Monday, November 4, 2013

More on the False Dichotomy

In my post comparing the winners of the Gold Glove Awards to the Fielding Bible Awards in the Bill James Handbook, I wrote, "The stats vs. scouts storyline is a false dichotomy." I mean, the idea that statistical analysis and more subjective analysis are incompatible resonates more in Hollywood (Moneyball, Trouble With The Curve) than in, you know, real life. Show me a scout who thinks that statistical metrics are useless, or an analyst who thinks that scouting reports are useless, and I'll show you someone who is willfully ignorant.

That's not to say the argument's been extinguished. Enter Murray Chass, former sportswriter for the New York Times and now a baseball blogger. Chass really, really dislikes modern statistical analysis. His latest target: The Gold Glove Awards, whose sponsor, Rawlings, he says "has gone over to the dark side" and "gone off the deep end" by "embracing the kind of statistics" (aka "gobbledy goop") that "are at the center of baseball's fiercest debate." I'm not making this up; his column is here.

Chass is upset that Rawlings is giving a 30% weight to statistical fielding measures in determining the Gold Glove winners. Did you know that fielding statistics "are at the center of baseball's fiercest debate?" Hey, DH, PEDs, hiring a manager who's never managed before, Yasiel Puig's baserunning, large market/small market payroll differences, Bud Selig's legacy--you're all off the hook! The biggest, most intractable issue in baseball is fielding statistics!

Anyway, over at Baseball Prospectus, Russell Carleton addressed Chass's contention that the statistical component of the vote is somehow unfathomable. (BP's a pay site, and well worth it, though Carleton's article is free.) Without using any differential equations or calculus at all, Carleton cuts to the quick:
All defensive metrics come down to three questions: "How many balls should he have gotten to? How many did he actually get to and convert into outs? What is the difference between those two numbers?" We start with the assumption that a good fielder is a guy who turns more balls than average into outs (that is, after all, his job description), and work from there. 
He also explains how prior years' Gold Glove votes reminded Carleton of the Homecoming Queen vote at his all-male Catholic high school. Carleton's an entertaining writer; his piece is worth a read.

The thing about modern fielding metrics: By and large, they're compiled by people tracking every ball in play and determining how frequently they're fielded cleanly. The word "metric" is scary and all, but these are basically tallies of the number of times each fielder makes plays compared to how other fielders do. It's the opposite of somebody deriving a complex formula that the anti-stats types loathe. It's using observation, nothing more, over and over.

No comments:

Post a Comment