MLB FanHouse Q&A: Bill James
FanHouse: How confident are you in the advanced defensive metrics relative to the offensive metrics?
James: Well ... they're not the same. We have a long history with batting statistics, which gives us an intuitive footing for making projections based on them. Let's say that a hitter has a breakthrough year, like Garrett Jones last year with the Pirates or Billy Butler with the Royals. You kind of know intuitively how likely that is to be real, and you can build on that with research. But with defensive metrics, you don't know intuitively what to make of Franklin Gutierrez' numbers ... somebody has a good year or a bad one, you don't have the same sort of frame of reference to build on it. We're gaining confidence, but it's not the same.
FanHouse: What will be the next big breakthrough in sabermetrics?
James: Physiognomy.
FanHouse: What are the most undervalued and overvalued skills in baseball?
James: A good example of "starting the discussion on page 20." The question is, what is the value of the skill, not "what is the value of the skill relative to what people think it is," because everybody has a different view of the value of skills, and there is no way to generalize accurately about what other people value.
FanHouse: Where do the Red Sox rank from a sabermetric standpoint within the industry?
James: No idea.
FanHouse: What was your role in the Sox new run prevention acquisitions?
James: Can't answer it.
FanHouse: Do you still believe pitching/defense to account for roughly 52 percent of team wins and if not where does that number now stand?
James: The 52 percent came from the Win Shares book, and that actually resulted from an accounting error on my part. In the win value schematic that I had set up, I was trying to account for pitcher's hitting within the framework of pitching. I figured if one pitcher allows five more runs than another but also creates 5 more runs with the bat, they're even, right? So why not account for pitcher's hitter as a part of pitching?
But what I didn't understand until I finished the book was that, by doing this, I was overstuffing the pitching area. Pitcher's hitting is not an insignificant element of the game. If you figure that hitting is 50% of the game, then pitcher's hitting is 3-4% of the game. If you account for that 3-4% as pitching when it is really hitting, that causes you to have too much value in pitching and not enough in hitting--in other words, it causes a 50-50 split between offense and defense not to work. I didn't understand that at the time; I knew that the 50-50 split wasn't working, but I couldn't figure out why. Or didn't figure out why until after I had finished the book.




