I've done what I did with the Red Sox for the A's. The results are here.

The recap:

Miller wins with Blanton, Duchscherer, Harden, Harville, Mecir, Redman, Rincon, and Saarloos.

Melhuse wins with Bradford, Dotel, Garcia, Hammond, Hudson, Lehr, Mulder, Rhodes, and Zito.

So the initial score is 9-8, Melhuse. We're going to throw out Blanton, though, because he only threw to 30 batters all year. Garcia only got a brief callup and faced just 32 batters, so he's out. Finally, Harville gets the same treatment: he was traded after facing just 11 batters.

With those revisions, then, we get an 8-6 edge for Melhuse. Combine that small edge with the fact that the differences in Redman, Mulder, Hudson, and Bradford's dERA's are pretty small (note that three of those fall on the Melhuse side) and we can conclude ... not much.

If we knew that catchers really did have an effect on pitchers, we'd conclude that Melhuse and Miller were about equal last year. Unfortunately, we don't know that, so this only adds a data point to the "random" side of the debate.