Chris Young And The Problem With Splits

Last night our friend Matthew Artus at Always Amazin' disagreed, for the first time ever, with the Sandy Alderson think tank in Queens. It seems Matthew doesn't like the Chris Young signing on account of Young's divergent performances at home -- the incomparably run-suppressant Petco Park -- and on the road. The article, as with most everything Matthew writes, is well worth your time, but it does overreach a bit in its apparent assertion that Young's modest success in San Diego was mainly due to the generosity of his home park to pitchers of all stripes.

The natural dualism of splits is that they can be at once curiously instructive and easily misconstrued, and it's tempting to draw conclusions from them without first taking stock of their many caveats. With respect to Chris Young's home/road splits, there are a number of things we must consider before deciding that he's simply a lousy pitcher away from Petco Park.

1. Most players have a natural home field advantage, and for a variety of reasons -- supportive crowd, proximity to family, elevated comfort -- players generally perform better at home than on the road. For example, take a look at the NL splits for runs allowed (runs per nine innings) over the last four years.

   Year	Home R/9	Road R/9	Diff
   2007	4.56	5.04	0.48
   2008	4.44	4.89	0.45
   2009	4.31	4.76	0.45
   2010	4.14	4.68	0.54

   I use R/9 instead of ERA here because of a known scorer bias in home games (that is, official scorers notoriously, even if unconsciously, favor home pitchers by charging more errors -- and therefore fewer earned runs -- to the home team). It's only four years of data, but it certainly suggests a fairly significant advantage in run prevention for pitchers in home games. The average "home field advantage" for pitchers has been on the order of a half-run since 2007. That's a substantial difference and, frankly, a larger one than I was expecting.

2. Young's road splits don't include games at Petco (obviously), which probably short-changes him a bit because other pitchers' road splits do include games at Petco. Since Petco is a much friendlier park for pitchers than even the average NL park, losing out on one or more road games there can have a small but real affect on road ERA (or FIP or whatever).

3. Cutting a player's statistical record in half -- in other words, tossing away half of Young's innings by only considering his road performance -- necessarily leaves you with a less reliable basis for projecting his performance going forward. Whatever conclusions you might draw about Young based on the 750 innings he pitched the past four seasons must bear even closer scrutiny when you've shaved off 370 of those innings. Fewer innings means a greater chance that the performance was influenced by random variation. This could mean a positive or a negative affect, of course, but the salient point is that the predictive value -- which is really the crux of the argument over Young: How is he likely to perform in 2011? -- is diminished.

4. Whatever advantage Young is losing by leaving San Diego's home park he is gaining back to a significant degree by playing his home games at Citi field as opposed to, say, Coors Field or even a neutral run-scoring environment. That is, losing the "Petco advantage" is mitigated considerably by gaining the "Citi Field advantage."

None of this is to suggest that Young is a great pitcher or that he's some kind of game-changer for the Mets, but it does say that leaning too heavily on Young's Petco performance is to ignore -- and perhaps be consumed by -- the pitfalls that are concomitant with splits in general and home/road splits in particular.