A statistic we should ignore...

The Nationals wasted a pretty good performance from starter Jason Bergmann last night when the bullpen blew a 3-2 lead and gave up four runs in three innings. Here is the story from the Post.

The Post's story right after the game made an interesting point in casting doubt on whether Manny Acta should have removed Bergmann after the sixth inning. The story asserted that Bergmann is very good at keeping opposing hitters' batting average down when they put the ball in play. In other words, hitters who hit the ball fair tend to make more outs against Bergmann than some other pitchers. This season's stats tend to suggest the Post is correct: the batting average of hitters when they put the ball in play (otherwise known as "BABIP") against Bergman is .224. To put this in context, Chad Cordero's BABIP is .414 and John Patterson's is .305.

So, does Bergmann's relatively low BABIP mean anything?

No.

In what was truly groundbreaking research, Voros McCracken showed that BABIP was a very poor indicator of pitching performance because pitchers have almost no control over the outcome of balls put into play. Defense, park dimensions, weather, and the apparent randomness of what can happen when a round ball meets a round bat all have dramatic effects on the outcome of a ball put in play. A pitcher cannot control any of those elements.

McCracken's research was very controversial at the time. Here's what he said in an article he published on the Baseball Prospectus site in 2001: "'You're insane.' That's generally the response I get when I present the information you're about to read." People eventually accepted McCracken's counterintuitive research, including Bill James, who said in his "New Historical Baseball Abstract" published in 2001 that he felt stupid for not having realized it 30 years ago.

Don't believe it? Well, consider this. Of all National League pitchers with 100 or more innings last year, Ian Snell had a BABIP of .327, which put him 58th out of 68 players. Not very good, right? But Snell is one of the best young pitchers in the National League. Livan Hernandez, a pitcher who had lost almost all of his power and was on the tail end of a career, had a BABIP of .322, good for 55th place. Two pitchers with similar BABIPs, but with dramatically different talent and productivity.

Still not convinced? Roy Oswalt, one of the best young pitchers in the game, had a BABIP last year of .310, good for 46th place. Oswalt's equivalent ERA (adjusted and converted to a neutral-park major league equivalent peformance) was 3.06. Finishing ahead of Oswalt were such luminaries as Woody Williams (.275, 11th place), Eric Milton (.275, 12th place), and Jason Marquis (.294, 26th place). The equivalent ERAs for these pitchers were 4.38, 4.61, and 5.90. No one in their right mind would conclude that Williams, Milton, and Marquis were better than Oswalt.

So, Bergmann may have a good BABIP now, but that statistic does not correlate to his success over time, and his BABIP will almost certainly increase over that same time.