Two Different Ways to See Melky Cabrera
Maybe the most important thing you learn early on in any basic stats class is that you can’t just throw out data. At least, not just because you want to, and not just because the data doesn’t fit. Just about all information is legitimate information, and you simply have to decide how heavily to weight it. Take Chase Headley, for example. No one figures he’s going to hit 31 dingers again, like he did in 2012, but the reality is that Headley did have a seven-win season just a few seasons ago, and we can’t justifiably ignore that. It’s a part of his record, and it hints at his true-talent level, or at least where it was in San Diego that one time. Because of that year, Headley gets a more favorable projection, and I don’t think you can argue that away.
If you’re going to eliminate data from a sample, you need to have a damn good reason. You need to be able to prove that the data is irrelevant. If you’re a research scientist, maybe the data came out of an experiment run you know you messed up. You accidentally buffered a solution to the wrong pH. As baseball fans, we’re not research scientists, but we’re still always looking for reasons to eliminate data. This is basically the same thing as having a disagreement with a given player projection. Overall, the projections do well, because they don’t eliminate data. But we’re always trying to beat them. Last year gave reason to eliminate prior data from J.D. Martinez. The White Sox saw reason to eliminate prior data from Zach Duke. And now the White Sox have also signed Melky Cabrera for three years and something like $42 million. Cabrera’s another interesting case, like Duke — he didn’t just overhaul his mechanics, but there’s something about his record that makes you wonder how much you should care about his 2013.

