Occasionally I listen to football pundits talk. I am naturally skeptical about everything I hear or read (I’d be terrible at my job otherwise) but sometimes the suggestions are so laughable that you’d struggle to believe their veracity even if you wanted to. Exhibit a:
|“Ultimately you’ve got to win matches…I saw them [Southampton] play West Brom, not one player got a yellow card in that game. You know that was a game they had to get something out of, from my point of view.”
“You know, if I was a manager you’d want to know that your players cared so much. Getting a yellow card, you can get a yellow card for anything, but, ultimately, get stuck in, show some urgency”
Graeme Le Saux, Five Live Football Daily, 10Nov12
Were I a manager then my personal preference would be that my players were doing something that would help my team win a match. Suggesting that picking up a frivolous booking would help achieve this strikes me as both superfluous and hilarious. Thankfully it’s something I can test relatively easily.
Firstly, as usual, lets establish the whether there’s a significant sample size. On record I have 4658 games and 2 teams play each game, giving a total of 9316 team games. How are they split in terms of games where teams receive at least one card compared to those where they received none?
So we’ve got a pretty decent number to play with in each category. Lets work with this split and see how these two sets of teams perform in terms of their average points per game:
Rationally I’d expected a split – I think a better team is likely to be better in all facets of the game, including committing less fouls and picking up fewer bookings as a result. That being said the gap a pretty significant one. To me the above deems Le Saux’s argument as an asinine one.
I actually think this can be taken a step further. Does the number of excess cards one team gets compared to the other have an effect on the number of points scored?
There’s a pretty clear trend here, the more yellow cards a team gets, the worse it performs. This, to me, is kind of interesting. The first ‘excess’ card costs a team ~0.21 points, the second ~0.14, the third ~0.16, and the fourth ~0.02. In other words, the further from evens in terms of the number of cards handed out, the smaller the effect on the outcome.
Finally lets compare the number of points they score compared to average (~1.37). This time I’m only taking the groups with a significant sample size (more than one season, or 76, team games).
I think there’s a couple of steps further to go with this. Watch this space.