What are the limits of a "Moneyball" -- expected value of a particular position -- approach to personell decisions? A Sports Illustrated piece on Albert Haynesworth (A Titanic Force) made me think about this issue. A lot can be said for the idea. Even before Moneyball became a common term and Scott Pioli used it to build the Patriots, guys like former Titan GM, Floyd Reese, used an implicit version. He thought the supply of defensive tackes, for example, high relative to their team value so that paying more than, say, $2 million for one didn't make sense.
One difficulty -- a player whose skills are extreme. Now, it's not enough to know the average contribution of a player at a given position, but a GM needs information on the abnormal or "upper tail" of values. The complication -- the upper values differ by position in ways not necessarily reflected by averages. The relative supply and contributions of a cornerback may make him more valuable, on average, than a defensive tackle, but the once-a-decade DT with skills like Haynesworth creates havoc right in the middle of an offense on every play that even the most skilled cornerback cannot replicate. The Titans defensive coordintor, Jim Schwartz (btw, a user of statistical regression analysis) puts it this way:
In fact, the more times Haynesworth can collapse a pocket or disrupt a running play with his size, strength and speed, the better the Titans will be. "We try to funnel stuff back to him, to keep him alive on every single play," Schwartz says. "Albert's not a one-trick pony. He's not the guy who can't rush—the guy an offensive lineman can't move off the line of scrimmage, but he can't move off the line of scrimmage himself. Albert is a big man who can do a lot, and those guys are extremely valuable."
Ironically, Schwartz doesn't really seem to get the mean versus tails aspect of analyzing someone like Haynesworth:
Schwartz says he chuckles every spring when the media's mock draft boards are heavy on wide receivers and light on defensive linemen. "It should be the other way around," he says. "General managers, head coaches, position coaches, coordinators—we all know how important those guys are and how hard they are to find."
Typically, the media and guys like Floyd Reese are right -- DT of typical skill levels are not all that worthy of big expenditures, but that's not true of a Haynesworth. In the 1990s, the same could be said of Warren Sapp or of Joe Green in the 1970s and Bob Lilly in the 1960s.