LA Galaxy have fallen into a familiar trap by signing Dan Kennedy

In 2011, mlssoccer.com named Dan Kennedy the most underrated player in the league, and, ever since, it's a tag that has stuck with the keeper. Over time, a certain narrative was built around this notion, and from 2011 to 2014, it was not at all uncommon to run across someone who'd make the argument that Dan Kennedy would be on the national team if he played for a better team.

Due to this "Chivas USA effect," it is often said that numbers don't tell the whole story when it comes to Dan Kennedy. While this is technically true (numbers don't tell the whole story of any player), I find the assertion to be a particularly lazy one.

Shutouts and goals against averages are fairly unreliable metrics to begin with, and their failure to fairly assess the skill of Dan Kennedy is simply a demonstration of this fact.  This should not be conflated with the notion that all goalkeeping metrics are equally incapable of stripping away the effect of playing in front of such a poor defense, or that numbers are incapable of measuring at least some aspect of his quality.

I have often criticized the LA Galaxy for their history of overpaying for players on the decline. Is Dan Kennedy such a player?

Save Percentage

Save percentage is a measure of what percentage of shots on goal are saved by a keeper. While it is not without its flaws (something we'll go into and control for in the next section), it has the benefit of controlling for the number of shots. In other words, if Dan Kennedy saves 16 of 20 shots on goal while another keeper saves 7 out of 10, we can use save percentage to determine that, despite giving up more goals than the other keeper, Kennedy saved a higher percentage of them. Since Chivas USA's horrid defense gave up more shots than most defenses in the league, this is a good control to.

Dan Kennedy's best season came in 2011, which remains one of the all-time best seasons for any goalkeeper in the modern MLS era. From 2011 to 2014, only 8 players who spent 2000 minutes or more in net logged a higher save percentage than Dan Kennedy's 74% mark in 2011 (he actually ranks 11^th since Nick Rimando and Bill Hamid both appear above him twice).

While 2011 was undoubtedly an amazing year for Kennedy, his save percentage has been in free fall ever since.

Going back to our list of single season goalkeeper performances from 2011-2014, Dan Kennedy's 62% mark in 2014 ranks 65^th out of 68, and €”only 3 percentage points ahead of Carlo Cudicini.

So how normal is this? There really aren't a lot of keepers in the league who have managed consistent minutes every single year from 2011-2014. Kennedy is  1 of only 5, but of this group, while we clearly see that save percentage tends to fluctuate from year to year, Kennedy's is the only  one whose numbers got worse every year.

Adjusting for shot quality

Of course, not all shots are created equal, and, given the nature of Chivas USA's defense, it could very well be that they were allowing greater and greater chances each year.  In order to test this theory, we can look at expected goals. Expected goals is a metric that essentially determines shot quality by taking things like distance, pass type, body part struck with, and type of build up into account in determining the likelihood of a shot being scored, according to league averages.

If opponents were truly getting better shots year to year against the Chivas USA defense, we'd expect to see a rise in Dan Kennedy's expected goals against (xGA) which would  inversely mirror the decline in his save percentage.  While something like this can be seen from 2011-2013, in 2014 he saw about the same level of xGA per 90 as his stellar year in 2011 but finished with a save percentage 12 points under that number.

There is also little discrepancy in how many shots he faced in 2011 vs. 2014, or in the breakdown of the types of goals.

It should be noted, however, that Dan Kennedy battled a hand injury in 2014, which could have a lot to do with how poor he was in that particular season. While this allows us to excuse at least one year of data, questions must still be asked about 2012 and 2013.

A finer of way of comparing xGA to save totals than simply eyeballing the above graph is to look at how many goals a keeper has allowed above or below what the model says a league average keeper would have allowed. This can be done by simply subtracting a keeper's goals allowed per 90 rate from his xGA per 90 rate. Here is how Dan Kennedy stacks up.

Unlike the other charts, this is one where trending up is a bad thing since anything above 0 is goals allowed per 90 over what is expected. Once again, we see that Dan Kennedy is the only keeper of the group whose numbers get progressively worse each year. While the hand injury might account for 2014, his numbers in 2012 and 2013 aren't that great compared to the quality of shots he was facing. This suggests that the slip in his numbers has more to do with him than it does to do with the defenses in front of him.

Conclusions:

At one point in time, Dan Kennedy was certainly an underrated keeper. Since 2012, however, I'd argue quite the opposite. It seems that in our haste to excuse Kennedy for the sins of the defenses in front of him, we've failed to notice that Dan Kennedy hasn't been truly elite since 2011. Considering his salary cost and injury struggles throughout 2014 and 2015, this once again feels like the Galaxy front office has overpaid for a player on the decline.