Apples and Oranges: The problem with LA Galaxy MVP voting

The MLS season is drawing to a close, and the internet is all abuzz with MVP talk. Who deserves it more? Is it Robbie Keane? Landon Donovan? Obafemi Martins? Bradley Wright Phillips? Everyone seems to have an opinion. Everyone with a twitter account anyhow.

I have a love-hate relationship with the entire concept of MVP. While I am constantly forming my own opinions about who is deserving of such a moniker, I have a distaste for the way it is awarded. I realize I’m being a total stick-in-the-mud, and completely understand why it’s such a revered award. People like definitive answers regardless of the validity of the apples and oranges comparisons the award forces, not to mention the award’s subjective nature and complete lack of selection criteria. People love knowing who the best player in the league is.

So what are the criteria for MVP? Is the aim to find the best player in the league or the player who is most valuable to his team? Landon Donovan and Robbie Keane are both putting up elite numbers—so much so that an absence of either wouldn’t hurt the team nearly as much as Lee Nyguen’s absence in New England or Fredrico Higuain’s absence in Columbus would.

Assuming MVP doesn’t actually mean Most "Valuable" Player, but the league’s best player, picking a winner is still troublesome. Despite the murkiness of the awards purpose and selection criteria a few front runners have emerged— Landon Donovan, Robbie Keane, Bradley Wright Phillips, and curiously, Obafemi Martins. I say curiously here because his stats put him no where near the other front runners in any of the metrics that matter (most of which I use below). In fact, of the front runners, his candidacy is probably the only on which statistics alone can be used to come to a decision-- he's simply not the MVP.

The others, however, are much harder to separate. The case for Bradley Wright Phillips has been made fairly well by Harrison Crow in an article here, so I won’t bother making one for him. The purpose of this article is to try and seperate Landon Donovan and Robbie Keane. Both Robbie and Landon are having career years, however, when it comes to the MVP there can be only one. If you really nerd out on the task, the answers still aren't very clear and exposes the inherent impossibility of fairly comparing a guy like Robbie Keane against a guy like Landon Donovan. Consider the following table.

While Robbie Keane has 7 more goals, a direct comparison is invalid as Donovan has played a majority of his minutes as a midfielder. Robbie Keane’s Expected Goals is higher by 6.83, but this can be attributed to field positioning as well. Robbie Keane takes more shots. Landon Donovan’s shot leverage is .136 while Robbie Keane’s is .128, meaning Landon’s expected goals would be higher than Keane’s if he continued to shoot from the positions he is shooting from now, but increased the number of shots he is taking to Keane’s rate. Of course, that is a hypothetical that assumes that Landon wouldn’t take more shots from worse positions if he was playing Keane’s role as the withdrawn striker. Regardless, these are not fair metrics to compare them on.

So, to recap, none of the stats that Keane wins in are fair comparisons. So does that mean that Donovan is actually a better candidate for MVP?— not exactly. The areas where Donovan bests Keane aren’t exactly fair comparisons either. Donovan creates nearly double the chances of Keane and bests him in expected assists by nearly double. But again, the problem is that it’s an apples and oranges comparison.

Donovan’s job is mostly playmaking and ball circulation. Robbie Keane’s job is mostly goal scoring. While Robbie Keane accounts for 32% of LA’s shots, Donovan creates 25% of their chances—apples and oranges. These are two incredible players playing two completely different roles, and there is really no firm ground on which to compare them. Even if you try to use a combined stat like Shots Created, which combines shots and chances created, you run into problems. While Keane leads the stat by 12, the stat weighs shots and chance creation the same, when it’s unclear whether these two things occur at a close enough rate to be given equal weight in a combined stat such as this one.

So who is LA’s MVP? My gut tells me it’s LD, and he deserves the league’s MVP as well; however, I recognize I have absolutely no empirical basis to place him above Keane or Bradley Wright Phillips or any other of the top contenders aside from Martins. The award is incredibly subjective, and, in my mind, carries less value because of that.

But I don’t want to get rid of it. Getting rid of the MVP award would be kind of like stopping the practice of voting for a Prom King and Queen. Sure it’s stupid and criteria less, but it’s also beloved and fun. The award is also not completely devoid of value. The process of cutting the league’s vast player pool into a short list of contenders is fairly sound. Sure players fall through the cracks, but the players who do make the list are without a doubt extremely talented and amongst the league’s elite. If a player makes that list, it’s a tremendous achievement. The problem comes in picking the one.

So when the leagues "M"VP is named this year, remember to take that "most" with a grain of salt and recognize the tremendous achievements of all those whose names were considered.