Determining a player's shooting efficiency is a difficult task. Previously in this series we have explored the concept of chance quality. Using the touch-by-touch event data that Opta collects, we can make a measurement of how likely a shot is to result in a goal based on a handful of observations.
In other words, if the shot we are measuring was attempted repeatedly, how often would it result in a goal?
From here, we can estimate how many goals a player should have scored (based on the quality of chances they accumulated over the season) and compare it to the number of goals that they actually scored. This is a concept called "goals added," and some of the top Golden Boot contestants scored exceptionally in this measurement.
Here are the top 10 Expected Goals Added performers for MLS 2013.
| PLAYER | EXPECTED | NON-PK GOALS | GOALS ADDED |
|---|---|---|---|
| Camilo | 9.6 | 17 | 7.4 |
| Marco Di Vaio | 14.1 | 20 | 5.9 |
| Mike Magee | 11.7 | 17 | 5.3 |
| Tim Cahill | 6.0 | 11 | 5.0 |
| Blas Pérez | 61 | 11 | 4.9 |
| Juan Agudelo | 4.2 | 9 | 4.8 |
| Robbie Keane | 6.2 | 11 | 4.8 |
| Diego Fagundez | 8.4 | 13 | 4.6 |
| Landon Donovan | 5.6 | 10 | 4.4 |
| Lloyd Sam | 1.1 | 5 | 3.9 |
Does not include penalty kicks
This is not perfect (but a far better indicator than raw shooting accuracy). It fails to contextualize for a few things. A player's chances aren't purely an individual skill, but instead a combination of teammate skill, system fit, opposition strength and even pure luck. And, to top it off, the model isn't exact!
This model controls for things such as the shot location and angle, if the shot resulted from a foot or a head, or if it came from contexts like open play, a free kick or even a counterattack and many others. The model is robust, trained on tens of thousands of data points, but it can't control for everything.
All that being said, expected goals is probably a better indicator for "Who should have won the Golden Boot?" race rather than overall shooting efficiency. Will Bruin had the league's lowest Expected Goals Added rate with -5.97. With the chances he had this season, he could have been deep in the Golden Boot race.
To remove some of this bias, we look to how a player performed relative to their average chance quality while adjusting for their shooting volume.
Here are this season's top performers in Expected Goals Added per shot.
| PLAYER | AVERAGE EXPECTED GOALS ADDED |
|---|---|
| Juan Agudelo | 16% |
| Landon Donovan | 12.1% |
| Tim Cahill | 9% |
| Blas Pérez | 8.8% |
| Diego Fagundez | 8.4% |
| Robbie Keane | 8.1% |
| Kekuta Manneh | 7.3% |
| Ryan Johnson | 7.3% |
| Will Johnson | 6.9% |
| Álvaro Saborío | 6.7% |
For players attempting at least 30 shots
Notice the trend toward lower-volume shooters and away from the players who were aspiring for the Golden Boot title. The New England Revolution, one of the premier defensive teams in MLS in 2013, did not generate as many scoring opportunities as many other teams. Instead, they were forced to rely on ruthless efficiency in front of goal. With Juan Agudelo and Diego Fagundez, the Revs had two of the most efficient finishers in 2013.
Again, context kills. This does not mean that Agudelo could maintain his ruthless efficiency in a situation where he would be provided a much larger volume of opportunities. What I am trying to say is, this is likely a non-predictive metric across teams and leagues.
But this measurement does show how exceptionally a player preformed given the opportunities he was presented in the system he was operating in.
