While we see every statistics of every player in every match, we don’t talk so much about referees statistics.
It’s mainly because players have much more influence on the result than officials, but like there is an unequal distribution of result according to playing game at home or away – thanks to supporters and some psychological reasons, it would be interesting to see if there are patterns according to who is arbitrate the game.
Here we are looking at referees in Premier League during the first half of 2017-2018 season.
Fouls per game
Firstly, let’s talk about fouls per game, which is maybe the most debated game fact. The graphic below shows the number of fouls whistled per game for each referee.
We can see that there are some differences between referees. While the mean is around 21 fouls per game, it’s interesting to see the two outliers : Anthony Taylor and Andre Marriner.
The first thing to remember is that the two men are among the most experienced referees in the Premier League (Taylor started in 2010 and Marriner in 2005). Nevertheless there is quite difference between the two referees in term of fouls whistled (24 against 17 per game).
What is interesting next, is to see if there is a correlation between the number of fouls whistled per game and the result at the end.
To explain a little bit this graphic : we look at final result for each referee and so project on fouls per game. For example, Kevin Friend result distribution is around 75% of home win and 25% of draw (he has not yet finished a game giving the win to the away team in 9 games), so he counts for two dots in this plot.
Therefore the graph above bring relevant information on the way fouls affect the final decision. More you have fouls in a match more you tend to finish with a draw of a win of the away team. It’s like the number of fouls make up for the home advantage we talked about earlier.
A short brief on final result distribution
Nevertheless, not only fouls counts and a lot of variables influence the final result. If we look at result distribution for each referee, we could see that finally the distribution is quite equivalent.
However, there are some referees that tend to advantage home team like Kevin Friend (9 games), Roger East (11 games) or Mike Dean (30 games).
Yellow (and red) cards
Yellow cards even more further than fouls have a big influence on team and coach strategies. Looking at yellow cards per game for each referee tells us how a official can be strict.
Another time, we can see that there are strong different between some referees (more than double yellow card per game between Michael Oliver and Lee Probert, respectively 30 and 4 games)
To get more deeper in this analysis, this a graph showing yellow cards per game according to number of apparitions.
This surprising trend – more you have apparitions more you give yellow cards each game – can be explained by the experience. In fact, a referee that has more experience and so who is in charge of lot of games knows when he has to sanctioned a fouls by a yellow card. Moreover a referee less experimented may tend to give oral warning instead of yellow cards because he doesn’t have this background, especially in Premier League which is very strict in term of arbitrary qualities and judgement.
This trend is even clearer looking at the whole 2016-2017 season
While it’s not very insightful, red cards chart is only interesting to know that some referees have given twice as much red card as others. But it’s often depend on special event during games.
Notes and perspectives
It’s very surprising to see that there are as much differences between referees. Like Arsenal fan when they see Mike Dean arbitrate Tottenham game, we can ask ourselves how a referee can influence a team during a season. Fortunately, there is a good distribution of “small” and “important” game between each referees, allowing a certain “equality” between teams.
I would be interested to look at other championship and compare each other, and maybe see differences between country in term of severity.
One last word : insights that we have found are to put into perspective because they are only statistics on a small number of games. As we have seen on the graph of yellow cards by apparitions, it will be interesting to analyse the complete data at the end of the season to see if these trends will be confirmed.
Data from WhoScored.com