0.08 s for the ® rst
sprint to 4.92 ± 0.07 s for the third sprint. Immediately
after the game, the times to complete the ® rst and
second sprint were 0.13 ± 0.04 and 0.07 ± 0.03 s longer
(P < 0.05) than before the game (Fig. 4). A signi® cant
correlation was noted between the mean time of the
three 30 m sprints performed immediately after the
game and peak distance to the oþ side line (r = 0.67,
P < 0.05).
Physiological responses
Heart rate and estimated oxygen uptake. The mean heart
rate of the assistant referees during the match was
137 ± 3 (117± 159) beats ´ min-1, which corresponds to
73 ± 2% (60± 88%) of individual HRmax. Mean heart
rate during the ® rst half was 140 beats ´min-1, which
decreased (P < 0.05) to 134 beats ´ min-1 during the
 Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015

Table 1. Match activities and distances covered by 15 top-class assistant referees (mean ± sxÅ )

Low- Sideways and Moderate- High-

speed backwards speed speed
Standing Walking Jogging running running running running Sprinting Total

Percent of total time

First half 41.4 ± 1.1 34.0 ± 1.2 8.8 ± 0.7 4.3 ± 0.2 7.4 ± 0.9 2.2 ± 0.1 1.3 ± 0.1 0.8 ± 0.1 100
Second half 46.2 ± 1.7* 31.7 ± 1.3* 7.8 ± 0.4 4.2 ± 0.4 6.0 ± 0.9* 2.1 ± 0.1 1.2 ± 0.1 0.7 ± 0.1 100
Overall 43.9 ± 1.2 32.8 ± 1.1 8.3 ± 0.5 4.3 ± 0.2 6.7 ± 0.9 2.1 ± 0.1 1.2 ± 0.1 0.8 ± 0.1 100

Frequency (n)
First half 119 ± 9 164 ± 5 81 ± 6 49 ± 3 63 ± 7 28 ± 2 17 ± 1 10 ± 1 532 ± 23
Second half 121 ± 8 164 ± 5 79 ± 8 50 ± 3 51 ± 6* 28 ± 2 16 ± 2 10 ± 1 521 ± 22
Overall 240 ± 17 328 ± 14 161 ± 14 100 ± 4 115 ± 13 56 ± 2 33 ± 3 20 ± 1 1053 ± 43

Mean duration (s)

First half 10.2 ± 0.6 5.9 ± 0.3 2.9 ± 0.1 2.4 ± 0.1 3.1 ± 0.2 2.1 ± 0.1 2.0 ± 0.1 2.0 ± 0.1 5.4 ± 0.2
Second half 11.7 ± 0.8* 5.8 ± 0.3 2.8 ± 0.1 2.4 ± 0.1 3.2 ± 0.2 2.1 ± 0.1 2.1 ± 0.1 2.1 ± 0.2 5.6 ± 0.2
Overall 10.9 ± 0.5 5.9 ± 0.3 2.9 ± 0.1 2.4 ± 0.1 3.2 ± 0.1 2.1 ± 0.1 2.0 ± 0.1 2.1 ± 0.1 5.5 ± 0.2

Distance covered (km)

First half 0.00 ± 0.00 1.56 ± 0.05 0.54 ± 0.04 0.39 ± 0.02 0.66 ± 0.08 0.25 ± 0.02 0.17 ± 0.02 0.16 ± 0.02 3.73 ± 0.07
Second half 0.00 ± 0.00 1.51 ± 0.06 0.50 ± 0.03 0.40 ± 0.03 0.56 ± 0.08* 0.25 ± 0.02 0.17 ± 0.03 0.15 ± 0.02 3.55 ± 0.12*
Overall 0.00 ± 0.00 3.08 ± 0.11 1.03 ± 0.06 0.80 ± 0.04 1.22 ± 0.15 0.50 ± 0.03 0.34 ± 0.03 0.31 ± 0.04 7.28 ± 0.17

* Signi® cant diþ erence between ® rst and second half.

 866 Krustrup et al.
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015

Fig. 2. Total distance covered (a), sideways running (b) and high-intensity running (c) for 15 min intervals during the ® rst
(j ) and second ( ) halves of the game (mean ± sxÅ ). * Signi® cant diþ erence (P < 0.05) between the ® rst and second halves.
#
Signi® cant diþ erence (P < 0.05) from the ® rst 15 min period.

second half (Fig. 5a). Heart rate was observed to be

Fig. 3. Number of high-intensity runs at given distances
during the ® rst (j ) and second ( ) halves (mean ± sxÅ ).
in the range 130± 170 beats ´min-1 for 60 ± 5% of total
match-play, whereas it was below 130 beats ´min-1 for
36 ± 6% of total match-play. Heart rate during the
second half was found to be in the range of 150± 170
beats ´min-1 for a shorter time (P < 0.05) and below
130 beats ´ min-1 for a longer time (P < 0.05) than in
the ® rst half (Fig. 5b). Peak heart rate during the game
was 170 ± 3 (148± 186) beats ´min-1, corresponding to
90 ± 2% (74± 100%) of HRmax. The intra-individual
variation in mean heart rate was 7 (2± 12) beats ´min-1
(coeý cient of variation = 6%). Mean oxygen uptake
estimated by the individual relationship between heart
rate and oxygen uptake obtained during the treadmill
test (n = 12, see below) was 2.32 ± 0.07 (1.96± 2.71)
l ´min
.
-1
, which corresponds to 65 ± 2% (53± 80%) of
V O2max.
 Physical demands of top-class soccer assistant refereeing
Blood lactate and glucose concentrations. Blood lactate
concentration was 4.7 (1.6± 11.0) and 4.8 (1.1± 13.7)
mmol ´l-1 after the ® rst and second half, respectively
(Fig. 6). Blood glucose concentration was 5.1 ± 0.1
(4.7± 5.7) mmol ´l-1 after the ® rst half, which was
higher (P < 0.05) than after the second half: 4.4 ± 0.2
(3.7± 5.3) mmol ´ l-1.
Muscle temperature. Quadriceps muscle temperature
increased (P < 0.05) from 36.5°C (35.7± 37.3°C) to
38.4°C (37.7± 39.2°C) during the warm-up and was
elevated further (P < 0.05) to 38.8°C (38.5± 39.2°C)
during the ® rst half (Fig. 7). At half time, muscle
temperature decreased (P < 0.05) markedly to 37.9°C
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015
(37.6± 38.1°C) and 37.5°C (36.4± 38.0°C) after 10 and
15 min, respectively. Thus, muscle temperature at the
Fig. 4. Times for three 30 m sprints performed before (d )
and immediately after the game ( ) (mean ± sxÅ ). * Signi® cant
diþ erence (P < 0.05) between before and after.
Fig. 5. Mean heart rate for 15 min intervals (a) and heart rate distribution (b) during the ® rst (j ) and second ( ) halves of the
game (mean ± sxÅ ). * Signi® cant diþ erence (P < 0.05) between the ® rst and second halves. # Signi® cant diþ erence (P < 0.05) from
the ® rst 15 min period.
867
start of the second half was 0.9 ± 0.2°C lower (P < 0.05)
than at the start of the ® rst half, whereas no diþ erence
was observed at the end of the two halves (Fig. 7).
Fluid loss and intake. Weight loss during a game was
0.81 ± 0.16 (0.49± 1.56) kg or 1.0 ± 0.2% (0.6± 2.0%)
of body mass. Fluid intake was 0.36 ± 0.05 (0.20± 0.50)
litres. Thus, estimated sweat loss during a game (n = 12)
was 1.17 ± 0.14 (0.83± 1.81) litres, which corresponds
to 1.5 ± 0.2% (1.0± 2.3%) of body mass.
Physiological capacity
The maximal oxygen uptake of the top-class assistant
referees (n = 12) was 3.71 (3.36± 4.50) l ´min-1 or 45.9
(40.9± 53.6) ml ´ min-1 ´kg-1. Oxygen uptake during tread-
mill running at speeds of 10, 12 and 14 km ´h-1 was
2.54, 2.95 and 3.35 l ´min-1, respectively, corresponding
Fig. 6. Individual and mean blood lactate concentrations
immediately after the ® rst and second halves (mean ± sxÅ ).
 868
to 69%
. (61± 78%), 80% (69± 89%) and 91% (77± 100%)
of V O2max. Heart rate was 147, 163 and 178 beats ´min-1
at running speeds of 10, 12 and 14 km ´h-1, respectively,
which corresponds to 77% (72± 84%), 86% (81± 94%)
and 94% (87± 100%) of the maximal heart rate of 189
(173± 200) beats ´ min-1. Time to exhaustion in the incre-
mental treadmill test was 4.21 (3.23± 5.25) min. The
distance covered in the 12 min run was 2889 (2760±
3175) m, corresponding to a mean speed of 14.4 ± 0.2
(13.8± 15.5) km ´h-1. The time for the best 50 m sprint
was 6.9 (6.2± 7.5) s, corresponding to a mean speed of
26.2 ± 0.4 (24.0± 29.1) km ´ h-1. The mean time of three
30 m sprints was 4.8 (4.4± 5.2) s, corresponding to a
mean speed of 22.4 ± 0.3 (20.7± 24.5) km ´h-1.
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015
Fig. 7. Quadriceps muscle temperature before, during and
after the game (mean ± sxÅ ). * Signi® cant diþ erence (P < 0.05)
between the ® rst and second halves. ¨ Signi® cant diþ erence
(P < 0.05) from pre-exercise values within each half of the
game.
Fig. 8. Individual relationships between (a) mean time of the three 30 m sprints and high-intensity running and (b) maximal
oxygen uptake and total distance covered during a game. The correlation coeý cients for the two relationships were 0.80
(F = 17.8, P < 0.002, n = 12) and 0.68 (F = 8.5, P < 0.02, n = 12), respectively.
Krustrup et al.
Test results in relation to match activities
Strong inter-individual relationships were observed
between the amount of high-intensity running during
match-play and performance of the repeated sprint test
(r = 0.80, P < 0.05; Fig. 8a), and between maximal
oxygen uptake (r = 0.64, P < 0.05) and the 50 m sprint
(r = 0.44, P < 0.05). No correlation was found between
high-intensity running during a game and time to
exhaustion in the incremental treadmill test (r = 0.43) or
time to complete the 12 min run (r = 0.25). In addition,
a signi® cant inter-individual relationship was found
between total distance covered and maximal oxygen
uptake (r = 0.68, P < 0.05, Fig. 8b). No correlation was
found between any of the test results and the amount of
sideways running during a game.
Test results in relation to physiological strain and match
performance
A signi® cant inverse relationship was found between
mean relative aerobic loading (percent of HRmax)
during a match and maximal oxygen uptake (r = -0.56,
P < 0.05) as well as time to complete the 12 min run
(r = 0.46, P < 0.05). In addition, a signi® cant relation-
ship was detected between peak distance to the oþ side
line during a game and mean time to complete three
repeated sprints after the game (r = 0.71, P < 0.05) and
the best time of two 50 m sprints (r = 0.60, P < 0.05).
Discussion
Our results show that the movement pattern of top-class
assistant referees is characterized by long periods of
low activity interspersed with short periods of intense
 Physical demands of top-class soccer assistant refereeing
exercise. The total distance covered during a match was
6± 8 km, of which more than 2 km was high-intensity
running and sideways running. The physiological
response to this movement pattern is relatively high
heart rates during a game and, in some cases, high
heart rates and blood lactate concentrations, re¯ ecting
a moderate aerobic energy contribution throughout
the game and episodes of considerable aerobic and
anaerobic energy production. Aerobic training status
was related to the total distance covered and relative
aerobic strain during match-play, whereas sprinting per-
formance was related to the amount of high-intensity
running and distance to the oþ side line. Towards the
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015
end of the game, high-intensity and sideways running
were reduced and the performance of repeated sprints
was impaired. This appears to have a negative impact on
the ability to keep up with play and achieve the best ® eld
of vision across the pitch.
Mean distance covered by the assistant referees was
7 km, which is only about two-thirds of the distance
covered by referees (Asami et al., 1988; Catterall et al.,
1993; Krustrup and Bangsbo, 2001) and soccer players
(Ekblom, 1986; Bangsbo et al., 1991). However, the
total distance covered during a game is a poor measure
of physical stress during a match, since an assistant
referee is standing and walking for more than 75% of
the game. A better indicator of the demanding periods
of the game is the amount of high-intensity running
(Bangsbo et al., 1991; Krustrup and Bangsbo, 2001) as
well as sideways running, which has been shown to be
more demanding of energy than forward running (Reilly
and Bowen, 1984). Also, the number of direction and
activity changes should be taken into account, since
this provides an indication of the work required for
accelerations and decelerations during a game. There-
fore, that top-class assistant referees on average perform
more than 110 high-intensity running activities and
100 bouts of sideways running, and that more than 225
direction changes and a total of 1000 activity changes
were registered during the matches, indicates that the
physical stress of an assistant referee is high during cer-
tain parts of a game. This is supported by blood lactate
concentration at half time and after the match being 5
mmol ´l-1 on average, with individual values exceeding
10 mmol´ l-1 (Fig. 6). These values are similar to those
recorded for referees (Krustrup and Bangsbo, 2001)
and soccer players (Ekblom, 1986; Bangsbo, 1994b),
as well as for moderately to well-trained individuals
performing 15 m sprints interspersed with long periods
of recovery (Balsom et al., 1992), and demonstrate that
the anaerobic mechanisms are highly stimulated during
a game. Even the two assistant referees who had low
blood lactate concentrations after both halves might
have produced high amounts of lactate during the game,
since previous studies have demonstrated that blood
869
lactate concentration only represents activities per-
formed during the last minutes before blood sampling
(Bangsbo et al., 1991; McInnes et al., 1995; Krustrup
and Bangsbo, 2001).
Information about the aerobic loading of assistant
referees during a game was obtained from the heart rate
measurements. In the present study, the mean heart rate
of the top-class assistant referees was 137 beats ´min-1,
which is lower than that for referees (Catterall et al.,
1993; Krustrup and Bangsbo, 2001) and soccer players
(Rohde and Espersen, 1988; Bangsbo, 1994a,b), but
still corresponds to 73% of individual maximal heart
rate. In addition, we found that heart rate was above
80% HRmax for more than 29 min or 31% of match-play.
This indicates that the aerobic system is under moderate
demand throughout a game and under heavy demand
at certain times. From the results of treadmill running,
it can be estimated that the . relative intensity during
a game was on average 65% V O2max. This is probably a
slight overestimation, since isometric muscle actions,
thermal and emotional stress . can elevate heart rate
beyond the normal HR± V O2 relationship (Bangsbo,
1994a,b). The emotional stress experienced by the
assistant referees in the present study was probably low,
as re¯ ected by low heart rates before the game and at
half time. However, some static exercise is performed
to hold the ¯ ag and to perform sideways running with
frequent changes in direction. Additionally, some
dehydration occurred during the matches, since sweat
loss during a game was, on average, 1.2 litres and almost
2 litres for some assistant referees, whereas ¯ uid intake
only amounted to 0.4 litres. Nevertheless, the mean
relative intensity may have been above 60% during the
game, which has been shown to cause decrements
in cognitive function (Reilly and Smith, . 1986). In
addition, exercise intensity reached 80% V O2max during
periods of the game, which has been suggested to cause
decrements in complex psychomotor functions (Reilly
and Smith, 1986), making decision-making prone to
error.
When comparing the top-class assistant referees’
activity pro® les in diþ erent parts of the game, total dis-
tance covered was less in the second half. This was due
to a decreased amount of walking and sideways running
in the second half; the amount of high-intensity running
was unaltered. When comparing each 15 min period of
the game, the amount of high-intensity running was
reduced at the end of the second half (Fig. 2). As the
movement pattern of assistant referees is dictated by the
second last defender, these changes could be due to
fatigue and alterations in the tempo of the game. There-
fore, we assessed whether the assistant referees were able
to maintain their sprinting performance throughout a
game. We found that 11 of 12 assistant referees were
slower in three 30 m sprints performed immediately
 870
after the game than before the game, indicating that
sprinting ability is impaired during a game, even though
less energy-demanding activities are performed in
between the high-intensity running bouts. Another indi-
cation of the development of fatigue was that the peak
distance to the oþ side line was greater in the second
half and exceeded 10 m for some assistant referees.
Although sprint performance also decreases during a
game for soccer players (Rebelo et al., 1998; Mohr and
Bangsbo, 2001), it would appear that the performance
of top-class assistant referees is reduced to the extent
that there is a negative impact on the ability to keep up
with play and on achieving the best ® eld of vision across
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015
the pitch. Large diþ erences were also observed in the
initial phase of the two halves. Thus, the amount of
high-intensity running was 33% lower and the distance
covered by sideways running was 30% less in the ® rst
5 min of the second than the ® rst half. This may be due
to the lower tempo of the game in the initial phase of the
second half, since the total distance covered by players
was lower at this time (Bangsbo et al., 1991). However,
Bangsbo et al. (1991) found that the amount of high-
intensity running for players was not diþ erent in the
initial phase of the two halves. It is somewhat surprising,
therefore, that high-intensity running and sideways
running were less for assistant referees in this phase
of the second half, since they had been resting for about
15 min and should have fully recovered. It could also be
that the muscles of the assistant referees became cold
during the rest period and they had to use the initial
phase of the second half to warm up a second time. We
found that quadriceps temperature was lowered by
more than 1.5°C during half-time and that it was about
1°C lower before the second half than before the ® rst
half. This may have in¯ uenced the activity pattern and
maximal sprint performance, since several studies have
demonstrated that maximal performance of short-
duration exercise is enhanced by an increased muscle
temperature (Asmussen and Bù je, 1945; Sargeant,
1987).
To achieve insight into the optimal physical prepar-
ation of assistant referees, it is important to what extent
the match performance is related to training status, the
tempo of the game oý ciated and the individual style
of the assistant referees. In the present study, we found
that there are large inter-individual variations in dis-
tance covered by high-speed running and sprinting
(two-fold) and sideways running (17-fold) during a
game. One possible explanation for this could be dif-
ferent tempos of the games oý ciated. In a recent study,
it was shown that the distances covered by assistant
referees are markedly lower in top-class women’ s
matches and third division men’ s matches (P. Krustrup,
M. Mohr and J. Bangsbo, unpublished data). How-
ever, when comparing the match activities of the same
Krustrup et al.
assistant referee in diþ erent matches in the top men’ s
league, we found that intra-individual diþ erences in
high-intensity running and sideways running were
relatively small. Furthermore, we found no correlation
between the high-intensity running of the ® rst and
second assistant referee in the same game (r = 0.03,
n.s.). Thus, it would appear that the observed diþ erences
in match performance are related to the physical
capacity or style of refereeing rather than the tempo
of the game oý ciated.
To allow us to focus directly on the in¯ uence of
training status on match performance, we carried out
several physiological tests and related the results to the
match activities of the assistant referees. From these
comparisons, we found that there is a strong correlation
between the performance of repeated sprints and the
amount of high-intensity running during a game (r =
0.80, P < 0.05). In addition, we observed that sprinting
performance both before and after a game was related to
the peak distance to the oþ side line (r = 0.71, P < 0.05).
Thus, it can be concluded that the marked inter-
individual diþ erences in performance of repeated sprints
is in part responsible for large variations in the assistant
referees’ ability to keep up with players, who at times
exceed sprinting speeds of 30 km ´ h-1 (Bangsbo, 1994a).
A high maximal aerobic power also appears to be
bene® cial for top-class assistant referees. Both the total
distance covered (r = 0.68, P < 0.05) and the amount
of high-intensity running (r = 0.64, P < 0.05) during a
game were related to the aerobic power of. the assistant
referees. In addition, we found that peak V O2 was nega-
tively correlated with the relative aerobic strain during
a match (r = -0.56, . P < 0.05), showing that assistant
referees with a high V O2max experience less physiological
stress.
The amount of sideways running was not correlated
with any of the physical characteristics of the assistant
referees. As the amount of sideways running was also
not correlated with the tempo of the game, the large
inter-individual variations in sideways running appear
to be caused by diþ erences in the individual style
of oý ciating. Performing frequent bouts of sideways
running should be an advantage for assistant referees,
since this gives them the opportunity to face towards
the pitch and to have a broader view of the play.
Interestingly, one FIFA-licensed assistant referee per-
formed less than 150 m of sideways running during a
game, whereas another performed more than 2.1 km.
One of the two assistant referees in the 2000 European
Cup Final has doubled the amount of sideways running
he performs during a match within the last couple of
years, so that he was the assistant referee that performed
most sideways running in the present study.
Taken together, these ® ndings suggest that the match
performance of an assistant referee is related to the
 Physical demands of top-class soccer assistant refereeing
ability to perform repeated sprints and aerobic power,
whereas diþ erences in the tempo of games within the
same league have a minor in¯ uence on match activities.
The individual style of assistant refereeing mostly aþ ects
the amount of sideways running during a game.
In summary, we found that top-class assistant referee-
ing is characterized by brief periods of intense exercise
interspersed with long periods of low-activity recovery.
The physiological demands related to these activities
were a high anaerobic and aerobic energy turnover in
parts of the game and a moderate aerobic energy con-
tribution during the rest of the game. In addition, we
found that sprint performance and the ability to recover
Downloaded by [University of California Santa Barbara] at 03:05 06 November 2015
between the intense exercise periods are essential for an
assistant referee to keep up with play throughout a
game. Thus, regular sprint training and match-speci® c
intermittent exercise training may improve the standard
of top-class assistant refereeing.
Acknowledgements
This study was supported by the Danish Football Association.
We would like to thank the assistant referees involved in
the study for their committed participation. The excellent
technical and practical assistance of Chris McDonald and
Preben Haarbye is appreciated. We would also like to thank
the camera crew: Per Hansen, Jan Malm-Hansen, Christoþ er
Krustrup and Lena Varntoft.
References
Asami, T., Togari, H. and Ohashi, J. (1988). Analysis of
movement patterns of referees during soccer matches.
In Science and Football (edited by T. Reilly, A. Lees,
K. Davids and W.J. Murphy), pp. 341± 345. London:
E & FN Spon.
Asmussen, E. and Bù je, O. (1945). Body temperature
and capacity for work. Acta Physiologica Scandinavica, 10,
1± 22.
Atkinson, G. and Nevill, A.M. (1998). Statistical methods
for assessing measurement error (reliability) in variables
relevant to sports medicine. Sports Medicine, 26, 217± 238.
Balsom, P.D., Seger, J.Y., Sj”din, B. and Ekblom, B. (1992).
Physiological responses to maximal intensity inter-
mittent exercise. European Journal of Applied Physiology, 65,
144± 149.
Bangsbo, J. (1994a). The physiology of soccer ± with special
871
reference to intense intermittent exercise. Acta Physiologica
Scandinavica, 151, suppl. 619.
Bangsbo, J. (1994b). Energy demands in competitive soccer.
Journal of Sports Sciences, 12, S5± S12.
Bangsbo, J., Nù rregaard, L. and Thorsù e, F. (1991). Activity
pro® le of competition soccer. Canadian Journal of Sports
Science, 16, 110± 116.
Catterall, C., Reilly, T., Atkinson, G. and Goldwells, A. (1993).
Analysis of the work rates and heart rates of association
football referees. British Journal of Sports Medicine, 27,
193± 196.
Ekblom, B. (1986). Applied physiology of soccer. Sports Medi-
cine, 3, 50± 60.
Foxdal, P., Bergqvist, Y., Eckerblom, S. and Sandhagen, B.
(1992). Improving lactate analysis with the YSI 2300 GL:
hemolyzing blood samples makes results comparable with
those for deproteinized whole blood. Clinical Chemistry, 38,
2110± 2114.
Krustrup, P. and Bangsbo, J. (2001). Physiological demands of
top class soccer refereeing in relation to physical capacity:
eþ ect of intense intermittent exercise training. Journal of
Sports Sciences, 19, 881± 891.
Mayhew, S.R. and Wenger, H.A. (1985). Time± motion
analysis of professional soccer. Journal of Human Movement
Studies, 11, 49± 52.
McInnes, S.E., Carlson, J.S., Jones, C.J. and McKenna, M.J.
(1995). The physiological load imposed on basketball
players during competition. Journal of Sports Sciences, 13,
387± 397.
Mohr, M. and Bangsbo, J. (2001). Development of fatigue
towards the end of a high level soccer match. Medicine and
Science in Sports and Exercise, 33, 160.
Rebelo, N., Krustrup, P., Soares, J. and Bangsbo, J. (1998).
Reduction in intermittent exercise performance during a
soccer match. Journal of Sports Sciences, 16, 482± 483.
Reilly, T. and Bowen, T. (1984). Exertional costs of changes
in directional modes of running. Perceptual and Motor Skills,
58, 149± 150.
Reilly, T. and Smith, D. (1986). Eþ ect of work intensity on
performance in a psychomotor task during exercise.
Ergonomics, 29, 601± 606.
Reilly, T. and Thomas, T. (1976). A motion analysis of work-
rate in diþ erent positional roles in professional football
match-play. Journal of Human Movement Studies, 2, 87± 97.
Rohde, E. and Espersen, T. (1988). Work intensity during
soccer training and match-play. In Science and Football
(edited by T. Reilly, A. Lees, K. Davids and W.J. Murphy),
pp. 68± 75. London: E & FN Spon.
Sargeant, A.J. (1987). Eþ ect of muscle temperature on leg
extension force and short term power output in humans.
European Journal of Applied Physiology and Occupational
Physiology, 56, 693± 698.