The aims of the present study were (1) to analyse the physical demands of top-class referees and (2) to compare their official FIFA fitness test results with physical performance during a match. The work rate profiles of 11 international referees were assessed during 12 competitive matches at the 2003 FIFA Under-17 World Cup and then analysed using a bi-dimensional photogrammetric video analysis system based on direct lineal transformation (DLT) algorithms. In the first 15 min of matches, the referees were more active, performing more high-intensity exercise (P < 0.01) than in the first 15 min of the second half. During the second half of matches, the referees covered a shorter distance (P < 0.01), spent more time standing still (P < 0.05), and covered less ground cruising (P < 0.05), sprinting (P < 0.05), and moving backwards (P < 0.001) than in the first half. Also in the second 45 min, the distance of referees from infringements increased (P < 0.05) in the left attacking zone of the filed. There was also a decrease (P < 0.05) in performance in the period following the most high-intensity activity, compared with the mean for the 90 min. Time spent performing high-intensity activities during a match was not related to performance in the 12-min run (r(2) = 0.30; P < 0.05), the 200-m sprint (r(2) = 0.05; P < 0.05), or the 50-m sprint (r(2) = 0.001; P < 0.05). The results of this study show that: (1) top-class referees experienced fatigue at different stages of the match, and (2) the typical field tests used by FIFA (two 50-m and 200-m sprints, followed by a 12-min run) are not correlated with match activities.

The first objective of the present study was to examine the accuracy of offside judgements of assistant referees during the 2006 World Cup and to compare overall results with performances of assistant referees during the 2002 World Cup. Second, we also examined underlying mechanisms leading to incorrect decisions. According to the definition used for the 2002 World Cup, the results revealed a decrease in the number of incorrect flag signals during the 2006 (24/240 or 10.0%) compared with the 2002 World Cup (58/222 or 26.1%) (P < 0.001). For accuracy of all potential offside situations, according to a new definition (with the attacker 2 m in front or behind the second-last defender), the error percentage was 7.6% (17 flag errors and 9 non-flag errors out of 342 situations). Overall, quality of offside decisions was not influenced by the position of the assistant referee relative to the offside line. In addition, flag errors can best be explained by the perceptual illusion induced by the flash-lag effect.