<not_sure_if_serious.jpg> But seriously, I have always found it interesting that baseball is full of averages (means) like BA and ERA, but statisticians in other walks of life use additional measures like variance or root-mean-square (not to mention kew and skurtosis etc) to answer exactly the question we're asking here. If pitchers always pitched 9 innings, then we could easily distinguish two 4.00 ERA pitchers, one who always pitches 4-run complete games, the other who alternates between shutouts and 8-run debacles - one has a RMS of 0 while the other has 4. (Which pitcher I'd rather have depends, perhaps on whether my offense averages more like 3 runs a game or 5, but the information seems worth having in its own right.) But I don't even know what computation to suggest for the actual case where starters pitch unequal numbers of innings from one game to the next, and where a 0.1 inning disaster by a reliever is to be weighed against several clean outings of exactly 1 inning each. Is Belisle's performance typical or not? I don't know of a handy dandy measure anyone's ever come up with for that. My own view would be to move away from ERA, and toward OPS-against, as a starting point for pitching effectiveness. Might have a better shot at comparing and correlating groups of games using that. Belisle's season OPS-against is .715, which is better than league average. Last night's outing was somewhat higher though.