A new Major League Baseball season began this week, and we'll soon find out which general managers were smart in their off-season acquisitions and which ones messed up. The pressure is on, especially as statistical analysis developed by Bill James and practiced most thoroughly by Oakland General Manager Billy Beane gives teams sophisticated ways to examine player performance.
Lots of teams, though, seem to misuse stats. A case in point: Two years ago, Beane's key assistant, Paul DePodesta, became general manager of the Los Angeles Dodgers and found his team in first place within its division halfway through the season. Podesta's computer, though, suggested that he could massively improve the Dodgers by trading away his starting catcher and others. Podesta did so despite warnings that he was tampering with "team chemistry," which some stats determinists see as a mystical notion that is unmeasurable and therefore irrelevant.
Result: The Dodgers staggered during the rest of 2004. After more stats-based signings of free agents not known for being team players, Los Angeles fell last season to 71 wins and 91 losses, the team's second worst record since 1944. Dodgers owner (and real estate magnate) Frank McCourt Jr. fired DePodesta. His wife, Jamie McCourt, who earned a B.S. degree in French from Georgetown University, is now president of the team.
How should general managers (GMs) use stats? I stopped in Orlando, Fla., last month and asked John Smoltz, the smart Atlanta pitcher with 17 years of Major League experience. Inside the Braves spring-training clubhouse, he explained how stats need to be augmented by a searching analysis of character.
As Smoltz put it: "Some GMs see a player with all the tools and go wild, but the tools will be misused if the character isn't there. ... It amazes me that some executives are up there cruising at 30,000 feet instead of doing research at ground level to find out what's in the hearts of players. ... They hand out huge multiyear contracts without knowing enough about the players and the families to determine whether a player given all that money and security is going to get lazy."
What should a GM look for? According to Smoltz, he should see whether a player loves baseball (so that he's willing to work hard at it) and has successfully "gone through trials and tribulations, because that's how we improve."
The most useful stats, he says, are those that show how a player performs in clutch situations: "I want gamers, guys who will lay it all on the line under pressure. I want to know whether a player is too high-strung, too wired-up, or whether he is calm under pressure."
Overall stats can be deceiving, Smoltz said, because "a guy may be impressive in relaxed situations, only to choke when you really need him." Team-spiritedness is also important, but impressions based on image can be wrong: "Players sometimes get negative because they're trapped in a bad situation. Sometimes they learn from that experience and change. ... You need to stay away from the troublemakers, but don't assume that a guy who's had trouble will always be a problem."
Patience is also crucial, but many managers and GMs "panic easily. They push a guy off a building if he's in a slump."
All of that sounds like not only a prescription for general managers, but a set of principles for managing generally: Use stats, but not stats only; examine hearts, and not just arms and legs; judge, but don't be judgmental; look for gamers who love the game and have learned from it. And, in examining both players and politicians, be skeptical about "a guy who brags a lot," as Smoltz notes, for players "who have lots of words often don't match that with their deeds."
