The Myopia of Statheads

Sabermetricians have taught us new and better ways to analyze the game of baseball. Sometimes, however, they fall victim to the allure of their own imperfect theories.

For example, there are various sabermetric player projection systems out there that are marginally better at predicting player performance than the average well-informed baseball fan. These systems know a few things that you might not; e.g., that a player was particularly lucky or unlucky last year (mostly a result of his BABIP -- batting average on balls in play) or how players tend to change as they age, or what happens to hitters following anomalous seasons. But baseball is too unpredictable to...well... predict, even with a computer.

Nonetheless, some sabermetricians make judgements about players as if their systems are infallible and know more than the coaches, managers and GMs who are acquainted with the players personally and watch them play everyday. These stat geeks express shock when something occurs that's outside their projection systems.

Last week Joe Sheehan, a highly respected writer for Sports Illustrated and Baseball Prospectus, asserted that the Braves should flip Jair Jurrjens to another team next year as the Braves deal with an abundance of starting pitching (Jurrjens, promising rookie Tommy Hanson, aces Derek Lowe and Javier Vazquez, Japanese import Kenshin Kawakami and rehabbed star Tim Hudson.) Sheehan observed that Jurrjens has outperformed the pitching indicators and could yield a return beyond his actual value.

Before I disassemble this leap of logic, those indicators merit some discussion. Statheads have noticed that pitcher ERA and won-loss records fluctuate wildly from year to year, but pitchers are much more predictable in their ability to walk and strike out batters, which, after all, are outcomes over which they have total control. Statheads have further noticed that pitchers who aren't fooling hitters, i.e., striking them out, tend not to be very successful. That doesn't mean pitchers have to throw heat, just that whatever they're serving up has hitters confounded sufficiently to keep them in the park and prevent large numbers of screaming liners.

This isn't a foolproof formula: there are pitchers who regularly succeed or fail despite walk/strikeout totals that suggest the opposite, and there are pitchers whose performance fluctuates wildly even though their walk/strikeout rates remain constant. Instead, it gives rise to some guidelines that prevent analysts from jumping to conclusions about pitchers who've been lucky or unlucky.

It's worth noting that for his career, Jurrjens has been the beneficiary of a very low BABIP. Pitchers with low one-year BABIPs often have either great defense behind them or serendipity smiling upon them. Either one suggests the pitcher is less accomplished than he appears.

So back to Sheehan, Jurrjens and the Braves surfeit of arms. First, haven't we learned anything from the '09 Red Sox? There is no such thing as an abundance of pitching. A blow up here, a tired arm there, a couple of Tommy Johns and before you know it, your six starters have melted away. The Braves should try to sign a couple of outfield/first base bats in the offseason, but they should definitely not trade any of the most valuable commodities in baseball.

Moreover, Jair Jurrjens' walk/strikeout and BABIP numbers are irrelevant, at least to the extent that they presage future accomplishments. Sure, Jurrjens has walked 140 batters against 272 strikeouts in his 383 innings of major league work, a sub 2-1 K/BB ratio that is often an indicator of future trouble. And his .310 BABIP is about 20 points lower than normal.

But Jair Jurrjens has been doing this for two-and-a-half years. In 2008 and 2009, he's been one of the 15 best starters in baseball, with a lifetime ERA of 3.43. He's relinquished less than a hit an inning and a brilliant 0.63 home runs per game. Jurrjens hurls in The Ted, not a particularly HR-averse field. He's also kept up this low BABIP, which suggests that he's enticing batters to put balls in play weakly.

The Braves traded for the Curacao native in the '08 off-season because they saw something they liked. Would you agree that that particular franchise might know a little something about good young pitchers? In short, he has proven that he is a superb pitcher despite a tepid walk-strikeout ratio and low BABIP.

The question has never been whether statistics are helpful to analyzing baseball games and players, just whether we understand the stats and use them appropriately. Sometimes even the stat guys forget that.
b