The Washington Post

Ahead of the World Cup, here’s why you should be skeptical of goals scored statistics

(AP Photo/Marcio Jose Sanchez)

Chris Wondolowski just made his first World Cup squad. When Clint Dempsey picked up a slight groin injury, he was pressed into service last night for the friendly with Azerbaijan. Wondolowski had a huge opportunity in the fourth minute with a free header from the danger zone, but he knocked it down into the keeper’s knee and the chance was lost.

Finishing has been a problem for Wondolowski recently. He has not had a terribly impressive scoring season so far with the San Jose Earthquakes, having notched four non-penalty goals in 810 minutes played. He was likewise disappointing in 2013, when he scored just 10 non-penalty goals in over 2600 minutes. Is that a national team striker, even a backup? Well, in 2012 when the Earthquakes won the Supporters’ Shield, their center forward was a dominant force. He knocked home 22 NPG in barely any more minutes than it took him to score 10 in 2013. What went wrong after 2012? Should U.S. fans be hopeful that Wondolowski’s decline can be reversed?

There are a variety of possible explanations. Perhaps it was age-related decline. Perhaps Wondolowski had problems in the service he received, or with the teammates and tactics surrounding him. Statistical analysis offers another possibility. Perhaps it was random variation.

Fancy stats for soccer offer an important insight into goal scoring. The rate at which an individual player finishes his shots rarely remains constant. Year to year, there is very little correlation in a player’s shooting percentage. To adjust for the quality of the chances a player gets, I am using the expected goals statistic that I introduced in my article on D.C. United. These are Wondolowski’s goals and expected goals per 90 minutes in 2012 and 2013.


Wondolowski finished his chances at an exceptionally high rate in 2012, and he was much less efficient in 2013. We can call the rate at which a player finishes his chances xG+. It functions like ERA+ or OPS+ in baseball. If a player scores more goals than expected based on his attempts, his xG+ will be above 100. Below, under 100. So Wondolowski would have an xG+ of 146 in 2012 and then 82 in 2013. If we look at all MLS players with at least 20 shots taken in both 2012 and 2013, we see likewise a ton of randomness and very little year-to-year correlation in their xG+.


There is next to no correlation. Now, this is not to say that shooting skill does not exist or that the rate at which soccer players finish their shots is just random. That flies in the face of experience and common sense. And indeed, over much larger samples, some signal can be found in finishing rates.

However, what this means is that we should be highly skeptical of very bad or very good xG+ seasons. As Alex Olshansky of Tempo Free Soccer recently pointed out, Joao Plata for Real Salt Lake has been on a crazy finishing run. He currently has an xG+ over 250, with six non-penalty goals scored compared to about 2.5 expected goals. We should most likely expect some decline. It is possible that Plata is an elite finisher, but simply having great finishing statistics in a one-season sample is not good evidence of that skill. There is too much variation.

Since Chris Wondowski has continued to get into dangerous positions and get his shots off, it is reasonable to hope that he can start scoring and justify his selection. Runs of bad finishing happen, even to good or great finishers, and they usually return to form.

Data supplied by Opta unless otherwise noted

Michael Caley writes for Cartilage Free Captain, where he analyzes fancy soccer statistics and bemoans Tottenham Hotspur’s most recent failures. You can follow him on twitter at @MC_of_A.



Success! Check your inbox for details. You might also like:

Please enter a valid email address

See all newsletters

Show Comments
Most Read



Success! Check your inbox for details.

See all newsletters

Your Three. Video curated for you.
Next Story
Neil Greenberg · May 28, 2014