Wednesday, March 9, 2011

Irony: Oops! Edition

In an article decrying a ranking system that puts Roger Federer #7 all-time, Bruce Jenkins writes:
Unfortunately for Mr. Radicchi, his findings come off as the tennis equivalent of statistical claims that Derek Jeter is a lousy shortstop, or that Kobe Bryant is a poor shooter in the clutch. You gotta be kidding, in other words.
Many analysts--maybe even most!--would claim exactly that. Look, Kobe's crunch-time record is inflated and any serious analysis can't come to any other conclusion:
In the final 24 seconds of playoff games, Bryant has racked up almost as many air balls as makes, making just below 30 percent of game-tying or go-ahead shots. He hasn't hit such a shot in a playoff game, in fact, since 2008, including key misses in the closing moments against the Jazz and Magic in 2009, and the Thunder and Suns last spring. He made one of his four shots in the fourth quarter of Game 7 of last year's Finals.
...
ESPN Stats & Information's Alok Pattani dug through 15 years of NBA data... -- Bryant's entire career, regular season and playoffs -- and found that Bryant has attempted 115 shots in the final 24 seconds of a game in which the Lakers were tied or trailed by two or fewer points. He connected on 36, and missed 79 times.

And baseball people will tell you that Derek Jeter is not such a great hitter and has often been a terrible fielder. While I'm not particularly familiar with baseball statistics, and I can't vouch for the veracity of the system, this is what many people who look at statistics believe.

The knock on sportswriters from statistically-oriented people is that they pump out lazy cliches without thought or analysis. It's ironic that in denigrating someone's effort, Bruce Jenkins unwittingly confirms that exact stereotype. He picks famous players, with great reputations. Well, where'd they get that fame and that great reputation? Largely from other sportswriters. Jenkins' work as a sportswriter is based on other people's work as sportswriters, and so on in a self-referential loop, without any reference to any outside performance measurement. Can it be too much to ask for more? If Bruce Jenkins is any guide, perhaps not.

No comments:

Post a Comment