I asked the nice people at Infostrada Sports Group for a whole slew of data with regards to penalties taken in the Premiership and they were kind enough to provide me with it. They can be found online and on twitter.
Long post this one. Get’s pretty cool by the end though. First up, which teams are at the extremes in terms of receiving and conceding penalties? (There’s a minimum of two seasons data required for each of these categories)
I’d suggest that there aren’t really any surprises here, maybe the sheer volume that Aston Villa have conceded but I don’t think it’s really that surprising. It’s also worth highlighting and adding some context to the difference between United and the rest of the big four. At about 1.4 penalties a season it’s worth roughly one goal. Now that’s nothing to sniff at but at first glance it looks like they gain a large advantage and that’s not really the case.
Ok, next: let’s see if better teams win more penalties. The two plots below are bubble plots, with the size of the bubble representing the number of penalties each team has had (for and against) during it’s time in the Premiership. As usual I’m judging a teams ability based on the proportion of shots it takes during its matches via total shots ratio, an explanation of which can be found here and here.
The centre of this first plot is a bit crowded and unlabelled. What I’ve done with the second plot is adjusted the axes to zoom in, tease these teams apart and label them there.
What do these tell us? Well the correlation between my proxy for possession and the proportion of penalties won is strong, for example the R2 value for the bubbles representing more than 20 penalties is 0.62, so there’s certainly a significant relationship here. The only team that really looks like an outlier is United – in other words their league best penalty differential isn’t because they’re seeing more of the ball in attacking areas than the rest of the big four.
Alright, so we know that better teams get more penalties, but are they better at scoring them? Conventional wisdom says yes, surely the average player from one of the big four teams is better at kicking a ball 12 yards than one from a team at the bottom of the league. Surely. Right?
Well let’s have a look: this plot only includes teams who’ve participated in at least 20 penalties in an attempt to remove some of the potential for noise.
Huh? Well there’s not a clear pattern here – in fact the R2 value is 0.01. In other words the best fit line describes about 10% of the behaviour here, the other 90% is due to random effects. Does the ability to save penalties tell the same story?
Pretty much; once more the R2 value here is 0.01, so we still have 90% random behaviour. Basically these plots suggest that there is no correlation between how good a team is and it’s ability to score, or save, penalties.
One last thing to look at; does a teams ability to score, or save, penalties improve as the number of penalties they take, or concede, increases? Time to break out some funnel plots.
For a better explanation of funnel plots see here (link) but briefly, the black solid line is the league average value and the two dotted lines represent two standard errors above and below the mean.
Nothing here: R2 = 0.02 (87.5% random). Only Chelsea (and to a lesser extent Southampton) may be suggested to have a potential talent for taking penalties. In each case one extra miss would have placed them within the bounds of two standard errors. And how about saving them?
An even worse correlation in this case: R2 = 0.00 (99.8% random). No team sits outside the two standard error limits. There’s no skill here. None.
So what to take from all of this? Regardless of which team is awarded the penalty, regardless of which team is penalised, a penalty is a penalty is a penalty. Our best guess is that it has exactly the league average chance of being scored.
So next time will be the last post on penalties – I’ll be using a very similar analysis to look at players. Alan Shearer was a damn good striker, intuition tells us he must have been better than most at taking kicking a ball 12 yards. Surely. Right?