Saturday, February 07, 2009

Calling All Sabermetricians-Again

Spring training and Barry Bonds' perjury trial will be getting underway very soon, which reminds me to offer a serious plea for help to baseball's quantitative researchers. Please leave off inventing new statistics with acronyms taken from the U.S. Department of Defense and put your admittedly good brains to a task that will serve humanity, or at least that subsection of humanity with a vote for the Hall of Fame.

We know that home run totals from somewhere in the early '90s to somewhere in the early 2000s (peak period: 1994-2002) were artificially inflated by the use of performance-enhancing drugs. We also know that other, less chemical factors contributed to the increase in homers, the most notable being expansion, which equals more lousy pitchers, and the rise of the Camden Yards-style bandbox ballyards. We also know that pitchers were taking performance-enhancing drugs, too-albeit to less striking effect.

So, my plea is this. Can you guys sort all that out for me? That is, can mathematical analysis figure out exactly, or even approximately, how much steroids helped hitters, especially prominent ones? Just what percentage of homers hit in the Steroid Era WERE artificial, anyhow?

Expansion, having happened several times before the 1990s, should be a quantifiable factor. I know ballparks are. That leaves the drugs. Surely, this issue cannot be beyond the reach of science.

So far, I've been using the following metric for home runs totals of the period. I cut them by 20 percent. This reduces, to take a not entirely random example, Mark McGwire's career homer total from 583 to a steroid-adjusted 460 or so, which I think is a rough estimate of just how good a slugger McGwire was-damn good but not unearthly.

Barry Bonds ruins my guessing method. Subtract 20 percent from 73 homers in a season, and you still get almost 60, far more than the pre-1998 Bonds ever came close to hitting. It would appear that drugs had a value-added effect. The better the player who took them, the more his performance improved.

So further study is needed. You numbercrunchers are always telling us old farts that guesstimates are no substitute for exact data, and you're right. So help a brother out here. This is about the most pressing issue in baseball historical research right now, and as the Hall-eligible classes of players roll on, it'll become even more pressing.

I regret I do not have any large grants to fund this project. I wouldn't bother calling the Hall of Fame, because it wants to pretend the 1990s never happened, and let us voters sort out this sticky problem for it.

Does Bill Gates like baseball?

1 Comments:

At 1:36 PM, Anonymous Barry Griffiths said...

Mike --

Funny you should ask that. It was actually done a couple of years ago by the economist and serious sports nut Arthur DeVany -- you can find the paper at http://www.arthurdevany.com/webstuff/images/HomeRunHitting.pdf. DeVany concludes that steroids made no statistically detectable difference in home run hitting. DeVany has done a lot of sophisticated statistical work on topics where most of the published stuff is...very bad. See, for example, his book "Hollywood Economics," which largely supports William Goldman's contention that in the movies, "nobody knows anything."

Best,
Barry

 

Post a Comment

<< Home