By Tony Attwood
PGMO – the organisation that runs refereeing in the Premier League, has generally stuck to its view that refereeing accuracy is around 98% to 99%.
For example on 13 August 2013 the Daily Mail ran a piece on referees which says, “According to the Premier League, 99 percent of offside decisions are correct, for the third season running. They also claim that 98.6 percent of decisions made in the penalty area are accurate.
“The assistant referee makes on average 50 decisions each game; 45 of these are pure offside judgments, with four of these resulting in offside flags. Their accuracy? Again, a staggering 98 percent.”
This figure became a little challenging when on 8 January 2018 the Guardian announced that Mike Riley had said that while Video Assistant Referees are not intended to make refereeing decisions 100% accurate, the game’s newest technological innovation will be viewed positively if it leads to a 2% reduction in errors by officials.”
Which would have made referees between 100% and 100.6% accurate.
This is not only impossible but contradicts the evidence. Ignacio Palacios-Huerta, was on the board of Athletic Bilbao for seven years and is professor of management, economics and strategy at the London School of Economics, and he wrote the report in the Review of Economics and Statistics (Volume 87 | Issue 2 | May 2005 p.208-216 if you are currently doing a PhD and want to look it up) into Favouritism under Social Pressure.
He started by looking at what happened in time added on in La Liga matches and found that when a home team was ahead by a single goal, the referee allowed almost 30% less additional time than average.
However, if the home team was behind by a single goal the referee added on 35% more time than average. And there was more, for the larger the crowd, the more biased the referee.
Furthermore, when the away team scored in time added on then time was added onto the time added on – but even this was variable. Referees were shown to end the game more quickly if the home team scored – responding to the whistling of the crowd.
The academic study then goes on to see which two teams benefitted from this refereeing bias. They were… Real Madrid and Barcelona and Real Madrid.
There was another study of this type conducted in the 2006/7 season in Serie A wherein various clubs had to play in empty stadia following serious crowd incidents. Here the significant crowd influence on the referee was removed and the home advantage with fouls, yellow cards and red cards awarded against the away side all reduced significantly.
Thus the argument made by Brighton directors recently that playing games at neutral stadia with no crowd would be unfair because smaller clubs benefit from their home advantage. That they do – but not necessarily for the reason that Brighton said. It is not because of the effect of the small ground and crowd noise on the away players, but its effect upon referees who can be influenced by the crowd.
The Guardian is now reporting on research in which 40 qualified referees considered 47 incidents from the Liverpool v Leicester game. Half watched with crowd noise, half without any crowd noise. The referees viewing the game with the crowd noise awarded more than 15% extra fouls committed against the home team compared with those watching in silence.
There is some evidence that video evidence reduces the effect of referee bias – American research shows a 4% decline in the number of home wins in matches following the introduction of video research and an 8% decline over three seasons.
According to the Guardian, in Europe’s top five leagues, “home advantage has fallen from 49% to around 45% – probably because of a combination of better referees, video replays and less hostile crowds.”
The problem for the PGMO is that according to their figures there was nothing wrong to begin with, since they were between 98% and 100.5% accurate, making improvement impossible.
But there is worrying information here. The level of hostility of the crowd, and indeed the tightness of the arena is shown to affect refereeing accuracy – whether or not the referee is inclined towards one club or not. However not all grounds will be the same. The rampant negativity towards Arsenal teams shown at the Arsenal stadium in recent years with placards, booing etc will have removed the home advantage in terms of referee bias for the home team, that Arsenal might otherwise have got.
Even if there is no bias against Arsenal, as some have argued (despite all the evidence from our 160 games analysis), there is a home / away bias caused by crowds. Except Arsenal are probably not benefitting much at home but getting the negative effect away..
Anyone who has watched Leicester v Arsenal in recent years for example will know just how vociferous a crowd can get and how that crowd can influence a referee. We’ve felt it being there. Now we have the stats to prove it.
The fact that PGMO stats ignore this factor shows how right we have been to dismiss them as a set of made-up figures of no relevance to the debate. Indeed these figures remove any vestiges of credibility that the PGMO might once have had.
- Arsenal v Wolverhampton: the club that gets cards at over twice the rate of Arsenal
- Arsenal v Wolverhampton Wanderers: where will each team finish?
- Arsenal v Lens: what we found, what we felt, what they did
- Arsenal v Lens: the team, the home/away form and the strange coincidences
- Arsenal v Lens: they had a poor start but are now flying