The man in the yellow shirt

18-10-2012 14:57


Database creation in football is mostly use for analyze player performance, but players are not the only ones (factors) involved in the game; like we see before, environment is also very important (Formation use by team; Home vs Away games etc…) one of the most important environment factors is the referees and usually we forget to analyze their performance. Of course the analysis of a referee performance will always be subjective according to different referees; but can we determine a decision-making tendency to take advantage and avoid some embarrassing situation? There is any advantage doing the analysis of referees? Table 1 shows us all referees from premiership season 2011/2012. Howard Webb; Martin Atkinson and Mike Dean refereed each 5 games of Manchester city team. Lee Mason; Mark Clattenburg and Phil Dowd refereed 4 games each. Total of game refereed 27 from 38 games (71%); so I think is important to know how will referee the game.

Table 1

Recoveries;Fouls conceded and Cards frequency for each principal referee

For this point I think the number yellow and red card per Tackles / Recoveries (interaction between Manchester City vs opponent team) provide a ratio important to analyses performance for each referee (Table 2). Howard Webb refereed 5 game; in this game Manchester City made 445 recovery; 11% from this recovery are fouls (51)1, 8 yellow card and no red card. For Martin Atkinson we have 961 recoveries made; 127 fouls; 20 yellow cards and 2 red cards in 5 games and for Mike Dean we have 891 recoveries, 81 fouls, 8 yellow cards and no red card.

Table 2

If we analyze carefully Howard data we can see:

·         A low number of recoveries made by Manchester City compared to other games (445).

·         Number of fouls is on average (11%)

·         High number of yellow card (16%)

If we analyze carefully Atkinson data we can see:

·         A high number of recoveries made by Manchester City (961).

·         High number of fouls (127/16%)

·         High number of yellow and red card (16%; 2%)

If we analyze carefully Mike Dean data we can see:

·         A low number of fouls (9%) compared to average.

·         Yellow card is on average (12%)


The same way we analyzed a player in a particular environment, I will do the same to referees. Howard Webb refereed 5 games with 445 recoveries, which mean less interaction between the two teams and less probability to fouls be made. The low number of recovery may be explained by characteristics of the game (A high percentage of ball possession by Manchester city means, they doesn´t need to recovery so many times the ball); the number of fouls are general average (11%) so having less interaction between teams, the number of fouls is proportional. Howard Webb gave a high number of yellow cards (16%) this may be a tendency in way the game are refereed, not essentially to prejudice Manchester city but more like a “style”.

Sometimes the referee give easily a yellow card in the beginning of a game to try control the players avoiding possible controversy decisions in the rest of match; or give yellow card´s consecutively to tries calm players at aggressive moment of the game, that may be the explanation for high number of yellow cards. For confirm that, I tried find when the yellows card´s are given: Man. City Vs Everton; yellow card 20´; 29´; 25´; 31´ Newcastle Vs Man. City; yellow card 26´; 30´; 32;´36´;38´.

Martin Artkinson referee also 5 games with 961 recoveries, a very high number if we compare to Howard Webb games, which mean many interaction between the two teams and more probability to fouls be made, confirmed with the number of fouls (127/16%), with also a high number of card´s (yellow 16%; red 2%). Can we say Martin Atkinson prejudice Manchester city? Or the “style” of referee is very strict?

Actually looking only for numbers leads us to errors, if we consider the high number of recovery we can say Manchester City fight for ball possession so is probable that deserve the cards; but why? If we look the opponent team refereed by Howard Webb (Everton; Newcastle; Norwich; Stoke City; Tottenham) and for Martin Atkinson (Liverpool; Arsenal; Wigan Athletic Queens; Park Rangers; Wigan Athletic) we can say the difficulties of the games are bigger for Martin Atkinson because he referee two game with team how fight for the title (Liverpool and Arsenal) and the game can be more aggressive giving more cards.

Mike Dean refereed also 5 games with 891 recoveries but with a low number of fouls (9%) compared to average, yellow card is on average (12%), so what can we say? If we compare to Martin Atkinson game we can see a number of recovery similar, but the number of fouls is very different and Chelsea are the only opponent team how fight for the title, so Mike Dean refereed only one difficult game; maybe Mike Dean have a style less strict and leave run the game (Game more fluid). So I think it is possible to use the database to analyze referees, and find tendencies to utilize them to our benefit.

1 When I said fouls, I mean not really fouls in each recoveries, but for X recovery (interaction between teams) we have Y fouls.