Monday, June 25, 2007

Rockies Home/Road Splits

The Cubs/Rockies game is on TV, and the announcers are on about how Matt Holliday has proven to be a good hitter even outside Coors Field, as evidenced by his statistics.

First off, even though I understand that Holliday is a good hitter, it's ridiculous to insist that he isn't getting a huge benefit from his home park. He's hitting .404/.443/.669 at home and .305/.352/.493 on the road. Basically, he's Ty Cobb in Coors and Bernard Gilkey in away games.

I actually have a broader bone to pick. Often, especially when dealing with the Rockies, you'll deal with splits that look off for some reason. One common example is that a Rockies hitter will have a higher OPS in away games, "proving" that he doesn't need Coors to hit well. This is a fallacy.

Suppose you find that you can consistently hit a golf ball 200 yards at your local driving range. You decide that you'd like to see how far you can drive on the moon, so you build a space shuttle and head up there. In the lower-gravity environment with no air resistance, your average distance per drive is...200 yards.

Does this mean that you derived no benefits from the moon's generous environment? Absolutely not. The differences in gravity and air resistance affected you the same way they would anyone else; you just performed worse on the moon. It's the same thing in Coors. If a Rockies hitter has a higher road OPS than home OPS, it doesn't mean Coors didn't help him, it means that he hit worse at home than on the road. This is why we use standard park factors to compare hitters rather than doubling everyone's road numbers or using individual adjustments.

That brings me to a good rule of thumb. If someone on TV tries to use split statistics to explain anything, mute your TV immediately or suffer the consequences.

No comments: