Ahead of the week-long officiating/refereeing/VARing/expected-penalties series (that will hopefully run next week), I thought I’d share part of the process – in this case, looking at who seem to be the most dangerous penalty-box players.
One of the aims is to find out why certain players (and teams) win a lot of penalties and seemingly similar players (in other similar teams) win very few.
There are lots of patterns that suggest different types of biases from officials in the years of data we have studied, that we’ll be reporting on. In order to not jump to conclusions, that meant continuing to refine the search, to add as many potential variables as possible, to see what other factors might be at play; which is why the task is taking longer than planned.
One thing I keep hearing is that a specific team will win penalties as they attack a certain way; but I’m not sure other teams attack so differently.
So yes, Player X wins a lot of penalties, but he is fast and attacks the box a lot. Okay, so what about Player Y, who has virtually the exact same style, but wins none?
As such, I worked with Andrew Beasley and professional high-level analyst and TTT subscriber CVT123 to find various potential ways to address these issues. Do the players who win the most penalties really attack in unique ways to those who do not?
We started with the quantity of penalty box touches. Who won penalties with very few penalty box touches, and who won them with lots? From this, it was possible to find players who won a lot of penalties but at a totally expected rate, based on their heavy box-based activity.
Indeed, within the group of the English players who have won four or more league penalties since 2017, there is an r2 relationship/correlation that suggests a 95% correlation with their number of penalty-box touches (the more touches, the more penalties). But this relationship does not apply for other players, and falls apart spectacularly for foreign players, who seem to need twice as many penalty-box touches to win the same number of penalties.
Even so, we can say that someone like Raheem Sterling, based on this r2 relationship, wins an absolutely fair and expected number of penalties, to lead the way with a hefty 13 won in the four-year period covered. He wins a lot of penalties as he’s fast, direct, skilful, and in the box a hell of a lot. (Indeed, only Mo Salah has touched the ball more times in the opposition box this season.)
But other players win penalties with far fewer penalty box touches (favourably treated), and some get zero penalties for almost as many penalty box touches as Sterling (unfavourably treated). We can’t say if these penalties were deserved or not, just that, based on the law of averages, there are those who win a fair number, those who win too many, and those who win too few.
I will be taking you through all the results next week (all being well, but this is a monster task), but for now, for TTT subscribers only, I thought I’d share the formula we devised to define ‘Penalty-Box Dangerousness’ and also the results, before I apply it to the penalty-winning data, to see if it is indeed just about certain playing attributes (but a quick preliminary glance suggests not).
This was to see if it’s not just the quantity of penalty-box touches – which clearly plays some role – but the nature of each players’ style; the quality of their final-third and penalty-box play.
Using data from fbref.com/en/, Andrew and I chose six metrics that we felt would help cluster a player’s activity into the danger zone. In other words, the number of shots taken outside the box are of no use.
We then applied these six key metrics to three distinct groups: the players who won the most penalties; the players whose standard penalty-box touches suggested they won an excess of penalties; and the players whose standard penalty-box touches suggested they won too few penalties. Note: standard = no analysis of the nature of those touches.
Some players could appear in at least two of the groups: those who had won 4+ penalties, for example, but whose expected penalties were closer to the 13 Sterling had won, or those who had won 4+ penalties but the data suggests they were more likely to have won just one or two.
The aim is to find some kind of connection between how penalties are won/awarded.
The series that will hopefully run next week, as well as video analysis, also includes examinations of the records of all of the referees; things like home and away biases; referees’ birthplace and location-based biases, and the ages of referees; the massive difference it seems to make being an English player versus being from overseas; teams that win too many penalties based on all the metrics, and teams that win too few; and all manner of individual player stats that show who has been fairly treated, harshly treated, and given a free pass to spot-kicks.
While it will be largely Liverpool-centric, it will also have a ton of data and info for other clubs. The data has been analysed in an objective manner, even if the conclusions will be from a Liverpool perspective.
As ever when examining this much data, we had hypotheses that seem confirmed and hypotheses that seem debunked.
There are rival players I felt got too many penalties who, it turns out, seem fairly treated, and rival players who are clearly harshly treated; there are refs I thought always punished Liverpool, but that was likely to due a big-incident bias (that sticks in the mind), and their overall records for the Reds look very fair. But there are also some refs that, if they do Liverpool games, Jürgen Klopp’s men would be better off with a dead sheep rotting on the centre circle.
There will always be some randomness to penalty decisions and sendings off (compared against those who do not get penalties, nor get sent off), but the aim is to try and make sense of game-changing officiating by using big data, and look for gaping holes and anomalies. I’ll also try and explain the different factors as to why certain clear fouls in the box are ignored, as I ponder the possibilities.
The series will start with a hive-mind piece from some of TTT’s top sages about the very notion of fairness in sport, and I’ll also look into some of the history of corruption amongst officials, and whether the data suggests there is anything beyond mere biases. (There is certainly no hard evidence of any conspiracies, but there is plenty of data to suggest favouritism.)
Anyway, for now, this is the list of the players who we rank as the most dangerous penalty-box players, listed on a per-season basis since the start of 2017/18 (followed by some of my thoughts, and an explanation of the methodology).
The top player in the whole study (based on an individual season) just happens to play for Liverpool, and his numbers from 2020/21 are insane. And various other Liverpool players rank highly, although on the 2017-2021 averages, which I’ll be using for the penalty-winning analysis next week, it is a Manchester City player who ranks top. Still, Liverpool clearly have a lot of players who should be winning a lot of penalties, and yet don’t.
This article is for subscribers only.