Introduction
In my last couple of posts, I have tried to create player profiles for passing and shooting. In each case, I have emphasised that this is a work in progress rather than the finished product. Now I have arrived at what I have found to be the most difficult profile: defending. I say this because the statistics associated with defending are fairly vague and not altogether user-friendly, in my opinion. Everyone knows what a completed pass or a shot on target is, but "duels" and other defensive terminology is not as clear. This means that the choice of categories is more subjective for this profile than for the previous ones.Once again, the aim is not to evaluate the success or otherwise of a player's defensive contribution, rather, it is to objectively categorise the attempted contribution.
Factors
Location details are not adequately provided to create a location factor, so the factors used are divided into two, rather than three, categories. These categories are familiar from the first two studies:- Involvement - This has been calculated by taking the total number of duels contested, tackles contested, clearances, blocks, interceptions and recoveries. This is divided by the minutes played and compared to the player's team's per minute average so that players who, for example, defend more than average in a team which does less defending is not penalised for that. As previously, players who fall in the top 20% for this category are described as "focal"; those in the bottom 20% are "peripheral"
- Behaviour - There are three behavioural categories:
- Type of defending - Defending often falls into two categories: "active", typified by tackles and duels, and "passive", typified by interceptions and blocks. To quantify this, the ratio below is used. Again, 20% bands at the top and bottom are chosen, called "passive" and "active" respectively:
(Interceptions + Blocks)/(Tackles + Duels)
- Duel type - Two types of duel are given in the MCFC Analytics dataset: aerial and ground. This factor is the proportion of duels which are ground duels. The top 20% are called "ground" and the bottom 20% are called "aerial"
- Clearance type - This is similar to Duel type, except the types of clearance are "aerial" and "other". The proportion of clearance with are "other" is calculated, with the top 20% being called "other" and the bottom 20%, "aerial"
Results
Again, drawing snappy conclusions from this work is not easy, but here are a few bullet points:- As might be expected, most of the "focal" players are defenders or midfielders. The only exception is Bryan Ruiz
- Strikers/attacking midfielders are the most active players. Defenders/defensive midfielders are the most passive - though Stephen Ireland is an exception to this
- Aerial duelers are most often central defenders or centre forwards. Failing that, certain midfielders are included (e.g. Tim Cahill). It is interesting to me that Bacary Sagna is also included in this category. He won over 2/3 of those duels so it surely isn't because opposition teams see this as a weakness to be exploited
- Almost every player in the "aerial" category for clearances is a striker. This might be expected because, even the most aerially dominant central defender must make ground clearances, whereas strikers often only defend when called back for set-pieces. An interesting exception is Ryan Giggs, who falls into the "aerial" category
Appendix - Presenting the Results
Presenting the results of these profiles has proved difficult, however I am now ready to propose a method. Show below is a plot comparing the defensive profiles of three Liverpool central defenders: Daniel Agger, Jamie Carragher and Martin Skrtel.
It is clear that, despite playing ostensibly the same position for the same team, each player plays it in a different way:
- Carragher's involvement is lower than Skrtel and Agger, whose involvements are similar
- Agger is more passive than Skrtel and Carragher
- Carrager's duel type is more ground-based than Skrtel and Carragher
- Skrtel's clearances are more aerial than Carragher and Agger
No comments:
Post a Comment