Foghorn! Bill James Sounds the Alarm

Sabermetricians are in for a bit of a surprise when they open up the latest issue of the Baseball Research Journal. Bill James, dean of baseball historians, stat gurus, and mentor to Theo Epstein, tells us that many of the conclusions that he and others have drawn about the game are either wrong or not demonstrable given available facts and models of analysis. The reason? Data fog. Too many of these studies—including his own—are based on the comparison of statistics with an already high degree of randomness that when compounded become useless. “Random data proves nothing—and it cannot be used as proof of nothingness.” His revised analysis of some sabermetric “truisms”:

-There is no such thing as clutch hitting. Not valid. “This should be regarded as an open question.”

-Winning close games is all luck. Not valid. “Probably not all luck.”

-Catchers don’t affect pitchers’ ERAs. Not valid. “I don’t think there is a scintilla of evidence that this is true.”

-Baserunning does not affect run scoring. Not valid. “Probably not true, and that’s probably mostly my fault.”

-Individual batters have no tendency to hit well/poorly against lefties. Not valid. “Just wrong. My mistake.”

-A pitcher has no control over his hits/ip, except by striking men out and allowing homers. Valid. “May have been overstated…[but]…more true than false.”

-Players don’t get hot or cold. Not valid. “At this point, no one has a compelling argument.”

-A good hitter does not improve those around him. Valid. “I still believe this to be true.”

This may wrankle some within the sabermetric community and embolden “traditionalists” (who now have a pretext for scoffing), but in our estimation this latest article only reaffirms James’s position as baseball’s foremost thinker, and a master of scientific method.

0 comments… add one

Leave a Comment

Next post:

Previous post: