1. Physical and perceptual-cognitive demands of top-class
refereeing in association football
WERNER HELSEN* and JEAN-BAPTIST BULTYNCK
Department of Kinesiology, Katholieke Universiteit Leuven, Tervuursevest 101, B-3001 Heverlee (Leuven), Belgium
Accepted 14 June 2003
The aim of this study was to examine the physical and perceptual-cognitive demands imposed on UEFA top-
class referees and assistant referees during the final round of the Euro 2000 Championship. To investigate the
physical workload, the heart rates during matches were monitored by short-range radio telemetry and translated
to different workloads expressed as a percentage of maximal heart rate. For measurement of the perceptual-
cognitive workload, video-recordings of games were used to obtain the average number of observable decisions
taken by a referee. On average, referees and assistant referees performed the matches at 85+5% and 77+7% of
their maximal heart rate, respectively. Over the 31 games, the mean number of observable decisions was 137
(range 104–162), 64% of which were based on communication with the assistant referees and/or the fourth
official. To optimize the physical preparation of top-class match officials, the results of this study support the
application of intensive and intermittent training sessions, which should place priority on high-intensity aerobic
stimuli. In addition, video training is discussed as an additional method for improving match officials’ decision
making.
Keywords: decision-making skills, heart rate, refereeing.
Introduction
Several researchers have examined the activity profiles
and physiological demands of professional soccer
players during competitive games (Bangsbo, 1994a,b;
Reilly, 1997; Reilly et al., 2000). Only a few researchers,
however, have examined the movement patterns and
physiological strain of referees (Krustrup and Bangsbo,
2001). To date, no attempt has been made to examine
both the physical and perceptual-cognitive demands on
top-class referees and assistant referees.
In Table 1, we provide an overview of the results of
studies that have examined the different activity profiles
of referees. Based on the average speed criteria of the
different activity categories suggested by Krustrup and
Bangsbo (2001) and D’Ottavio and Castagna (2001a),
a generalized speed range is presented for the different
activities. In line with Bangsbo (1995), we compiled all
these activities into five main categories, expressed as a
percentage of individual maximal heart rate (HRmax).
For several reasons, it is difficult to compare the
findings of the studies in Table 1. First, there are
substantial differences in the methodology the authors
used to categorize the different activities and in the
technologies they employed to quantify the physiolo-
gical demands. Recent researchers (Krustrup and
Bangsbo, 2001; D’Ottavio and Castagna, 2001a), for
example, used twice as many activity categories as the
researchers in the late 1980s and the early 1990s.
However, even for those studies which involved the
same number of activity categories, as in the 2001
studies, different speed criteria have been applied to
obtain the amount of time spent in each of the
different activities. In terms of the physiological
demands, there are also quite important differences
between studies. Johnston and McNaughton (1994)
used activities over 85% HRmax as the upper range,
whereas Krustrup and Bangsbo (2001) and D’Ottavio
and Castagna (2001a, b) used activities over 90%
HRmax. On a related topic, we also need to consider
how the authors used HRmax as an indicator of the
physical demands. Krustup and Bangsbo (2001) used
peak values to determine maximal heart rate, while
both Johnston and McNaughton (1994) and D’Ottavio
and Castagna (2001a, b) used the ‘2207age in years’
rule to obtain HRmax. Because in all these studies the
physical demands were expressed in relation to HRmax,
it is clear that the way HRmax was determined may
impact on the results.
* Author to whom all correspondence should be addressed.
e-mail: werner.helsen@flok.kuleuven.ac.be
Journal of Sports Sciences, 2004, 22, 179–189
Journal of Sports Sciences ISSN 0264-0414 print/ISSN 1466-447X online # 2004 Taylor & Francis Ltd
DOI: 10.1080/02640410310001641502

2. A second reason why comparisons between studies
may only be done with caution is that differences in the
competitive standard of both referees and players may
be of importance for the results that have been
obtained. Harley et al. (1999), for example, reported
total distances of approximately 7 km for English
county standard referees, which approximates to the
distance an Italian referee in Serie A covers during a
single half (Castagna and D’Ottavio, 2001). In the same
way, players in the English Premier League cover
1.5 km more than their South American international
counterparts (Drust et al., 1998). Evidently, these
differences also impact on the referees’ performances.
A third factor that should be taken into consideration
when comparing different studies is the fitness of the
referees. As Krustrup and Bangsbo (2001) and Cas-
tagna and D’Ottavio (2001) have pointed out, the
fitness of referees clearly affects the distance they cover:
the better the physical condition of the referees, the
greater the amount of activity they will perform at both
high intensity and maximum speed during the game.
Evidently, fitness differences between groups may
partly explain the differences found between the studies
published in 2001 listed in Table 1.
Finally, differences in the styles of play (the ‘between-
country effect’) should also be taken into consideration
when comparing the results of different studies.
Specifically, it may be hypothesized that the patterns
of play in matches in the Italian Serie A are quite
different from those in the top Danish leagues. Hence
this difference in play will lead to a difference in the
activities of the referees. In the same way, it is difficult
to make a like-for-like comparison between the 1994
Australian results of Johnston and McNaughton due to
the standard of the football played, especially back in
1993 and the limitations to their study mentioned
previously (i.e. HRmax being calculated as 2207age in
years). Obviously, each country has a slightly different
story to tell with regards to the demands upon both
players and referees.
If only those studies using a similar methodology are
taken into consideration, it becomes clear that the total
distance covered by referees has increased significantly
from approximately 9 km in the early 1990s (Johnston
and McNaughon, 1994) to over 11 km in the most
recent studies (D’Ottavio and Castagna, 2001a,b).
However, the total distance covered during a game is
not the best measure of the physical load imposed on a
referee during a game. As Bangsbo et al. (1991) have
pointed out, the amount of high-intensity work is a
better indicator of the demanding periods of the game.
From this perspective, D’Ottavio and Castagna (2001a)
combined maximal ( 424.1 km × h71
), high (18.1–
24.0 km × h71
) and medium (13.1–18.0 km × h71
)
speed runs to represent the high-intensity activity
category. Within this high-intensity activity category,
Italian high-level soccer referees covered 4.8 km or
42% of the whole match distance. Heart rates
amounted to 89% of the estimated maximal heart rate
over the duration of a full game. These results are in
line with those of D’Ottavio and Castagna (2001b), but
twice those found in the early 1990s (e.g. Johnston and
McNaughton, 1994).
To date, limited research has examined the activity
patterns and physiological demands of soccer referee-
ing, and only one study has focused on assistant
refereeing. Using the same methodology Krustrup and
Bangsbo (2001) used for referees, Krustrup et al.
(2002) performed computerized time–motion analyses
and measured the heart rate and blood lactate
concentration of assistant referees in the top Danish
leagues. Compared with referees, who covered a total
distance of 10.1 km and performed 1265 different
activities, assistant referees covered 7.2 km and per-
formed 943 activities, resulting in a change in activity
every 4 and 6 s, respectively. For referees and assistant
referees, the mean heart rate was 162 and 137
beats × min71
, corresponding to 85% and 73% of
maximal heart rate, respectively. Taken together, the
results of Krustrup et al. (2002) show that top-class
assistant soccer refereeing is characterized by brief
intense bouts of forward and sideways running inter-
spersed with long periods of low activity.
The present study was the first to examine the
physical loads imposed on the two categories of match
officials. To extend the existing knowledge base using a
more homogeneous sample of games, we investigated
the physical load imposed on UEFA top-class referees
and assistant referees during matches in the final rounds
of the Euro 2000 Championship.
Perceptual-cognitive demands
The most important aspect of refereeing is the decision-
making process. Throughout the game, a referee has to
make correct decisions under time constraints. To date,
only two studies have examined the correctness of
referees’ and assistant referees’ decisions.
Van Meerbeek et al. (1987), for example, counted the
number of correct and incorrect decisions related to
specific laws of the game during the 1986 World Cup in
Mexico. They found that of all the decisions made in 16
games, 17% were incorrect (range 11–35%). Unfortu-
nately, these authors did not explain how a small panel
of researchers were able to conclude what was the
correct response for each decision (and decide that the
top-class referees were ‘wrong’). Without this informa-
tion, it is difficult to assess the accuracy of the findings
of their study. That it is very difficult to achieve an
errorless implementation of the laws of the game has
180 Helsen and Bultynck

3. Table 1. Relationship between different activity categories of referees and corresponding intensity categories expressed as a percentage of maximal heart rate
Intensity categories Asami et al. (1988)
Catterall et al.
(1993)
Johnston and
McNaughton
(1994)
Krustrup and
Bangsbo (2001)
D’Ottavio and
Castagna (2001a)
Speed
range
(km × h71
)
Maximal effort
(4 95% HRmax)
— Sprinting (11.8%) Sprinting (6.2%) Sprinting (0.5%) Sprinting (5.3%) 4 24.1
High-intensity
(86–95% HRmax)
Running (17.7%) — Running/striding
(12.1%)
High speed (1.6%)
Moderate speed (4.5%)
High speed (13.7%)
Medium speed (22.7%)
18.1–24.0
13.1–18.0
Low-intensity
(76–85% HRmax)
Jogging (48.5%) Jogging (47.0%) Jogging (46.6%) Low speed (9.0%)
Jogging (15.5%)
Low speed (36.7%) 9.1–13.0
6.1–9.0
Active recovery
(66–75% HRmax)
Walking (33.8%) Walking (23.0%) Walking (18.9%) Walking (40.3%) Walking (7.8%) 3.1–6.0
Passive recovery
(5 65% HRmax)
— — — Standing (22.6%) — 5 3
No specifications Backward (10.8%) Reverse run
(18.2%)
Backward (16.2%) Backward (5.9%) Back and sideways
(13.2%)
10.0

4. also been demonstrated recently by Oudejans et al.
(2000) for judging offside by assistant referees. Speci-
fically, they showed that the observation point of
assistant referees relative to the offside line is an
important determinant of incorrect decisions in judging
offside.
To our knowledge, no study has dealt with the
perceptual-cognitive demands of top-class refereeing,
revealing how many and what kinds of decisions a referee
has to make while ‘reading’ the game. Therefore, we
wished to gain a better understanding of this particular
aspect of decision making, as well as examine the physical
demands of football refereeing during top games.
The specific aims of this study were to: (1) examine
the physical load imposed on both referees and assistant
referees during the matches of the Euro 2000 Cham-
pionship; and (2) assess the perceptual-cognitive
demands of top-class refereeing during all the 31
matches of the Euro 2000 Championship, expressed
as the number and type of visible decisions.
Methods
Participants
The participants were all experienced UEFA top-class
referees (n = 17) and assistant referees (n = 17)
selected from different European countries. After a
brief explanation of the nature of the study, written
consent was obtained from the UEFA referees’
committee and from all the referees involved in the
tournament. The study was designed and conducted
in accordance with the ethical standards laid down in
the 1964 Declaration of Helsinki and approved by the
Committee for Ethical Considerations in Human
Experimentation of the Faculty of Physical Education
and Physiotherapy from the Katholieke Universiteit
Leuven.
Back-up team
For the Euro 2000 tournament, the refereeing team was
considered in a manner similar to that as the 16 national
sides. First, there was the assistance of the members of
the referees’ committee. To further assist the 34
referees, there was also a back-up team that consisted
of a medical doctor, three physiotherapists, a video
technician, a press officer, one assistant coach and one
head coach, who was also the coordinator of the back-
up staff. All these people were present full-time for the
group matches. Thereafter, one physiotherapist left as
well as the assistant coach, resulting in four full-time
staff members for the second tour. From the quarter-
finals until the final, three staff members were present
full-time.
Anthropometry
The age of the participants was recorded to the nearest
0.01 year by subtracting the decimal year of the
participant’s date of birth from the decimal year of the
day of the measurement (Baxter-Jones et al., 1995).
Measurement of height and weight was performed
during the week prior to the start of the tournament. All
measurements took place in the medical service room
and were carried out by the UEFA referees’ team
physician. Height was measured to the nearest milli-
metre with a Harpenden stadiometer (Holtain Ltd,
Crosswell, UK). For accuracy, the measurement was
performed twice and the results averaged. The mean
arithmetical difference between these two measure-
ments for standing height was 0.04 cm. Body mass was
measured to the nearest 100 g on a Soehnle electronic
scale (Soehnle, Bonn, Germany).
Physical workload during match-play
Heart rate measurement was used as an indication of
the physical workload imposed on referees and assistant
referees during match-play. Heart rate was recorded
using short-range radio telemetry (Polar Accurex Plus,
Kempele, Finland). A chest monitor transmitted the
heart rate signal to a receiver on the wrist. The heart
rate was recorded every 5 s. The data were stored in the
watch and then downloaded on a computer using the
Polar precision 3.0 software (Polar, Kempele, Finland).
After each and every game, the individual heart rate
data were printed out to inform each and every referee
about his performance and/or progress. During the
UEFA top-class referees’ course in Malta (February
2000), and the pre-tournament clinic in Brussels and
Leuven (April 2000), each referee was given a heart rate
watch. For those who were not familiar with the
equipment, these courses were used to provide them
with the necessary experience before the start of the
Euro 2000 tournament. The wearing of the heart rate
watch during matches was not mandatory. However, 47
match observations were obtained (13 for the referees
and 34 for the assistant referees). In Fig. 1, typical
examples are given of the graphical output of the heart
rate recordings of the referee during the Euro 2000 final
between France and Italy and one of his assistant
referees.
Perceptual-cognitive demands
To measure the number of ‘observable’ decisions a
referee made during a match, each game was recorded
on videotape. Using an interactive programme (Video-
Coach 2.0, Sport4S, Heverlee, Belgium), it was
possible to make an objective notational analysis of
182 Helsen and Bultynck

5. the referees’ decisions in all 31 matches. This analysis
was based on the observation of the body language of
the referee on the video replays of the games. A
distinction was made between three main categories.
First, there were those decisions that the referee had to
make, including signalling to let the play continue even
when an infringement had occurred (i.e. playing
‘advantage’), awarding direct and indirect free kicks
and penalties, and dealing with ‘simulations’ (i.e. when
players try to force a penalty or a free kick when there is
no real offence). The second category consisted of those
decisions in which the referee had to communicate with
one of his assistants before making a decision, as in the
case of corner kicks, goals, goal kicks, offsides, team
changes and throw-ins. Finally, the decisions under the
heading ‘others’ referred to dropped balls and the
12010560450
80
100
120
140
160
180
200
Time
First period Second period
85-95% HRmax
Extra time
Heartrate(beats.min-1)
(a)
60
80
100
120
140
160
180
Heartrate(beats.min-1)
12010560450
85-95% HRmax
First period Second period Extra time
(b)
Fig. 1. A typical example of the graphical output of heart rate recordings, showing heart rate versus time. The graphs show data for
the referee (a) and one of his assistants (b) during the final of the Euro 2000 Championship between France and Italy.
183Demands of top-class refereeing

6. awarding of red and yellow cards. The 31 match
analyses were all performed by the same experienced
observer, and were tested for intra- and inter-observer
reliability. During a match, however, a referee can also
decide not to interfere. The numbers and types of these
‘non-observable’ decisions were not taken into account.
Data reduction
In line with Krustrup and Bangsbo (2001) and
Krustrup et al. (2002), individual values for maximal
heart rate (HRmax) were determined as the peak values
reached in any of the 5 s periods observed during
matches, training sessions or incremental exercise tests
in the laboratory.
For the analysis of the physical load imposed on both
referees and assistant referees during the actual
matches, a distinction was made between five different
activity categories expressed as a percentage of maximal
heart rate: (1) maximal effort ( 495% HRmax); (2)
high-intensity (86–95% HRmax); (3) low-intensity (76–
85% HRmax); (4) active recovery (65–75% HRmax); and
(5) passive recovery ( 565% HRmax). These five
categories were based on findings in the literature
(Bangsbo, 1994a; D’Ottavio and Castagna, 2001a,b;
Krustrup and Bangsbo, 2001; Weston and Brewer,
2002) as well as on empirically defined standards that
were based on our experiences with top-class referees.
An interactive programme (Polar precision 3.0 soft-
ware, Kempele, Finland) was used to identify the
amount of time that was spent in the different activity
categories.
Changes in physical and mental load were evaluated
for 15 min intervals during the first and second half.
Data analysis
The anthropometric data for the referees and assistant
referees were compared using unpaired t-tests. For the
analysis of the physical load during match-play, a two-
way analysis of variance (ANOVA) was performed on
the heart rate data expressed as a percentage of HRmax
using a 2 (group: referees, assistant referees)66
(period: 15 min intervals) design with repeated mea-
sures on the last factor. To further examine the relative
contribution of the different activities performed during
match-play by referees and assistant referees, a two-way
ANOVA was performed using a 2 (group: referees,
assistant referees)65 (activity category: maximal effort,
high-intensity, low-intensity, active recovery, passive
recovery) design with repeated measures on the last
factor.
For the analysis of perceptual-cognitive demands,
the total number of decisions was reduced to the
mean number of decisions per round as given in
Table 3. Because of the non-significant number of
indirect free kicks, we merged this category into the
general free kick group for the statistical analysis.
Also, the number of decisions that were taken in
added time at the end of the first and second half
were not included in this analysis. In line with Nevill
et al. (2002), a chi-squared goodness-of-fit test was
used to determine whether the number of observed
decisions was uniformly distributed throughout the six
15 min intervals.
Statistical significance was set at P 50.05 for all
statistical tests. Fisher’s LSD post-hoc procedure
(P 50.05) was used to further evaluate any significant
effect suggested by the analyses of variance.
Results
Participants
Table 2 presents the mean age, height, body mass and
body mass index (BMI) of the participants. In addition,
some basic heart rate parameters are provided, such as
the heart rate at rest, the maximal heart rate during
match-play and the maximal heart rate during an
incremental exercise test the referees performed in an
exercise laboratory in their home country. Finally,
maximal heart rate is also shown using the ‘2207age
in years’ rule.
Of all the anthropometric measurements, only the
mean difference in standing height approached the
adopted level of significance (P50.06), with referees
tending to be taller than assistant referees. For both
groups of match officials, the mean BMI was less than
the reference value of 25 (Eismann, 1996), although
individual differences were apparent with values ran-
ging from 20.5 to 31.4.
Table 2. Anthropometric characteristics and cardiovascular
responses to matches and laboratory exercise tests of the
referees and assistant referees at the Euro 2000 Championship
(mean+s)
Referees
(n = 17)
Assistant
referees
(n = 17)
Age (years) 40.2+3.9 41.3+2.8
Height (m) 1.82+0.06 1.77+0.08
Body mass (kg) 79.7+9.2 75.3+8.9
BMI (kg × m72
) 24.2+2.6 24.0+1.7
HR at rest (beats × min71
) 52+7 59+6
HRmax (beats × min71
)
match 179+9 181+9
laboratory 172+9 177+9
‘2207age’ 180+4 178+3
184 Helsen and Bultynck

7. For heart rate, there was a significant difference
between both groups for heart rate at rest (P = 0.0054),
referees showing a lower heart rate than assistant
referees. The maximal heart rate (range 164–200
beats × min71
) seemed to be consistently higher during
match observations than during exercise tests in the
laboratory. Although the ‘2207age in years’ rule
approximates, on average, the maximal heart rate, it
does not hold true on an individual basis (i.e. deviations
from 711 to +18 beats × min71
).
Physical load during match-play
Figure 2 provides an overview of the mean physical load
imposed on match officials during match-play. The
overall mean for referees and assistant referees was
85+5% and 77+7% of the individual maximal heart
rate, respectively. A two-way ANOVA yielded a main
effect for group (F1,45 = 16.48, P 50.0002) and period
(F5,225 = 10.41, P50.0001). The group6period inter-
action effect was also significant (F5,225 = 5.45,
P50.0001). Specifically, referees performed at a higher
percentage of their HRmax than assistant referees
throughout the game. This was also the case for mean
heart rate, referees recording higher values (155+16
beats × min71
) than assistant referees (140+14
beats × min71
). For both referees and assistant referees,
the physical load was higher at the beginning of the
first half than at the beginning of the second half.
Finally, from Fig. 3 it becomes clear that the
significant interaction effect was mainly explained by
a significant increase in physical workload for the
referees towards the end of each half, while for the
assistant referees there was a progressive decrease in
heart rate to a baseline activity level that was
maintained throughout the second half.
In Fig. 3, the activities performed by referees and
assistant referees are shown expressed as a percentage
of total match time. A two-way ANOVA yielded a main
effect for activity category (F4,180 = 13.27, P50.0001).
The group6activity category interaction effect was also
significant (F4,180 = 14.91, P50.0001). Specifically,
referees appear to spend most of their efforts in the
maximum effort and high-intensity categories, while
assistant referees perform mostly at high and low
intensity.
Perceptual-cognitive demands
Table 3 provides an overview of the type and number
of observable decisions for all 31 matches during the
final rounds of the Euro 2000 tournament. These
decisions were clearly observable on video replays of
the actual games. The video analysis of the obser-
vable decisions revealed an acceptable intra-observer
reliability of 0.98 and an inter-observer reliability of
0.96.
On average, 137 observable decisions were made per
match (range 104–162). The average number of
decisions a referee had to make by himself was 44.4.
Throw-ins made up the highest proportion of decisions
a referee had to make. During the tournament, the
mean number of direct free kicks was 37.3, while there
were only 0.6 indirect free kicks. However, as off-side is
presented as a separate category, it is not included in
these indirect free kicks. Of all the decisions, 64% were
based on communication with the assistant referees
and/or the fourth official. Finally, the average number
of decisions under the heading ‘others’ was 12.5. In
added time after the second half, twice as many
decisions were taken than in added time after the first
half (2.6 vs 5.2).
Physicalload(%HRmax)
74
76
78
80
82
84
86
88
0-15
Time (min)
15-30 30-45 45-60 60-75 75-90
Fig. 2. Physical load imposed on referees (&) and assistant
referees (.) for 15 min intervals expressed as a percentage of
individual maximal heart rate (mean+sx).
0
5
10
15
20
25
30
35
40
Activity categories
%oftotaltime
ACTIVE
RECOVERY
LOW
INTENSITY
PASSIVE
RECOVERY
MAXIMUM
EFFORT
HIGH
INTENSITY
Fig. 3. Activities performed during match-play by referees
() and assistant referees (.) expressed as a percentage of
total time (mean+sx).
185Demands of top-class refereeing

8. The calculated value of the chi-squared goodness-of-
fit test (0.323) clearly showed that the number of
observed decisions was uniformly distributed through-
out the six 15 min intervals (the critical w2
value
was 11.07).
Discussion
In this study, we used an interdisciplinary approach to
gain a better understanding of both the physical and
perceptual-cognitive demands of top-class soccer refer-
eeing. We chose to examine the match performances of
the UEFA top-class referees and assistant referees
during the final rounds of the Euro 2000 Champion-
ship.
That the overall pattern of activities of a referee varies
in parallel with the players’ actions is clear from the
graphs of the heart rate recordings of the referee and of
one of his assistant referees during the Euro 2000 final
between France and Italy. As a result, it is not
surprising that in our study referees and assistant
referees are subjected to an average physical workload
of approximately 85+5% and 77+7%, respectively.
Reilly (1986) observed English league footballers
during friendly matches and reported that they ex-
ercised at an average of 80% of their HRmax. Van
Meerbeek et al. (1987) found that Belgian university
players exercised at 87 and 84% of their HRmax in the
first and second half, respectively. Recently, Helgerud et
al. (2001) reported an average increase in work intensity
during a football match from 83% to 86% of HRmax
after a training programme that was designed to
improve aerobic endurance. In the literature (Reilly,
1996), the last 15 min of each half have been suggested
to be the most frenetic periods in a game. Obviously,
players do whatever they can to have an impact on the
final outcome. Quite clearly, this final ‘power-play’ also
affects the referees’ efforts. In our study, too, the load
imposed on referees increased significantly towards the
end of each half. All together, referees spend most of
their efforts in the maximum effort and high-intensity
categories, while assistant referees perform mostly at
high and low intensity.
These observations on the total distance covered, the
amount of time spent at high intensity, the number of
changes in activity and the physical workload, reveal
that referees performing at the highest international
standard show work rate profiles that are as high as, if
not higher than, those of players. This may be explained
by the more active behaviour the referee has to show to
implement the rules of play. As D’Ottavio and Castagna
(2001a) noted, players potentially have a much wider
choice of undertaking high-intensity efforts ‘off the ball’
Table 3. Average number and type of observable decisions for all 31 matches during the final rounds of the Euro 2000 tournament
Group Quarter Semi Final
(24 games) (4 games) (2 games) (1 game) Mean Total
Referees
Advantage 4.1 1.8 2.5 1.0 3.6 111
Direct free kick 35.0 40.5 56.0 41.0 37.3 1156
Indirect free kick 0.7 0.3 0.5 0.0 0.6 19
Penalty 0.4 1.0 2.0 0.0 0.5 17
Simulations 2.5 1.3 2.5 2.0 2.4 73
Total 1376
Assistant referees
Corner kick 10.7 8.3 13.5 15.0 10.7 332
Goal 2.8 3.5 1.5 3.0 2.8 86
Goal kick 19.5 21.3 20.0 25.0 19.9 618
Offside 6.5 6.5 7.5 11.0 6.7 208
Team changes 5.1 4.8 6.0 6.0 5.2 160
Throw-in 43.0 35.3 56.0 31.0 42.4 1315
Total 2719
Others
Dropped ball 0.1 0.5 0.5 1.0 0.2 6
Red card 0.3 0.5 0.5 0.0 0.4 10
Yellow card 3.5 4.5 8.5 3.0 3.9 121
Total 136.5 4233
186 Helsen and Bultynck

9. than referees, who have to follow the game quite closely
from beginning to end without any down time. In fact,
to apply the laws of the game, match officials are
expected to follow the game closely irrespective of the
intensity of their previous movements because good
positioning is a pre-requisite for accurate decision
making. Given that the match officials at this standard
of competition are on average about 15–20 years older
than the professional players, they have to attain
excellent age-related fitness to make the appropriate
decisions and to control the game to the very end.
Heart rate clearly seems to be a useful index of the
physical strain imposed on match officials during
games. Heart rate increases to facilitate the transport
of oxygen to the working muscles. Nevertheless, heart
rate can also be affected by stress mediated by the
secretion of the ‘stress hormones’ adrenaline and
noradrenaline. The psychological component of heart
rate is illustrated by the pre-match heart rates of referees
being much higher (approximately 85% HRmax) than
their heart rates before the start of the second half
(approximately 75% HRmax). Evidently, this increase is
not only caused by the pre-match warm-up, as match
officials have to be back in the dressing room
approximately the same amount of time (20 min)
before kick-off as during the break. As Krustrup et al.
(2002) have pointed out, heart rates can be elevated
beyond the normal HR–V˙ O2 relationship, during static
contractions and during exercise with small muscle
groups, under thermal stress and emotional stress.
These findings may also explain why heart rates
observed during matches are generally higher than
during speed endurance or high-intensity training
sessions.
We have found that HRmax is consistently higher
during matches and high-intensity training sessions
than during exercise tests in the laboratory. In line with
Krustup and Bangsbo (2001), we therefore used the
peak values reached in any of the 5 s periods observed
during match performances or training sessions to
determine the maximal heart rate for each and every
match official. It has also been shown that the
‘2207age in years’ rule approximates, on average, the
maximal heart rate, but does not hold true on an
individual basis. Although this rule is still used in many
textbooks and fitness programmes, its use for calculat-
ing work rates and for creating individual target zones
should be questioned, at least for this particular age
group. As also shown by Shephard (1987), the formula
using the age-correction on the estimated maximal
heart rate may exaggerate the ageing effect.
As a wide range of physiological attributes appears to
be required by referees, including high aerobic and
anaerobic capacities to meet the demands of officiating
a match (see Fig. 3), they need to prepare very carefully.
The results of this study support the adoption of
specialized intensive and intermittent training sessions
by referees, which should in the main place priority on
high-intensity aerobic exercise in view of the need to
maintain a high work rate throughout the game. The
anaerobic system, although involved to a lesser extent,
also plays an important role as short and multiple
periods of high-intensity activities are particularly
important throughout the game (Krustrup and Bangs-
bo, 2001). As training programmes have to be designed
in a task-specific way, taking into consideration the
differential demands, these findings are important to
optimize the content of training sessions for both
referees and assistant referees.
In terms of perceptual-cognitive demands, a top
referee makes approximately 137 observable decisions
per match. Besides these decisions, which can easily be
categorized using video replays, there are also ‘non-
observable’ decisions (i.e. when a referee decides not to
interfere with play). It is difficult to quantify these
invisible decisions, but from our discussions with top-
class referees and members of the UEFA referees’
committee, we speculate that non-observable decisions
brings the total number of decisions to approximately
200. Given an effective playing time of approximately
56.3% or 51 min (Miyamura et al., 1997), a top-class
referee takes 3–4 decisions per minute. As there are no
data available for comparison in other team sports, it is
not possible to verify if this workload is similar to, or
higher than, that in other sports. What we do know,
however, is that the number of observed decisions is
uniformly distributed throughout the match. Match
officials have to be completely focused from the start of
a match to the very end. In addition, 64% of all
decisions are based on teamwork between the match
officials. In line with the results of Krustrup et al.
(2002), our data also suggest that separate training
programmes for referees and assistant referees are
required that are in line with the specific requirements
during the game.
Further applications
For the trainability of decision-making skills, video
training offers a unique tool to assist both referees and
assistant referees in controlling the game to the same
high standard but with greater economy of effort.
Indeed, research in industry, traffic and sport has
indicated that, with increasing experience, experts ‘read
the environment/game’ more efficiently, more selec-
tively and more rapidly the more expert they are
(Helsen and Starkes, 1999a). In similar circumstances,
such as police officers making decisions under stresses
of time when they intervene in potentially dangerous
187Demands of top-class refereeing

10. situations, video training has proven to be an appro-
priate way of providing complementary experience ‘off
the field’ (Helsen and Starkes, 1999b). In the same way,
refereeing does have a very definite structure and
cognitive demands, which account in part for what
makes a top-class referee. However, research on
perception and decision making of football referees is
limited (Oudejans et al., 2000). For players, laboratory
settings have predominantly been used in which
simulated game situations have been presented as video
clips (e.g. Helsen and Starkes, 1999a; Williams and
Grant, 1999). To optimize the decision-making skills of
referees, similar set-ups can be used. These are
concerned primarily with the way in which visual
information registration and processing in general,
and perceptual performance in particular, can be
improved. It must be stressed that it is not possible to
elevate a novice referee to the level of an expert simply
by perceptual imitation, or by improving his fitness
alone. Although the availability of specific information
is the same for different skill groups, only experienced
referees can make the important relationship between
informative items (e.g. a simulation or an offence) and
their implications (e.g. giving advantage or a direct free
kick). It is thus important to disclose the relationship
between specific informative items and their sport-
specific significance, which also develops their knowl-
edge structure. Further research must therefore not
only be directed to the disclosure of the information or
knowledge that is used by experts in a specific situation,
but also to the way in which it can be best transmitted.
The findings of the present study suggest that a
dynamic and sport-specific context is needed to
adequately represent both physical and mental match-
play demands. In this respect, Abernethy and Wood
(2001) have recently revealed the limited contribution
of general visual training programmes, which are being
promoted by sports optometrists in the USA and
Canada under the motto ‘Eyerobics’. They stimulate
only the visual information system in non-specific and
often static environmental contexts, which, in our
opinion, cannot lead to more successful sports partici-
pation and performance.
As has been shown by Farrow and Abernethy (2002),
training programmes – including anticipatory tasks and
other audio-visual aids such as video and 16 mm film –
may offer a fruitful alternative to improve the decision-
making skills of referees and assistant referees.
Conclusions
This is the first study to address empirically the physical
and perceptual-cognitive demands of top-class referee-
ing during the final tournament of a European
championship. The present results indicate that match
officials are highly stressed during top matches, both
physically and cognitively.
Acknowledgements
We would like to thank the members of the UEFA referees’
committee for their support in this project. The cooperation of
the Euro 2000 match officials is gratefully acknowledged. We
are also grateful to Bert De Bruyn and Tim Vanmarcke for
assistance with data reduction. Finally, we wish to thank
Matthew Weston and Carlo Castagna for their valuable
comments on earlier drafts of the manuscript.
References
Abernethy, B. and Wood, J.M. (2001). Do generalised visual
training programmes for sport really work? An experi-
mental investigation. Journal of Sports Sciences, 19, 203–
222.
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.
Bangsbo, J. (1994a). The physiology of soccer – with special
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. (1995). Fitness Training in Football: A Scientific
Approach. Bagsvaerd, Denmark: HO+Storm.
Bangsbo, J., Nørregaard, L. and Thorsøe, F. (1991). Activity
profile of competition soccer. International Journal of Sports
Medicine, 13, 125–132.
Baxter-Jones, A.D.G., Helms, P., Maffulli, N., Baines-
Preece, J.C. and Preece, M. (1995). Growth and devel-
opment of male gymnasts, swimmers, soccer and tennis
players: a longitudinal study. Annals of Human Biology, 22,
381–394.
Castagna, C. and D’Ottavio, S. (2001). Effect of maximal
aerobic power on match performance in elite soccer
referees. Journal of Strength and Conditioning Research, 15,
420–425.
Catterall, C., Reilly, T., Atkinson, G. and Coldwells, A.
(1993). Analysis of the work rates and heart rates of
association football referees. British Journal of Sports
Medicine, 27, 193–196.
D’Ottavio, S. and Castagna, C. (2001a). Physiological load
imposed on elite soccer referees during actual match play.
Journal of Sports Medicine and Physical Fitness, 41, 27–32.
D’Ottavio, S. and Castagna, C. (2001b). Analysis of match
activities in elite soccer referees during actual match play.
Journal of Strength and Conditioning Research, 15, 167–171.
Drust, B., Reilly, T. and Rienzi, E. (1998). Analysis of work
rate in soccer. Sports Exercise and Injury, 4, 151–155.
188 Helsen and Bultynck

11. Eismann, H.J. (1996). The 23rd Man: Sports Medical Advice
for Football Referees. Leipzig: Gerso¨ne-Druck.
Farrow, D. and Abernethy, B. (2002). Can anticipatory skills
be learned through implicit video-based perceptual learn-
ing? Journal of Sports Sciences, 20, 471–485.
Harley, R.A., Tozer, K. and Doust, J. (1999). An analysis of
movement patterns and physiological strain in relation to
optimal positioning of association football referees. Journal
of Sports Sciences, 17, 813.
Helgerud, J., Engen, L.C., Wisloff, U. and Hoff, J. (2001).
Aerobic endurance training improves soccer performance.
Medicine and Science in Sports and Exercise, 33, 1925–1931.
Helsen, W.F. and Starkes, J.L. (1999a). A multidimensional
approach to skilled perception and performance in sport.
Applied Cognitive Psychology, 13, 1–27.
Helsen, W.F. and Starkes, J.L. (1999b). A new training
approach to complex decision making for police officers in
potentially dangerous interventions. Journal of Criminal
Justice, 27, 395–410.
Johnston, L. and McNaughton, L. (1994). The physiological
requirements of soccer refereeing. Australian Journal of
Science and Medicine in Sport, 26, 67–72.
Krustrup, P. and Bangsbo, J. (2001). Physiological demands
of top-class soccer refereeing in relation to physical
capacity: effect of intense intermittent exercise training.
Journal of Sports Sciences, 19, 881–891.
Krustrup, P., Mohr, M. and Bangsbo, J. (2002). Activity
profile and physiological demands of top-class soccer
assistant refereeing in relation to training status. Journal
of Sports Sciences, 20, 861–871.
Miyamura, S., Seto, S. and Kobayashi, H. (1997). A time
analysis of men’s and women’s soccer. In Science and
Football III (edited by T. Reilly, J. Bangsbo and M.
Hughes), pp. 251–257. London: E FN Spon.
Nevill, A.M., Atkinson, G., Hughes, M.D. and Cooper, S.
(2002). Statistical methods for analysing discrete and
categorical data recorded in performance analysis. Journal
of Sports Sciences, 20, 829–844.
Oudejans, R.R.D., Verheijen, R., Bakker, F.C. et al. (2000).
Errors in judging ‘offside’ in football. Nature, 404, 33.
Reilly, T. (1986). Fundamental studies on soccer. In
Sportwissenschaft und Sportpraxis, pp. 114–121. Hamburg:
Verlag Ingrid Czwalina.
Reilly, T. (1996). Motion analysis and physiological
demands. In Science and Soccer (edited by T. Reilly),
pp. 65–81. London: E FN Spon.
Reilly, T. (1997). Energetics of high intensity exercise
(soccer) with particular reference to fatigue. Journal of
Sports Sciences, 15, 257–263.
Reilly, T., Bangsbo, J. and Franks, A. (2000). Anthropo-
metric and physiological predispositions for elite soccer.
Journal of Sports Sciences, 18, 669–683.
Shephard, R.J. (1987). Physical Activity and Aging, 2nd edn.
London: Croom Helm.
Van Meerbeek, R., Van Gool, D. and Bollens, J. (1987).
Analysis of the refereeing decisions during the world
soccer championship in 1986 in Mexico. In Science and
Football (edited by T. Reilly, A. Lees, K. Davids and W.J.
Murphy), pp. 377–382. London: E FN Spon.
Weston, M. and Brewer, J. (2002). An investigation into the
physiological demands of soccer refereeing. Journal of
Sports Sciences, 20, 59–60.
Williams, A.M. and Grant, A. (1999). Training perceptual
skill in soccer. Journal of Sport and Exercise Psychology, 30,
194–220.
189Demands of top-class refereeing