Are referees biased? Further strangely suggestive figures….

By Tony Attwood

This article continues from the articles two days ago relating to how many cards each team gets.

This table below shows the number of cards per game involving different clubs.  The club with the most cards per game is Arsenal with 4.90 cards on average.  The lowest is Manchester City with 2.40 cards per game.  That is quite a big difference.

So with matches involving Arsenal there are more than twice as many cards per game as there are with matches involving Manchester City.  And not far off twice of the number of cards as with games involving Burnley.

And remember these are not just cards against Arsenal but cards in total in matches involving Arsenal.

And just out of interest in the list below in order of clubs with cards against them, I have listed the London teams and Watford in brown, and we see that in a cards per game analysis the top four places are given to London and home counties teams.

So I wondered, given that the clubs that are getting the most cards in their games are from the Home Counties, where do referees come from.  The referee data sheet I used did not have everyone’s home area, but since the two refs without a home (as it were) handled just six games between them we can leave them out for the moment.

Now there is a thing.  The teams with the most cards against them are generally from London and the south.   But the number of referees from London and the south are, well, Tim Robinson from West Sussex (1 game) and Simon Hooper from Wiltshire with six.

We are however looking at six teams from London and the home counties, and we find that four of the teams occupy the top four positions.  In a normal situation these London and home counties team might expect to be spread out across the league in terms of cards, but four of the six occupy the top four places.

Of course it is easy for supporters of clubs from outside the region to say that London clubs tend to be cheating bastards who need to be carded – and yet you might expect such clubs to be aware of this, and so restrain their players.  Yet they are still being carded – which is odd.

Now when we combine this with the data from our articles earlier in the week we know that, for example, London based Arsenal were at that moment 13th in terms of the number of fouls committed by Premier League clubs with 211, compared to Leicester’s 198.  But Arsenal have the most cards with 52 against Leicester’s 20.   So these non-London referees are in this case two and a half times less likely to penalise Leicester for a foul with a yellow card than they are London based Arsenal.

Obviously we do not have videos of non-London referees being handed payments by non-London clubs to prove corruption, and of course if rewards are being given they will be made very secretively.   And of course we do not have definitive evidence that anything is amiss.  Rather, in each analysis what we have are a very curious set of statistics – here a set which show London clubs being penalised with more yellow cards by non-London refs, even when the number of fouls is similar.

There have been comments of course, during this series, suggesting that we have proven nothing, and I’d agree with that.  What I am trying to say (obviously not very clearly given some of the comments) is that if something is amiss with refereeing by the PGMO, it is going to be well hidden, so all we can do is look for unusual patterns.  None of the unusual patterns is of itself going to be proof that anything is wrong, but a collection of unusual patterns would be suggestive that something might be wrong.  If one finds enough unusual patterns then the chances that there is something odd happening are increased.

To give a simplistic example, if referee X wants to ensure a victory for club Y and club Y commands the game of its own volition and is sailing to an easy win then not only will referee X not have to bias the game for club Y he might, in order to balance his own figures give a few more dubious opinions in favour of club Z, since they won’t affect the result.

This means we have to look and see if there are any trends which creep through, despite such possibilities.  Those who have chosen to disbelieve there is anything wrong with PGMO, are of course fully entitled to their own opinion.  The difference between us I suppose is that I am looking at certain oddities and wondering whether they really are likely to be happening by chance, or whether they are the by-product of bias.

And as always wondering why PGMO choose to have such a small number of referees, with such a geographic bias, and with their regulations (not replicated in some other countries) about no interviews with referees.