08/04/2009 1:46 PM ET
Yanks shouldn't shut Joba down
Forget the Verducci effect; team needs right-hander
By Michael Salfino / SNY.tv
Jon Lester has put up excellent numbers for the Red Sox, who are not worried about his innings and the Verducci Effect. (AP)

Joba Chamberlain has been the Yankees' best pitcher of late. Even so, the front office may be winding his season down due to adherence to a quasi-scientific pitcher workload theory called the Verducci Effect, named after Tom Verducci, the Sports Illustrated writer who coined it.

"Quasi-scientific" is too kind. The theory is anti-science -- something proffered with no evidence because it sounds true. In non-scientific, Woody Allen parlance, "It's a travesty of two mockeries of a sham."

The two biggest advocates, Verducci and Baseball Prospectus' "injury guru" Will Carroll, have not yet cited any studies of it to the best of my knowledge. Carroll told me earlier this season when I was writing a broader piece on this subject that "he's working on another study now." We're still waiting for those results to be published.

Well, all of us except Brian Cashman, who has clearly seen enough. Never mind that the only real multi-season study to date by David Gassko of the great HardballTimes.com in 2006 debunked it. VE pitchers under age 25 fared no worse in the following year than non-VE pitchers (those whose workload did not increase 30 or more innings) who also were under age 25.

Verducci and Carroll are not statisticians, nor does either have any medical training. Neither do I. But I'm calling out the VE, anyway, for the following reasons:

VE proponents have rigged the data. Why is 30 innings a magic number? Because that was the number that best supported the point that Verducci was trying to prove when he looked it up. And don't be fooled by their argument that this is not meant to be taken literally -- that 29 more innings than the prior year is fine but at 31 your arm falls off. Either 30-plus innings is a proven effect or it is not. Clearly the VE people say 30 innings, and not the number of pitches, is the line in the sand. The latter makes more sense when you think about it for even two seconds.

And if 30 innings were a meaningful cutoff, we should see a trend of even greater injury frequency the further you go past that threshold. But the proponents haven't yet done this and merely lump together the entire pool of all 25-and-under pitchers with 30-plus more innings than the prior year.

Using poor(er) performance in the year after the 30-plus innings of use makes the target too large. We're looking at, by definition, young, unproven pitchers. The assumption that the VE makes is that the year of "overuse" represents some true level of ability for the player. Typically, proponents center on guys coming off of good years. Regression should be expected of all young, relatively successful pitchers. For that reason, I believe it's a fluke that the VE candidates of 2009 have fared so well.

The VE ignores not only regression to the mean, but also entropy. Brian Burke, a former Navy pilot with a bachelor's degree aerospace engineering, is the author of the always-interesting AdvancedNFLStats.com. He's recently offered an illuminating take. on the VE.

"Injuries in sports are like entropy, the inevitable reality that all matter and energy in the universe are trending toward deterioration," Burke writes. "Players always start out healthy and are progressively more likely to get injured. Pitchers don't enter the Major Leagues hurt and gradually get healthier throughout their career."

VE proponents may be confusing the altering of probability caused by the alteration of the sample space with cause and effect. Burke applies the famous Monty Hall Problem to the VE. Note that Carroll vehemently dissents on Burke's message board.

You can see Burke's piece and the accompanying discussion and his follow-up piece for suitable detail.

Burke argues that if pitchers have a 1-in-5 chance of getting hurt and you take one healthy year away looking back at a five-year sample, the odds now are 1-in-4 that the following year is an injury year. Depending on how someone looks at the data, the supposed increase in injury risk may be merely an illusion.

Burke's point about entropy and pitching says that unhealthy is always more likely to follow healthy for pitchers regardless of whether they pitch 29 or 31 more innings than they did the year prior to the healthy season. The healthy year and next year are not independent like, say, two flips of a fair coin.

So what about Joba? It's a joke that the Yankees are being so cautious. The Red Sox have Bill James in their baseball boardroom. Note how they ignored the VE when it came to Jon Lester last year. If there was any statistical merit to the VE, James sure would know about it and Theo Epstein sure would listen.

The Red Sox must be laughing over this. There's no magical solution to the inevitable destruction of 95 percent of pitching arms. Chamberlain is very likely to have the same injury/poor performance risk in 2010 irrespective of the innings he throws in 2009.

It's not like all of this babying of young arms has decreased pitching injuries. You can look at historical disabled-list data from any angle and see that.

This is a healthy Joba year. He's finally in rhythm now and peaking. The Red Sox are looming and the defending American League champion Rays dangerously lurk in the shadows of the AL East.

When it comes to young, healthy arms, I say use 'em or lose 'em. Or, in Joba's case, quite possibly, "Use him or lose."

Michael Salfino is a nationally syndicated columnist and a regular contributor to SNY.tv.
Write a Comment! Post a Comment