By Ashley Yeager
The question, and response, prompted Mukherjee, a statistician who studies machine learning, to launch into a discussion about intuition and statistics in sports, specifically in baseball. Mukherjee spoke on June 12 as part of Periodic Tables: Durham’s Science Café.
He admitted he was a Yankees fan, which elicited some booing. Laughing it off, he then showed a complex statistical equation his colleague, Shane Jensen at the University of Pennsylvania, and others use to calculate a player’s success at fielding ground and fly balls.
On the next slide, Mukherjee showed the results. Rodriquez was clearly on top, and Jeter closer to the bottom. “Jeter doesn’t have as big a range as other players, that’s all I’m suggesting,” Mukherjee said.
Of course, these statistics, called sabermetrics, aren’t new to Jeter and other players. The numbers, based on Bayesian statistics, are exactly what the Oakland A’s baseball team used in 2002 to build a winning team. And, when new numbers came out in 2008, the stats ranked Jeter fairly low as a defensive player. He responded by saying there was a “bug” in the model.
“He has a point. The exact conditions for each play are not the same, so it’s hard to truly compare them,” Mukherjee said. The equation, however, is a way to measure factors of the game, rather than rely on intuition, and statisticians are trying to add more factors to make the model more realistic. The next factor they want to add will account for the different designs of ballparks, Mukherjee said.
He added, though, that these stats don’t really put players’ jobs at jeopardy. Judging by the crowd’s first response, people obviously still rely on intuition when it comes to picking their favorite players. The cold, hard numbers therefore affect how players approach their game – ie Jeter’s post-2007 season focus on a training program to combat the effects of age, Mukherjee said.
The data also affect people betting on the games. “Betting is huge, in any sport,” Mukherjee said, and the numbers, it seems, can affect how people choose to risk their money, but not their team loyalty.