Tuesday, January 10, 2012

Transfer Window Benchmarking II: Defensive Performance in the Premier League

As a follow-up to yesterday's post about offensive performance in the Premiership so far this season, here's a simple graph summarizing clubs' defensive performance this season. Naturally, it's limited in what it can tell you - all stats are - but the picture does provide some interesting pieces of information. On the vertical axis, we see the average number of shots conceded per match; the horizontal axis allows us to locate clubs' defensive goal to shot ratio (the ratio of goals to shots conceded) per match.

As before, recall that the different quadrants reflect different performance profiles. Clubs in the upper right hand corner both conceded lots of shots and allowed their opponents to convert them to goals very efficiently. In contrast, teams in the lower left hand corner didn't allow too many shots and prevented their opponents from scoring efficiently (the green lines denote the league medians - 50% above and below). So, on defensive performance, we want our club to be in the lower left hand corner and certainly not in the upper right hand one.


So what do we see? As on offensive performance, Manchester City are this season's standouts so far. They have allowed the fewest shots and typically haven't allowed their opponents to score off the chances they did create. While City are in a league of their own, Liverpool, too, has had a very strong defensive first half of the season.

At the other end of the table, we see that Blackburn, Fulham, QPR, Wolves, Norwich, and Bolton have allowed similar numbers of shots per average match (between 15 and 16), but Blackburn (as well as Wigan) have been particularly permissive when it comes to being scored on. Almost 1 of every 6 shots taken on the Rovers' and Latics' goal found the net; in contrast, only one in 12.5 shots crossed the line for Manchester City, Tottenham, Manchester United, and - surprise - Swansea. Clearly, the Swans' defense (and their terrific goalkeeper Michel Vorm) have contributed greatly to the club's league position so far this year.

I'll mention two other clubs: first, Everton. The Blues are an outlier in the lower right hand corner, suggesting that they have managed to prevent their opponents from generating too many chances. In fact, Everton lead the league in allowing the fewest number of shots on goal (at 9.6). But curiously, the chances opponents did create stood a better than average chance of going in (pun intended). On this dimension of defensive performance, Tim Howard & Co. are 4th from bottom in the league. The other team that's worth taking a look at is Sunderland - as on offensive performance, the Black Cats haven't been an outlier at all. They find themselves right smack in the middle of the pack.

So, what to do in the transfer market? Perhaps one way to look at the information above is to consider what clubs should not do. Consider Blackburn. Given their performance so far, one thing they should not do is sell Chris Samba - unless they have resigned themselves to playing in the Championship next year. If not, they would do well to consider buying other help in the back, or perhaps in goal (my sense is that Paul Robinson hasn't been playing as well as Rovers need him to). Wigan, too, need all the help they can get, both in preventing opponents' shot creation - so perhaps help in defensive midfield - and the back.

City look like they don't need any help - they look outstanding, statistically speaking, and should continue to do what they've been doing - until you consider that they won't have the Toure brothers and Vincent Kompany available for some time. This may tempt them to spend some of the money a Tevez sale may generate to sign defensive reinforcements.

What about the rest of the clubs? They implications of the data are less clear here, if you ask me. Some of the performance may be due to tactics or coaching, rather than ability, so I would be more hesitant to suggest buying new players to deal with a bad defensive goal to shot ratio (though good goalkeeping has never hurt anyone). And if you are a selling club, make sure you ask for a premium if you're selling to the teams in the upper right quadrant.