Bill James mentioned recently that he didn't think the Catcher ERA discussion should be closed just yet. This made me wonder whether the year-to-year randomness in catcher ERA seen by researchers in the past was perhaps more due to the DIPS problem, i.e. that balls in play don't tend to be terribly well-controlled by the pitcher (nor by the catcher, you'd assume) and thus can cause undeserved fluctuation in a pitcher's runs-allowed record. Where a catcher could have effect, possibly, is precisely where the pitcher has effect: the DIPS stats. People talk all the time about catchers "framing" pitches, and it stands to reason that some catchers might be better at this than others, resulting in fluctuations for the pitchers in walks and strikeouts. How they might affect home runs is harder to see, but if a catcher can, overall, make a pitcher better, then they might help their pitchers avoid the long ball.

The process of getting some preliminary data for this study is tedious, but I did it anyway: I broke down the Boston Red Sox pitchers in 2004 by their catchers and calculated the DIPS ERA (using Larry Mahnken's worksheets) for each combination. The results are here. It should be noted that I couldn't figure out how to get the left vs. right checkbox to do anything, so that adjustment hasn't been made. The same goes for knuckleballers, which is particularly important in this case, since Tim Wakefield is in the data set. That said, the adjustment wasn't made for any of the catchers involved, or any of the pitchers involved, so comparisons between the data are still fair.

The usual weaknesses of this kind of study persist: the backup catcher (Doug Mirabelli) caught far fewer innings than the starter (Jason Varitek) did. That said, here's the list of pitchers where Mirabelli's DIPS ERA was lower than Varitek's: Adams, Arroyo, Astacio, Brown, DiNardo, Embree, Foulke, Kim, Leskanic, Lowe, A. Martinez, P. Martinez, McCarty, Mendoza, Myers, Nelson, Schilling, Timlin. Varitek wins with Williamson, Wakefield, Malaska, and Anderson. Four pitchers out of a shared set of 22.

Are we on to something? Maybe. It'll take someone smarter than me to really nail something down here, but at the very least, assuming someone else hasn't done this already and I missed it, hopefully I've created something we can run with.