Statsfindings

Whenever I play Fifa I get very frustrated. I always find myself in good positions, such as in the opposition penalty area. However, sometimes you just want to smash your controller, since it looks obvious you are about to score and have your own ‘Agueroooooooo’ moment, only for a defender to come out of nowhere and tackle you, or block your shot. This fascinated me to see whether or not you could learn anything from the proportion of penalty box touches that result in shots, as this could indicate which players are (more) able to make space in the box for a shot, as opposed to being tackled or having a shot blocked. I originally submitted this as a proposal for the OptaPro forum, but since it didn’t get selected I’m writing about it here. This is just a little bit of fun research to follow up on my thoughts, not me claiming that I’ve ‘found the solution to football’.

As always, collecting the data was rather tedious. Since data on touches in the box is rather difficult to find if you don’t have easy access to opta data, I only have data on touches and shots in the box for select players this season, and I cannot compare to past seasons. This is quite a nuisance since I can’t compare between different seasons. The expected goals data is from Paul Riley’s shot map (https://twitter.com/footballfactman) and https://twitter.com/ChallengersPod.

To summarize, A players percentage of penalty box touches resulting in shots correlates strongly with their expected goals total. However, there’s little pattern between a team’s XG, or with shots and shots on target.

On the other hand, this is what a team’s percentage looks like compared to their expected goals:

As you can see there is no pattern here, unlike the previous graph. Sbt% stands for ‘shots (per) box touch percentage’. Let me know if you think of a better name! I’ll just refer to this as ‘the percentage’ to make things easier.

The main discrepancy is that strikers obviously have a lot more involvement in the penalty area, so the percentage can be very applicable to them, whilst midfielders or attacking midfielders are less involved ‘in and around’ the box. This weighs against them, and to a bigger extent it makes the percentage not very useful for comparing teams. When a team has the ball in the penalty box, it may often be in not so dangerous areas, like the wider areas, or there are so many defenders in the box it’s difficult to shoot. This massively skews the percentage as there will be so many more touches in the box for a team than shots, so there is a much smaller range of percentages for teams than players. Furthermore, these shots are likely to be taken by the strikers, further boosting their percentage.

To demonstrate, here is the graph for players again.

The lowest point and the point furthest to the left, which don’t match the trend so much, represent Sigurdsson and Pogba, with Sigurdsson the furthest to the left. These are both midfielders who aren’t always so advanced or in teams who aren’t doing a lot of attacking, and hence their percentage doesn’t tell us much. Likewise, Philipe Coutinho only has a percentage of 17%, lower than both Andre Gray and Alvaro Negredo.

Another thing which seems to stick out to me is the fact that it may favour ‘target men’ more, as they do the majority of their attacking in the penalty area. For example, Benteke and Rondon both have percentages of 44% and 47%. Crystal Palace have put in the third highest amount of crosses this season, so Benteke could easily be on the receiving end of a lot of these crosses to benefit his percentage.

Whilst the percentage may have some problems, it obviously has value if it players with a high percentage have a high expected goals total. It just needs to be used in the right context. Ibrahimovic, Aguero and Firmino have very high percentages which is good to know, whilst almost half of Aguero’s touches in the box result in a shot. If some of the Premier League’s best players are performing well in this measure, it is obviously useful. This could definitely be another tool used to compare strikers.