June 3, 2013
Why Sabermetricians Should Watch Their Language
Smoking cigarettes is bad for your health. This is not a secret. Six-year-olds know that. And smoking is not just bad, it's really super-duper bad for you. Yet, tobacco companies aren't going out of business any time soon. There are millions of smokers in the United States and more start each day, even knowing all the risks. Why?
I don't know whether Seattle Mariners manager Eric Wedge smokes, but he inadvertently illustrated that sabermetrics has the same problem as United States culture. (Full disclosure: I began doing some freelance work with the Cleveland Indians in 2009, which was also Wedge's last year as manager of the Indians. We never spoke.) Last week, in discussing the troubles of Dustin Ackley's 2013 season (and let's face it, his 2012 season as well), Wedge dropped in this curious statement:
"Oh, yeah, I have. It’s kind of the new generation. It’s all this sabermetrics (pause) stuff, for lack of a better term. You know what I mean? People who haven’t played since they were nine years old and think they’ve got it figured out. It gets in these kids’ head [sic]."
Okay, we all had a good laugh at the thought that I (and others like me) somehow destroyed Dustin Ackley (sorry!) Let me attempt to be the adult in the room and take the statement at face value. Earlier in the interview, Wedge suggested that Ackley had become too passive at the plate, not swinging early in the count, and letting himself fall behind too often. He seems to believe that Ackley had heard/read too much about Sabermetrics, particularly the stuff from 10 years ago (i.e., Moneyball) where walks were the Most. Important. Stat. Ever.
(Side note: The lesson of Moneyball was not that if you want to be a good player, you should try to walk more. It was that there were certain players who were already good at (among other things) drawing walks, and this provided value that was not being recognized because batting average treated walks as a non-event. Walk rate is not the cause of success. It's an indicator of a successful approach that was already there.)
As a result of ingesting all of this information, Ackley apparently changed something about his approach in an attempt to become something that he thought he was supposed to be, and in doing so, messed everything up. In fact, Wedge said this the next day:
“The internet and everything else, the information that’s out there, they’re human beings, too. If you’re on it, they’re on it, too, I’m sure. You hear all the baseball experts say you’ve got to do this, you’ve got to do that. Again, there’s a way to go about doing it where you can have the best of both worlds. You’ve got to be ready to hit. You can be both ready to hit be disciplined at the same time. That’s the mental approach.
“I’m all about getting on base, but I’m about hitting, too. People have to understand: You can’t go up there looking for a walk and expect to be a big-leaguer very long. Nobody’s stayed up here by just walking. You’ve got to hit, too. You can get deep in the count all you want, but eventually you have to hit. It’s just not a black and white thing like some people think. I can’t explain it any better than that.”
Wedge actually makes a good deal of sense here. The whole episode brings up an interesting question that sabermetrics has never really asked itself. It's not that any of the information that Ackley got his hands on was wrong. A walk certainly isn't a bad outcome of a plate appearance, and it has been very undervalued historically. A walk is better than a strikeout, and if Ackley (or anyone) can replace some of his strikeouts or ground outs with walks, he'll probably come out ahead in the deal. Assuming that the problem really is that some misplaced focus on walks (sabermetrically-induced or not) messed up Ackley's head, how did accurate information produce such a bad result?
Perhaps the worst thing that one can do when trying to help someone quit smoking is to make the matter about quitting smoking. Focusing on the behavior alone often gets you nowhere. Plenty of people who smoke know full well that they should quit, and they want to quit. But they don't. Quitting smoking is only tangentially about the smoking. What actually has to change are all of the structures in the person's life that support smoking. Smokers often structure their days so that they can have smoke breaks. They make friends with other smokers and bond with them over cigarettes. You have to change the whole environment, not just one decision.
There are plenty of noble, but misguided, people who begin to try to help a friend quit smoking by finding some inspirational story about someone who was able to quit cold turkey. If that guy can do it, you can too! It's such a seductive message, but it will probably backfire. Most people are not able to quit cold turkey, and when the friend who's attempting to quit has another cigarette, he might very well become discouraged and say, "Well, I guess I'm just not good enough at quitting."
Again, the message isn't incorrect. If one is able to quit, that's a good thing. It's just not that easy. There are plenty of other variables to address. Information can be misinterpreted, and things that seem sensible can backfire. It's not enough to have good information. It has to be presented in a way that makes the person want to make a change. How does one get from "smoking is bad from you" to the actual behavior of becoming smoke-free?
I don't believe that the sabermetric movement has properly grappled with this issue either. Thankfully, sac bunts and three-run saves don't increase the risk of heart disease, but if there's one thing that sabermetricians like to whine about (myself included), it's that even when we show the run expectancy matrix and control for all the confounding variables, people don't buy the overall idea that we're trying to push. Maybe worse is that sometimes the conclusions are misinterpreted (and if the Ackley story is true, that would be a good example).
Are we presenting the information to people in a manner that speaks to the way in which they understand the game? If not, can we really be surprised when our theories aren't more broadly accepted? The way in which a message is presented tells a lot about the way that the speaker (or writer) conceptualizes the problem. When Sabermetricians speak, we often frame our answers to questions in terms of numbers. When I hear scouts speak, I often hear them talk in the language of geometry (arm angles and planes and such), and it's clear that many of them see players in terms of mechanical principles. The body parts and motions fit together to produce a result. It's all very visual-spatial, and my wife will tell you that I have no visual-spatial intelligence. And there are times when scouts get really into it that it sounds like it's in Russian. I have to wonder if I sound the same way.
This is why I would propose that sabermetrics needs to engage in something that's known as translation research. If we believe that the ideas that we have are good ones (and we should always be humbly open to the idea that they might not be), and we want people to change in some way as a result, we need to take it upon ourselves to present the information in a way that can be easily understood by the people whom we want to reach. It's not a matter of dumbing things down or doing shoddy work. We need to think more clearly (and perhaps research) who needs to hear the information that we discover, and beyond just showing the run expectancy chart, how to present it so that it makes the most difference.
It sounds like Eric Wedge believes that Dustin Ackley would have been better off not hearing about the sabermetric fascination with walks. He might be right.