Critics of sabermetrics often cite its impenetrability, or its "arbitrariness," or its newfangledness, or just its like math-ness. Too much math, they say. Too many nerdy numbers. Who needs numbers? What can they ever tell us? Facts? Phooey. I will rely on my eyes, my gut, and a different set of equally arbitrary mathematical cutoffs which aren't nerdy because I am familiar with them.
To those people I say: relax. Because Jon Weisman over at SI.com has invented a new kind of mathematical system for player evaluation. It's a veritable Principia Mathematica for the baseball fan. Here's how it works.To get a sense of which teams had the strongest rotations, I awarded each pitcher points based on the following categories (looking at three-year trends, with the most weight on the 2007 season):
• 0 points: below-average pitcher
• 1 point: mystery pitcher -- wildly inconsistent pitcher or above-average recent track record but with dubious health
• 1 point: young, up-and-coming minor-league pitcher with above-average potential in 2008
• 2 points: average to above-average pitcher
• 3 points: above-average pitcher
• 4 points: super above-average pitcher
Simple! Now you just assign the points to their cut-and-dry, indisputable, category-based participants, and you get to see who has the best staff. This system of pitcher evaluation will soon be all a GM needs to evaluate trades, draft choices, or free agent signings.
"Should we pick up Aaron Fultz?"
"Why? Our bullpen is already set, and he's a below-average pitcher."
"Wrong! He technically qualifies as an average pitcher. Two points, not one. This will put our pitching staff at 16 points, and the next best team in our division only has 15."
"Get it done."
Weisman even anticipates that some nerds might have a problem with this system:
Now, there are certain to be quibbles about the choices I made in assigning point values -- in fact, the entire point system is rather arbitrary. There's a built-in margin for error -- because of how difficult it is to predict future performance, even with the best projections. These are not meant to be scientific.
Don't worry, people. This mathematical formula is not meant to be scientific. Which means there is no benefit to it beyond the benefit you might get from like looking at ERA trends or K/BB trends. Whatever. The point is: it's a system!
My personal favorite point of the apologia is that he says "There's a built-in margin for error." I disagree. I think this is a rock-solid problem-free scale. When you consider that an "average to above-average pitcher" gets two points, and the completely different rating of "above-average pitcher" gets three points, and when you further consider that the difference between most of the teams in the final rankings is one point, well, then you come to the inevitable and time-wasting conclusion that the built-in margin for error is: the entire system.
I know that this kind of off-season article is mostly just goof-off time-wasting before pitchers and catchers report, but goodness gracious.
EDIT:
Many of you have nervously emailed me to report that Weisman created and mostly hangs out over at dodgerthoughts, and is generally a very smart and statistically-inclined kind of a dude. (I have cruised the site in the past and enjoyed it, FWIW, which probably isn't very much.) And like I said in the post, I know that fun distraction-type point attribution systems are all that baseball fans have, sometimes, in the off-season, and I know he caveated the hell out of it, but the problem, to me, is the forum.
If Weisman had limited this baby to dodgerthoughts as a goof-around boredom killer, then that's cool with me. Murbles, dak, some other dudes and I once had a mock draft of rock albums, books, and movies for 9 hours in my apartment in New York. But when these things go national, and people read what they assume is gospel on a mainstream site like si.com, then I get a little annoyed. The same simplistic evaluation could have been done with actual illuminating information on the pitchers' histories instead of a point-attribution system that the author himself admits is arbitrary. And the caveats, while extant, come post-facto, when the system has already been laid out.
(I see that Weisman has just posted about our post. I will conclude only by saying that I like the dude's writing, generally, and I wouldn't anticipate the "jon weisman" label getting too much bigger anytime soon.)
Labels: jon weisman, math, principia mathematica
