February 24, 2004
Sabermetrics vs. Psychology
The other day I referred to an article about the Orioles use of psychological tests to help them select players. In the article, they talk about Arthur Rhodes:
Meanwhile, Flanagan had become the Orioles' pitching coach. In 1995, while struggling to find a role for talented, inconsistent Arthur Rhodes, he sought from Ritterpusch's advice. He said Rhodes' psychological testing indicated he should neither start nor close, roles the Orioles wanted him to fill.
Rhodes was made into a middle reliever and setup man, and is still pitching in the major leagues today.
So what is Arthur Rhodes going to be doing this year?
Closing for the A's. What Beane sees is someone who strikes out a lot of batters, doesn't walk too many, and keeps the ball in the park. In other words, he does the three things pitchers can control very well. So Rhodes will be an interesting data point for a sabermetrics vs. psychology debate.
Does talent matter more than makeup? In the past I would have said yes. But I'm willing to be open minded about the subject. Rhodes won't prove anything by himself, but we will be able to watch the Orioles starters, closers and middle relievers from now on, so we should get a good idea of how this theory plays out.
Posted by David Pinto at
08:46 AM
|
TrackBack (0)
I doubt the psychological tests did anything but confirm the suspicions Gillick and crew had about Rhodes as an Oriole. In the late 90s, he often whined about the way he was used. He blamed poor performances on his being brought in at irregular times. He expressed a preference for middle relief and set-up - sixth, seventh, and eighth-innings only please. When Gillick went to Seattle, he brought along Rhodes to set-up Kaz. When Kaz got hurt, Rhodes expressed no interest taking over the closing duties. When Rhodes did poorly, he complained about the way he was used and talked about how he didn't like to pitch when he was tired out. He also whined about not making the All-Star team for his killer numbers as an eighth-inning left-handed set-up guy.
I expect Rhodes to scuffle as the closer unless he's babied. He won't be able to handle all the save opps. If he gets "tired out," he won't pitch well and he'll blame it on his handlers. He's suited for closing in that he won't be able to blame irregular use for his struggles. He's not suited for closing in that he's not a pitcher who can go three days in a row or ten times in twenty days without feeling injured or abused.
I'm not sure what the psychological profile says about Rhodes, but he hardly fits the stereotype of today's closer. And good for him. It's a stupid idea to have one reliever handle all the save opps, and if his character provides an excuse for the A's to do otherwise, then good for the A's. From a business perspective, it's even better if Rhodes will leave less than half the saves opps to Bradford and thus reduce Bradford's salary for 2005.
Although I believe that the psychological tests are accurate I'm not so sure about their longevity. Can a pitcher's "makeup" change after a few years of success or failure? I think it can. The tests may be useful for the here and now but I firmly believe that a pitcher can step into a new role and conquer that which had previously held him back.
With Rhodes though, I don't think he has changed. Nothing that I know of in his career has prepared him for making the jump from lefty setup to closer. He has been babied and he has enjoyed the success that he has found in the role he has had over the last few years. I don't think he wants to be a closer in his heart of hearts because he doesn't want to fail again where he has once failed before. It's a shame Beane didn't think of that and had him tested again. I don't think there is any doubt that he'll shuffle between closer and set up and be marginal at both. However, if he does find some success at closing then perhaps he'll take the first steps in altering his "profile".
I guess Alex beat me to it. It's not clear how permanent these psychological traits are. While I suspect certain aspects of our character/psychological makeup probably don't change much over time and therefore the test's predictive ability likely holds over time, I think several of the domains measured may change in response to experience. It probably argues for testing people in the organization over time, not necessarily annually but at least a second test sometime in their lifetimes.
The dichotomy you refer to is known as the state/trait dimension. The issue is whether given behaviors or personality types (depending on your theoretical persepctive) are enduring or transitory?
One way to increase the predictive capacity of these tests would be to assess the subject's ability to use mastery to overcome his psychological makeup in a variety of contexts.