Tuesday, September 22, 2009

(Extra) Extra Time for United

The Guardian shows off their statistical prowess with on this article describing how ManU gets more added time when they're not winning than when they're winning.

Here's my letter to The Guardian regarding their awesome, yet flawed article:

Dear Guardian Football Section,

Your recent article "Revealed: Manchester United get more injury time when they need it" provides a fascinating first cut at a very interesting problem. Mainly, is there systematic referee bias in favor of the big four clubs. I applaud you data collection effort and sincerely hope that you continue in these types of studies. The lack of parity in football is one of the most pressing issues in the sport today, and systematic referee bias is one potential facet that could need to be addressed.

That said, I think your argument suffers from several serious flaws.

First, consider this alternate explanation: when the big four are not winning, more added time is called for because of the tactics used by their opponent. If MUFC is losing, it's likely that the other team has found a goal somewhere (via United mistake, counter, etc.) and has promptly sat back to defend in the trenches. Obviously, this doesn't occur every time United aren't leading, but this is a frequent occurrence. Often, this "defending" involves constantly putting the ball out of play, rolling on the ground for several minutes after every single challenge, goalies who forget that they need to return the ball to play eventually, etc. The Big Four throw around the term "anti-football" (which honestly usually just comes across as whining). But with regards to extra time, these tactics warrant more extra time.

Second, a relevant question that the article omits is "United get more extra time compared to whom?" Fundamentally, the article is concerned with the ratio of "added time when not winning to added time when winning." But looking at this number for United alone doesn't tell us anything. We need to know what the rate is for every team in order to compare United's number to something relevant. If anything, the numbers for the other Big Four members suggest United gets less added time than their peers. What about home and away splits? Maybe home teams get more added time when not winning in general? In any study like yours, it's important to consider the relevant "counterfactual" to your argument.

Third, your study does not control for important ommitted variables like number of goals scored, number of cards awarded, etc. In econometrics, we call this "ommitted variable bias." Since you aren't considering all of the relevant information, it's possible that you're drawing the wrong conclusion.

Once again, I sincerely applaud your effort. As a suggestion for further research, I think it would be interesting to answer the question "Are referees more likely to award fouls/penalties to the Big Four than to other clubs?" We would need data on the number of challenges initiated against players and the number of times they're awarded a foul/penalty, so that we can tap into the probability that any challenge results in a foul, and then compare this across teams and players.

1 comment:

  1. Did they ever publish this letter? I hope they made their data available somewhere too.

    ReplyDelete