Premier League 2011-12: Position Analysis SM/WA

Wide midfielders/attackers. This time reviewing the Opta data for those players who started a match on either side of a ‘flat’ midfield 4 or 5 or an advanced 3. As with my earlier post on CMs, I have to admit from the start that the roles these players have can differ widely from team to team – some teams who play a quartet in midfield see the need for defensive-oriented side midfielder(s), whilst in other cases a player may be given a ‘free’ role and won’t actually spend as much time on the flank. As a result, because my system firstly groups the players then looks at the strongest correlations between those player’s actions and winning, the more influential midfielders’ statistics give a sort of bias to the results that mean some midfielders with rarer abilities may not be noticed. This isn’t a defence of Gabriel Obertan, who as we will see later finished at the bottom of the rating, but rather a disclaimer that Obertan might be superb in some areas that:

a) are not covered by the statistical fields Opta have provided or

b) my analysis doesn’t pick them up because it focuses on attributes dominated by players at the top of the rating

Or perhaps Obertan just isn’t very good. Note b above is particularly interesting because at times during my position analysis reviews certain players have been surprising underperformers. I am always sensitive to the results of Newcastle, but it has been rather surprising that for a team which finished 5th in the league so far only 1 player has made the top 10 of any position rating so far (Yohan Cabaye, CM). Newcastle’s uncanny ability last season to score goals despite a low final third pass success rating – as highlighted by Ravi Ramineni’s post here, suggests that Alan Pardew’s team have a unique but effective style that may not be comparable to other teams in the EPL – or maybe Newcastle had an incredibly ‘lucky’ season, as suggested by Mark Taylor.

So, with that kind of disclaimer noted, back to the analysis. This time I introduced some new measures to try to avoid bias for certain players or certain teams since goals, chances created and pass success in opponent’s half were all very significant contributors to performance here – but frankly comparing those statistics directly between a Wolves player and a Man City player would unfairly bias the City player. The measures used are noted below:

Goals rating: goals per big chance per minute (rewarding clinicism, goals out of nothing and profligacy)
Chances created per min: key passes + assists per min
Pass success opp half: as a ratio of unsuccessful passes
Long pass success: as a ratio of unsuccessful long passes
Dribble success: as a ratio of unsuccessful dribbles
Lay-offs success: as a ratio of unsuccessful lay-offs
Through ball per touch
Touches per min
Ground duel ratio
Recoveries per min
Team bonus: 1-20 additional points based on the player’s team

I weighted these factors according to their relative effects on winning/producing goals/conceding goals – then added a bit of a lazy attempt to reward players who play for losing teams by adding a ‘team bonus’.

46 players played more than 1000mins having started in the position of wide midfield or attacker, notable absentees are numerous but include Sessegnon, Modric, Arshavin, Park, Milner, Drenthe, Walters, Adam Johnson, Ramires, Bellamy, Ben Arfa. Some of these players came agonisingly close to qualifiying but just didn’t quite make it:

And the winner is Juan Mata, leading a Spanish top 2 with Silva in second place. Manchester United have an impressive three players in the top 10: Young, Valencia and Nani, whilst QPR’s Adel Taarabt makes 6th place – very impressive considering that without the team bonus he’d still be in the top 10.

Mata leads his peers on chances created, and otherwise scores very well in the key offensive statistics. In dribble success, a quality that doesn’t seem to matter that much but is certainly fun to watch, Nani leads a group of only 8 players who made more successful dribbles than failures (2.3, head and shoulders above the rest). The rest of the group, in order of success, were Larsson, Mata, Pennant, Bennett, Dyer, Valencia and Silva (ranging from a ratio of 1.5 to 1.0).

Arsenal fans may be surprised to see such poor scores from Gervinho and Walcott (or maybe not!). Gervinho may be forgiven considering that it was his first season in England but Walcott’s position just ahead of Obertan is a disappointment despite the fact that he scored 8 and made 8 assists from games in which he started. It may well be that Arsenal’s attack tend to focus through the centre – and as such Gervinho and Walcott do not get on the ball as much (backed up by their relatively low touches per min stats). In any case, this season Wenger has significantly strengthened his attacking midfield with Podolski and Cazorla – whilst Walcott’s contract is due to expire at the end of the season. If Walcott cannot force himself into a central striker’s role at Arsenal then a move away from the club at the end of the season could even be mutually beneficial.

Lastly I have provided a view of the only 2 defensive measures I included to give a different picture of how useful this same group of players are defensively:

Gutierrez leaps up this table, highlighting his defensive contribution to Newcastle’s performances last season, whilst Pienaar, Young, Dyer and Nani make the top 10 in both tables – proving their admirable all-round capabilities. Jermaine Pennant is 2nd, strongly outperforming his Stoke counterpart Matthew Etherington.

Advertisements

One thought on “Premier League 2011-12: Position Analysis SM/WA

  1. Pingback: MCFC Analytics – Summary of blog posts #5 « Analyse Football

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s