Why the media’s new statistical analyses of football is just a trick to stop you noticing what’s going on

 

 

• Yesterday’s game: how Arsenal won, and where the journalists got things wrong
• French authorities issue arrest warrant over awarding of World Cup to Qatar

By Tony Attwood

It is a rather curious thing but of late more and more football commentaries have appeared in newspapers and on the internet with statistics in them.  And not just statistics, but stats of the type we’ve not seen before.

Which is odd, because for years we have been complaining about the lack of statistical evidence produced by and used by the media.  Has our message finally got across?

Well, yes and no.  Because in making this complaint we were particularly thinking about things like the number of tackles and fouls a side made, and the number of times the referee showed a yellow card in relation to tackles and fouls.

But now our measuring system is having to be reworked, for it seems this season around half of the yellow cards shown are not for fouls at all, and so we are having to do some pretty nifty re-thinking.  (Indeed five of Arsenal’s 11 yellow cards have been for non-fouling offences so far this season).  So we are working on a new model.

But at the same time as the “fouls per yellow card” system that we’ve pioneered needs re-working (because about half of the yellows are now not waved for fouls) the outpouring of football statistics elsewhere has become utterly overwhelming, although the media continues to ignore the basic issue of how different referees treat different teams. 

In fact, one might almost be excused for thinking that all these new statistics have been deliberately thrown into the mix by PGMO and their media pals, as they are anxious, as ever, for fans not to start pondering why some referees are so different from others.  

Indeed the new range of statistics is now so vast, that I suspect even the most mathematically educated supporter watching a game can’t take them all in.   Expected goals has been there for a while as a rather crude analysis of how good a player is at putting the ball in the net from a decent position, but now we have…

PPDA which apparently means opposition passes allowed by the other side outside of the pressing team’s own defensive third, divided by the number of defensive actions by the pressing team outside of their own defensive third.

That definition comes from Sky who also are now considering (at least sometimes)…

• Goals prevented.
• Carries (The total number of times a player runs with the ball at their feet for a distance of five meters or more).
• High turnovers (The number of sequences that start in open play and begin 40m or fewer from the opponent’s goal).
• Defenders bypassed.
• Player Pressures (Here I am going to have to quote Sky because still can’t get my head around it: “A defensive event attributed to an individual player. Pressure can be applied directly to the ball carrier (direct pressure) or indirectly to potential receivers (indirect pressure). Pressure is characterised by the defender’s sustained speed, direction of movement and distance from the opponent.”
• Off-the-ball runs
• Sequences,
• Possessions,
• High defensive actions,
• Directness and direct speed,
• Attacks – either build-up attacks or direct attacks,

Now I am sure that there are people in the clubs who take all this, and decide on ways of changing how the team plays; that’s fair enough.   But to start writing articles about how turnovers affect a result while ignoring the much simpler fact that referee David Coote has, this season, called 45% more fouls per game (averaged across every game) than Anthony Taylor, seems really to be looking at the wrong thing.   Turnovers might affect a result, but if they do, they do not affect the result nearly as much as which referee is in charge.

What’s more not only is the number of fouls called by a referee on average per match easy to understand (it is, after all, the number of fouls called by a referee on average per match!) it is also a lot easier to understand than “pressures” or “high turnovers”.   And yet the number of fouls called by a referee per match is utterly ignored while more and more commentators are quoting these new statistics.

But we can’t ignore all this, and so we ask, out of all these new figures that are suddenly becoming available, which ones are going to be important?

I must admit, I don’t know, but by way of exploration take a look at this chart which includes a bit of the data others are now citing, but with which no one is trying to show the actual, (rather than possible or theoretical) significance.

 

Team	Short passes (pos)	Goals	Short passes per goal (pos)	Goal Difference (pos)	Lge pos
Arsenal	549 (3)	11	49.9 (5)	5 (4)	5
Chelsea	601 (2)	5	120.2 (7)	-1 (6)	14
Liverpool	532 (5)	15	35.5 (2)	10 (2)	2
Man C	642 (1)	16	40.1 (4)	13 (1)	1
Man U	437 (7)	7	62.4 (6)	-3 (7)	9
Newcastle	474 (6)	16	29.6 (1)	1 (5)	8
Tottenham	534 (4)	15	35.6 (3)	8 (3)	4

 

This is the sort of table that Untold often produces, but quite often never publishes, simply because quite often the figures show nothing significant or the slightest bit interesting.  

And yet at the moment some of the media seem to be pumping out data by the ton, when much of it tells us nothing other than the statistic itself!

So what is to be made of that table above?  The most obvious thing it tells us is that the number of short passes is pretty irrelevant.  It also shows that a club like Chelsea can indulge in a huge number of short passes and not score goals, while a club like Manchester City can indulge in a huge number of short passes and score goals.  Which suggests the number of short passes is not a key factor.  

If I were paranoid and prone to thinking that Untold’s constant analysis of tackles, fouls and yellow cards has really been noticed by PGMO and their pals in the media, I’d say all this meaningless new data is produced to try and stop anyone taking notice of Untold’s analyses.

Fortunately, my paranoia level remains pretty much average for a middle-class old man living in the East Midlands.  At least, I think it does.