Tuesday, January 25, 2011
Voros McCracken changed the game
Below, a long excerpt from a much longer article about Voros McCracken, who merely changed baseball analysis as we knew it, practically overnight. Jeff Passan:
When a Baseball Prospectus book suggested pitching and defense were too intertwined to be separated, the defeatism bothered McCracken. At the gifted school, he had taken a critical-thinking class. One lesson he gleaned: divide and conquer. By isolating the elements into independent and dependent groups, McCracken was able to prove the most accurate assessment of a pitcher comes not from the hits he allows but his strikeout, walk and home run tallies.
The year-to-year instability of batting average on balls in play (BABIP) remains the most difficult concept to fathom. Groundball-to-flyball ratio correlates for pitchers over their careers. BABIP correlates with groundball-to-flyball ratio. Thus, the averages on balls in play should be consistent. Only they're not. Not close. It wasn't just Martinez. Greg Maddux's ebbed and flowed. So did Curt Schilling's. And just about every other pitcher.
The incongruity kept others from discovering DIPS before McCracken. [Bill] James had come close, as had a rec.sports.baseball poster named Dale Stephenson. Another poster, Keith Woolner, broke down the question similarly to McCracken, only to answer it incorrectly. Woolner, now an analyst with the Cleveland Indians, described a scenario in which "there's no difference, on average, between a ball hit off of Tim Belcher and one hit off Roger Clemens, assuming that it lands in play." His response to it: "To my mind, that isn't a reasonable assumption."
McCracken's findings proved otherwise, and they took one of the most important statistics for a pitcher – hits allowed – and essentially absolved pitchers.
"I knew right away that the article was groundbreaking, that it was going to change the way we looked at baseball permanently," says Rany Jazayerli, one of the founders of Baseball Prospectus. "Like, I imagine, everyone else, I was skeptical about his conclusion when I first read it, but it only took a few minutes of looking up the stats for various great pitchers -- and the stats for mediocre pitchers who had lucked into a great season -- to realize Voros was onto something."
The response from the rest of rec.sport.baseball, and later Baseball Prospectus readers, mimicked Jazayerli's: disbelief, followed by a eureka moment. The acceptance of DIPS took longer for some than others, and McCracken fought every question, every cry of fraud and hack and con man, to explain that however someone felt about him, the data didn't lie, and his spreadsheets told him this was the truth.
"He was in hand-to-hand combat to convince people," says Dan Szymborski, a sabermetrician and friend of McCracken's. "People think sabermetricians are this community where everyone agrees and it's this monolithic view of statistics, and statheads can't leave an argument. Trust me. They argued with Voros."
There were small flaws to McCracken's research, most of which came from lack of data. Knuckleball pitchers, follow-up studies showed, give up fewer hits on balls in play, and there is a difference between groundball and flyball pitchers. Perhaps McCracken's quick-and-dirty take from the beginning of pitchers having no influence was premature. The essence of the theory remains: it is minimal, far less, certainly, than baseball presumed before McCracken's research.
"Voros' realization," James says, "has become one of the pivotal points in the history of sabermetrics."
Just one small correction: DIPS does not "absolve" pitchers for hits allowed. It absolves pitchers who give up more hits than theory suggests they should, based on their strikeout rate. With that out of the way ...
I played my own small role in the popularization of DIPS.
I am, I must regretfully admit, highly suggestible. If you can throw some facts at the wall and connect them with reasonably good writing, I'm probably going to buy your conclusion unless I've got some reasonably contradictory facts -- or prejudices -- on hand.
Well, when it came to DIPS I didn't have any other facts, or prejudices. I just figured Voros McCracken must be some sort of genius, and I devoted most of a week to McCracken's theory. For the only time in my career, I went way out of my way for credibility, e-mailing both Bill James and Craig Wright for their thoughts. Both veteran sabermetricians were impressed enough and gracious enough to essentially pen guest columns. Both tried but failed to see any significant flaw in McCracken's logic.
DIPS would have hit big without me. But I helped a little, because my ESPN.com column gave DIPS its biggest exposure to that point. Of course, later came "Moneyball," etc.
And I certainly can't take even a dollop of credit for any of it. McCracken did all the work, engaged in the hand-to-hand combat, has suffered for years from his inability to come up with another once-in-a-lifetime revolutionary ground-breaking earth-shaking insight. I was just sitting there, waiting for someone to blow me away.