Early this season, I investigated shooting efficiency within the context that not all shots are of the same relative difficulty. However, this analysis suffered from a small sample size, noisy findings and results that had not yet regressed to the mean.
Since the season is now over – a season in which the all-time single-season scoring record was tied – it’s an appropriate time to revisit this analysis with some more meaningful numbers (and a more robust prediction model).
By looking at shot variables such as distance, location, shooting angle and things such as if the shot was the result of a header, a logistic regression is used to estimate the likelihood of each particular shot resulting in a goal.
It is imperfect – as of now, all these models are – in that it fails to control for things such as defensive pressure and game context.
It does, however, do a relatively decent job at predicting goals. For example, for shots estimated to have more than a 50 percent chance of resulting in a goal, 68 percent ended up being goals. For shots estimated to have less than a 50 percent chance of resulting in a goal, only 10 percent resulted in goals. Not bad.
Here are the top 10 over-performing goal-scorers for the 2012 MLS season based on this model:
It shouldn't come as any surprise as to who came on top here. Chris Wondolowski's 27 regular-season goals is one of the league's most impressive feats in recent memory, and he will absolutely run away with this year’s MVP award.
New York’s Kenny Cooper, who finished second to Wondolowski in the Golden Boot race with 18 goals, finishes second here as well. Which would seem to beg the question of Hans Backe, “Why isn’t his name written in the starting lineup in ink?”
Here are the top 10 players in terms of their expected goals for the 2012 MLS season based on this model:
Kei Kamara, like in previous article at the beginning of the season, remains at the top – scoring 8.4 less goals than he would have otherwise been expected to score. Like Wondolowski, he is a savant at putting himself in position to score. Unlike the San Jose star, his technique simply lets him down too often.
And that may very well be the biggest difference between San Jose and Sporting KC, the league’s two best teams this season. The Quakes, led by Wondolowski, usually finish what they start. Sporting simply haven’t.