Thursday, October 05, 2006

Liver follow-up

Tuesday, I argued that academics show little respect for the work of us amateurs. Specifically, I said that a recent academic paper by J. C. Bradbury on DIPS should have cited non-academic work that looked at the same question.

Many readers disagreed and said I was out of line.

Some of that discussion is
here. Dr. Bradbury’s response is here.

Also, some people objected to my description of the DIPS consensus, saying that’s not the consensus at all, and if it were, the consensus would be wrong. While the specific view on the DIPS question didn’t affect my argument from yesterday, I do want to get it right. I’ll investigate the more recent DIPS work and post on that soon.


At Monday, October 09, 2006 12:11:00 PM, Anonymous Anonymous said...

Good catch on the pattern of results on BABIP. It does look like there is an impact. However, the study's flawed methodology of using the prior year's performance data, rather than something like career performance prior to signing of last contract (and annualized value of current contract as dependent variable), makes it hard to even infer the size of the relationship, and I don't think you can draw any conclusions about the last few year's coefficients.

Finding a relationship also doesn't necessarily mean the market was "wrong." The market should give some weight to BABIP, but not a lot. For example, I'd argue that Zito will, and should, be rewarded this winter for his career .269 BABIP. It's part of his skill set. However, paying more for a pitcher because of one or two years of low BABIP, w/o any other evidence it was a real skill, would clearly be a mistake.

One quibble: you describe the difference between a BABIP of .320 vs. .300. as "only a moderate increase in BABIP." That's actually a very substantial difference: it translates into an ERA difference of about 0.50. If that were a true talent difference, 20 points of BABIP would separate an average pitcher from a star, or a star from a HOFer. That's important to keep in mind as you review the rest of the DIPS literature. A common error is to dismiss BABIP differences as "small" because they are small as a percentage of the mean, compared to a stat like K/9 (where a player can commonly be 50% above average). But what matters is RUNS, and superficially small changes in BABIP have a large impact on runs.

At Monday, October 09, 2006 3:43:00 PM, Blogger Tangotiger said...

I agree with Guy on the .300 and .320. A 20 point difference with 500 BIP is 10 hits, or 8 runs. A nine-inning game would have 28 BIP, so this works out to about 0.45 runs per game (or .40 in ERA).

To get that kind of difference, you need to increase your walks by +1.4 walks per 9 innings, or say going from 3 walks to 4.4, or a 47% increase! The change required in K is even higher.


Post a Comment

<< Home