Unfortunately, the page you’ve requested cannot be displayed. It appears that you’ve lost your way, either through an outdated link or a typo on the page you were trying to reach. Head back to the homepage or try searching the site below.
James adds insight as insider
by Alan Schwarz
When the Red Sox hired Bill James as a consultant several years ago, some complained that allowing an outside stathead influence over player moves would run the club into the ground. They don’t seem to be complaining anymore. James’ moving from the outside to the inside has had other effects, though—including a recent essay that repudiates some of his theories.
ALAN SCHWARZ: What can you learn from the inside of a baseball operation that you could never learn from the outside?
BILL JAMES: You see the reasons for things more. You learn how difficult it is to move players between roles if the players are uncomfortable moving between roles. You encounter a lot of barriers to action that I didn’t know were there. Sometimes those barriers are dollar signs, and sometimes they’re a lack of confidence on the part of the organization, or on the part of a key person in the organization that something will actually work.
AS: What kind of humility comes along with learning these things?
BJ: When people have strongly different opinions than mine on an issue, and you have to work together to reach some sort of consensus, then you have to try to understand the way they are seeing it. And sometimes when you shift over to their perspective on the issue, you see that you’ve been missing something. You have to take very seriously the possibility that you didn’t really understand the issue. As an insider you see the contribution of so many other people. If you look at the Red Sox, it’s not just the contributions of the current front office, but the contributions of the people going back a long, long period of time.
AS: You mean like Wayne Britton drafted Trot Nixon in ’93, that type of thing?
BJ: Exactly. You see those threads coming together from a million different sources, going back to the people who built Fenway Park. The success of the organization rests on the things done by an incalculable number of people over a very long period of time.
AS: Then again, it was the current people who decided to keep Trot Nixon—they built the club with their sensibilities.
BJ: Yes, the decisions that you make are very critical, but you can’t make good decisions unless you have good options—and we have been placed at a moment in history where we’ve had a lot of good options.
AS: I think that the most important divide in the area of statistical analysis is that some statistics do well describing performance and others do better predicting performance. Would you agree that a lot of the important work is being done on that second island of predictive statistics?
BJ: Going back to 1975, I was always focused on prediction. The runs created formula is a way of predicting runs scored from other offensive events. The Pythagorean relationship between runs scored and runs allowed to wins is a predictive relationship—if the relationship is out of whack in the past, you can predict that there will be an adjustment in the future. So I think that I’ve been concerned about that since I’ve started doing this, and I still am.
AS: You recently wrote an article in the Baseball Research Journal called “Underestimating the Fog,” in which you say that many of your longtime conclusions—that the ability to hit in the clutch or win one-run games are not real traits among players or teams, for example—are actually invalid. Can you explain how you came to writing that essay?
BJ: The point of the article in the simplest terms is that a method that we sabermetricians have used for over 25 years to distinguish between what is real and what is a statistical illusion, I have realized, often does not work at all.
There is an analysis that attempts to prove that there is no such thing is clutch hitting ability, and that analysis is as follows: If you look at the guys who hit well in the clutch one year, and the guys who hit poorly in the clutch the same year, then you look at what they did next year, there is no difference. There is no predictive significance to clutch performance deviation. And I thought that that was true for 25 years, but I have now realized that the way we were studying the issue, the way we were looking for consistency, is very prone to error. A lot of consistency could be there that we’re not seeing because of randomness.
AS: What you’re saying is that even if we feel like we have an adequate sample size—100-200 at-bats, say—while looking at hot and cold streaks or clutch hitting, there are still too many variables at work to make any confident statements about any specific one? That clutch hitting could still be a batter’s inherent skill but it’s drowned out by all the other randomness?
BJ: We all know that baseball statistics are fraught with random variations that look meaningful, but aren’t. But we thought that we had a filter for these problems by looking at large numbers of small samples, and what I am concluding is that that filter doesn’t work. We thought that randomness was a proof of nothingness, when in fact it proves nothing.
AS: It’s interesting that you seem to be pulling off a belief that’s held by outside-type people and going with the view of the insiders. Could you have come to the conclusion about clutch hitting, for example, without working for the Red Sox?
AS: Would you have?
BJ: I don’t know. It is certainly related to Red Sox problems. But it is also related to the problem of catcher ERA. Studies of consistency and the impact of a catcher on the pitcher’s ERA have led to the false conclusion that a catcher doesn’t have an observable impact on the ERA of the pitchers he works with. And before I went to work with the Red Sox, I realized there was something wrong with that analysis.
Working with the Red Sox, I’ve realized that I was having the same problem with clutch hitting and the same problem with platoon differentials for individual batters. So putting it together, I realized that there was a general problem as opposed to a specific problem.
AS: What has the reaction been from people who take you very seriously, to have you repudiate many of your oft-stated beliefs?
BJ: I think more puzzled. People want experts. When I set out to do this 30 years ago, I was not trying to be an expert—I was trying to say that rather than sorting out these questions by relying on expertise, let’s rely on the facts, look at what the facts themselves are telling us. The irony is that by doing that, I became a sort of expert.
AS: Now that you have an inside expertise, you’re saying the insiders have a point!
BJ: An expert worries a lot about consistency, but a scientist understands that a lot of the things he believed 20 years ago were just wrong for one reason or another—and the fact that you said something 10 years ago doesn’t mean it’s true. I was wrong about a lot of things 10 years ago. I am wrong about a lot of things now. The facts are still more important than what I thought, just as they are more important than what somebody else thought.
AS: You know, Einstein basically said “never mind” to some of his theories, and he ended up having been right in the first place.
BJ: (Laughing) I have found it’s not a good idea to compare myself to Einstein.
You can reach Alan Schwarz by sending e-mail to firstname.lastname@example.org.