My previous post on substitutions at UEFA Euro 2012 drew a comment on Soccermetrics Research’s Facebook page that asked if I could control for the substitute’s position in my analysis. I was curious about the results myself, so here they are.Below is a table of the default positions of players who were substituted in and out during the Euros. It’s true that a player can be brought into the match and asked to play a more defensive role in midfield or drop deeper up front, but I think the default positions are useful to get an indication of the manager’s mindset as he made the substitution.
In general, substitutions tended to be like-for-like. Of the four basic field positions in soccer, the midfielders were replaced more than any other position at the Euros. Given the physical demands on these players, I can’t be surprised at that. Roughly the same number of midfielders and forwards were subbed into Euro 2012 matches (80 midfielders vs. 78 strikers), and a slightly higher proportion of midfielders were replaced by strikers than strikers by midfielders (34.7% vs. 31.3%). Defenders in the Euro were replaced at a much lower rate than midfielders or strikers, and they were equally as likely to replace a midfielder or a striker as they were to be replaced by those two positions.
So who were the 11 substitute goalscorers in the Euros? The table below displays the positions of the players who scored and the positions of the players that they replaced.
Most of the substitutes who scored goals were strikers by default, on raw numbers and on proportion relative to the number of strikers subbed into the match. Again, that’s not much of a surprise and is actually quite desired. I did find it interesting that no substitute defender scored a goal at the Euros.
Even if a substitute doesn’t score, it’s possible to contribute in other ways toward the final scoreline. Most substitutions fail to move the needle, and the Euros were no different: 118 substitutions did not change the scoreline. The remaining substitutions affected the scoreline almost evenly. Twenty-eight substitutions were on the field when their team registered a positive goal differential relative to the scoreline at the time of the substitution (one goalkeeper, three defenders, 13 midfielders, and 11 strikers). Twenty-seven resulted in a negative goal difference (two defenders, 13 midfielders, and 12 strikers).
(I have the list of substitute players who were present when the scoreline changed. If there’s sufficient interest I’ll post it.)
So with the goal differentials one can start to form some very rough outlines of a plus/minus rating system, and my internal FMRD library is at the point where I can start to carry out such an analysis. I hope to give more details on my contributions in the near future.