The subhead for Michael Lewis's
review of Daniel Kahnemen's
Thinking, Fast and Slow in
Vanity Fair reads:
Billy Beane’s sports-management revolution, chronicled by the author in Moneyball, was made possible by Israeli psychologists Daniel Kahneman and Amos Tversky.
Then Lewis goes on to lament in amazement how could anyone have made such super gigantic mistakes as all the baseball executives in the history of the world until Bill James and Billy Beane came along, or something like that.
I'm a big fan of advanced baseball statistics, but its actual impact has been pretty marginal (other than the message: take lots of performance-enhancing drugs, but that didn't really require a degree in stats to figure out). Consider this shocking revelation I discovered in my voluminous readings of sabermetrics:
Q. Who was the greatest baseball player of all time?
A. Babe Ruth.
Of course, Babe Ruth was also the most famous and popular baseball player of all time, in real time. Baseball fans went nuts over him the moment he started hitting huge numbers of homers. Decades later, sabermetricians fired up their computers and figured out: hey, the bleacher bums were right.
Q. Who was the greatest American League player of the 1950s?
A. Mickey Mantle.
So, the nine year old boys of America were right in 1956. Mickey Mantle was enormously famous throughout his career. I suspect many of my foreign readers don't believe that Mickey Mantle was a real baseball player. They can vaguely recall the name, but he sounds like a fictitious folk hero, like Yankee Doodle or Jack Armstrong or Horatio Alger, made up to symbolize American post-WWII dominance. (Similarly, I suspect foreigners sometimes get Babe Ruth and Paul Bunyan confused.)
In other words, baseball fans's views were pretty accurate. And that's not terribly surprising: if you listen to most of your teams's games on the radio, much less have season tickets, you'll figure out pretty accurately who are the best players on the team and who are the worst. For example, sabermetricians like to claim that LA Dodger first baseman Steve Garvey was overrated because he was handsome but didn't get a lot of walks. Statistics prove, they like to say, that Reggie Smith was better than Garvey in the Dodgers' World Series years of 1977-78. Indeed, but Dodger fans knew that already. Fans at Dodger Stadium voted Smith the team MVP both years over Garvey.
What advanced statistical analysis did was improve what national baseball intellectuals had to say. What sportswriters had to say about their local team had always tended to be pretty reasonable: they watched all the games and, as Yogi Berra said, you can observe a lot just watching. But when sportswriters went to vote for the Most Valuable Player award for the whole league, since they didn't see many games played by the likely candidates, on other teams, they tended to overvalue dumb statistics like runs batted in. They felt they needed some statistical evidence to justify their votes, and it was traditional to overweight the RBI number.
Bill James argued that people who denounced his emphasis on statistics weren't free of statistics, they were just slaves to dumb statistics. But that was mostly true of their evaluations of players not on their own team. You know how at hockey games, the announcer comes on after the game is over and announces the third-best, second-best, and best player in tonight's game? They don't say that out-loud in baseball, but fans kind of do it in their heads, so if they listen to 100 games on the radio in the season, they have a pretty good idea in their head that in those 100 games Reggie Smith was the best player in about 18 of those games and Steve Garvey in about 12, so Reggie is the team MVP.
But league MVP awards are a pretty minor part of the game.
Similarly, when I was a kid in the 1960s, Babe Ruth was the most famous ballplayer ever, but it was a mark of intellectual sophistication to say that his homerun hitting was vulgar and that real experts all knew that the line drive hitting and base stealing Ty Cobb was better. But that's just mostly an epiphenomenon. Nobody benched Ruth because they didn't understand baseball statistics. They just watched games and Ruth clearly dominated over the other players.
So, while the rise of sabermetrics had some impact on how baseball was played (much of its impact malign), in the big scheme of things, it's pretty small change. The big impacts of better statistics are on the post-season awards and who gets into the Hall of Fame, not on the field of play.
Remember how it drove sabermetricians crazy in 2001 that sportswriters gave the AL MVP award to elegant Ichiro Suzuki rather than lumbering Jason Giambi, who almost died a few years later from all the PEDs he was taking in order to get his Billy Beane-approved surfeit of homers and walks? (For some reason, those statistical geniuses never tried to figure out which players were on the juice.)
For example, who were the best players on the
2002 Oakland A's, the team featured in
Moneyball, as ranked on Wins Above Replacement? Oddly enough, you can't answer that question accurately from reading Lewis's book. The main reason the team did well was little mentioned in the book: its three ace pitchers Tim Hudson, Barry Zito, and Mark Mulder were 15.4 wins above replacement. Three of the bad guys in the book for not taking enough pitches, Miguel Tejada, Eric Chavez, and John Mabry, were 10.8 wins above replacement. And the three heroes of the book, Scott Hatteberg, Chad Bradford, and David Justice, were 5.5 wins above replacement: useful acquisitions, but pretty marginal in the big picture of things, which is a pretty accurate evaluation of sabermetrics role in baseball: modestly useful.