Monday, April 11, 2005

Fatal exception?

The Nats are on computers now? Nope.

As Scoopy McScoopscoop pointed out, it's not exactly shocking that Frank Robinson disassociates himself from the "stathead" movement or even disavows knowing how to operate a computer. (As an aside, I consider District of Baseball and William World News beyond "scoop"; they're so early and current, they might as well be writing the news.)

I can't really add much to what Chris has already said; while my normal inclination might be to say pretty much the same thing in far more words (part of my charm, I suppose), I'll just add a few (hopefully) quick comments here:

1. Robinson might be conflating "statistical/analytical ideology" with "minutaie." For instance, did Earl Weaver know what Robinson was hitting on a given day? Probably not. Did Robinson's general "statistical profile" fit what Weaver looked for in a player? You bet. (Well, anybody would be crazy to say, "Robinson? Nah. Don't want him.)

2. Can't say I disagree with Robinson one bit here:

"Numbers don't win you ball games," he said. "Many times in a game it comes down to what is here [pointing to his heart], here [pointing to his gut] and here [pointing to his head]. Numbers don't mean a thing a lot of times out there on the field. . . ."

That's right, of course.

While many statheads deny the existence of particular, uniquely "clutch hitters," it is of course true that "clutch hits" happen all the time. It is axiomatic of the variable nature of an at-bat, a ball-in-play, a pitched ball, etc. that flukey things will happen. Luis Sojo will even knock RBI singles up the middle during the World Series. While it's more likely that, you know, a legitimate major league player will come through in such a spot, it's hard to tell an experienced major league manager, who in a sense has collected his own "data" thoughout his career, to play against his own hunches in a specific situation.

3. Robinson puts a noteworthy number of batter/pitcher confrontations at fifty. That sounds good to me, for all I know. I recall around 1999 or so Joe Morgan was on a kick of saying, "Anything 20 or above is meaningful." Maybe that's true; maybe it's not. Morgan sure thought so. I recall posing the question on rec.sport.baseball, and no one there---you name 'em, Voros McCracken, BP guys, whoever---answered with any certainty. The reason, of course, is that any batter/pitcher confrontation figure, unless the batter and pitcher both play until we colonize Mars, will invariably represent a rather small sample size. Then again, it has to be true, I suspect, that some batters have pitchers' numbers, and vice versa.

4. I think it's funny that the subject of "computers doing the job of baseball managers" comes up, considering that Papa Stats, Bill James himself, decried the possibility of computers in dugouts two decades ago. Consequently, I conclude that either the criticisms are uninformed, or that the statheads have gotten bolder.

5. Finally, yesterday I placed the words "Mike Wise" and "Laura Vescey" in the same sentence. Today, I apologize that Mr. Wise; that was cold. And today's Wise WaPo column on the Robinson-computers angle was quite well-presented, in my humble opinion.

Comments:
I refuse to read Wise since he stole my "Will the real Esteban Loaiza please stand up?" line on Friday. ;)
 
That's a good line, too. ;-)
 
In all seriousness, though, I don't think it's possible to pick a specific number of appearances to say it's significant.

I mean, if a usually .275 hitter is 4-for-20 against a guy, that might just be a couple of unlucky bounces. But if a guy is 7-for-9 with 5 homers, it's probably safe to say he's figured a pitcher out. Conversely, if somebody's 0-for-6 with 5 strikeouts and a popup, I'd say the pitcher's got him.
 
Well, it depends on the context of those appearances, of course.

A pitcher, for instance, might have had a hitter's number years ago (when the hitter was just coming up), then the pitcher went to the other league, and now he's back, and the hitter is by now a mature major league batter.

Or the previous 2-for-19, or what have you, might have occurred before the pitcher tore his labrum.

It's just tough to tell. Even aside from sample size concerns, you have to know the context of the previous encounters, too.

I would agree, though, that 7-for-9 with 5 homers might justifiably light up the "DANGER!" sign. ;-)
 
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?