Proceeds from the ads below will be donated to the
Bob Wuesthoff scholarship fund.
Tuesday, May 24, 2005 |
Two From Hardball Times
A couple from Hardball Times:
- First, this Aaron Gleeman piece about the Hee Seop Choi and Dallas McPherson, in which he quoted my earlier brief piece. Thanks for reading, Aaron.
- JC Bradbury goes another round dissecting DIPS, and basically, concluding yet again that McCracken was right. Is this really necessary, though? I'm beginning to think there should be a class of sabermetric "off limits" categories for lack of being able to show any useful work, and one of them is disproving the fundamental thrust of DIPS: namely, that the pitcher has only a very limited ability to control the behavior of a batted ball in play.
Comments:
Studes, I think the problem is that nobody has bothered to actually show any useful results other than the original work. There is a great deal of huffing and puffing around the periphery, but every time I read stories like this one, I get several distinct sets of commentaries:
1) Fooling about at the edges, sometimes with selective sampling (a la Tom Tippett)
2) Increasingly fine (some would say heteroskedastic) subgrouping of data, which may or may not prove anything
3) Redundant "I couldn't find anything to disprove DIPS" articles.
For all the gas exhaled on the Primer thread, I notice none of them has published a groundbreaking study systematically disproving DIPS. The basic papers have been published for three years or more, yet still it stands. Anti-Moneyballers and the general gloria mundi awaits him who accomplishes it.
1) Fooling about at the edges, sometimes with selective sampling (a la Tom Tippett)
2) Increasingly fine (some would say heteroskedastic) subgrouping of data, which may or may not prove anything
3) Redundant "I couldn't find anything to disprove DIPS" articles.
For all the gas exhaled on the Primer thread, I notice none of them has published a groundbreaking study systematically disproving DIPS. The basic papers have been published for three years or more, yet still it stands. Anti-Moneyballers and the general gloria mundi awaits him who accomplishes it.
Studes -- did I say McCracken had stuck to his earliest statements of the problem? No, and that's important to understand. But the realization that lefthanders and knuckleballers have differing rates of BABIP control is something that Voros knew. I don't think there's been appreciable progress in this field since then. How about MGL in BTF, 2/29/04:
"That's right - once we take defense and home park out of the equation, there appears to be almost no skill in a pitcher's ability to prevent hits on balls in play! Voros was right!"
He goes on to say that frequency of various types of batted balls "suggest" there might be a BABIP control, but he never actually proves it. My timeline begins at essentially Voros II, where he backed off his initial claims a bit and expanded on his findings (i.e., knuckleballers seem to have somewhat more control on BABIP than other types of pitchers). The big problem may simply be that we don't have adequate descriptions of batted ball trajectories and the scorers are just too subjective for us to get decent understanding of subcategories, and/or the smaller sample sizes of each type of ball masks other difficulties in making empirical statements.
Thanks for the discussion, Studes.
"That's right - once we take defense and home park out of the equation, there appears to be almost no skill in a pitcher's ability to prevent hits on balls in play! Voros was right!"
He goes on to say that frequency of various types of batted balls "suggest" there might be a BABIP control, but he never actually proves it. My timeline begins at essentially Voros II, where he backed off his initial claims a bit and expanded on his findings (i.e., knuckleballers seem to have somewhat more control on BABIP than other types of pitchers). The big problem may simply be that we don't have adequate descriptions of batted ball trajectories and the scorers are just too subjective for us to get decent understanding of subcategories, and/or the smaller sample sizes of each type of ball masks other difficulties in making empirical statements.
Thanks for the discussion, Studes.
Newer› ‹Older
 |
Post a Comment
Note: Only a member of this blog may post a comment.