SkyKing162's Baseblog

A fan of the Yankees, Red Sox, and large sample sizes.

Back to baseball, like I promised...

People are slowly catching on to the whole DIPS thing. Even mainstream columns are using BB rates, SO rates, and HR rates to analyze pitchers. Two bad things, though - they're also using H rates, which is often meaningless, and they're using 9 IP as the denominator in the rates (BB/9, SO/9, HR/9). I want to rant about the second problem right now.

Innings Pitched is not the best measure of a pitcher's playing time - batters facing pitcher (what we call plate appearances for hitters) is the best. IP is dependant on how many batters a pitcher faces get out, which is dependant on the pitcher AND the defense AND the ballpark. If a pitcher has a lousy defense behind him, or an extraordinary number of hits drop in just by chance, the IP total for that outing will be lower than it "should" be. If Bob the Pitcher faces 24 batters, strikes out 8, walks 0, gives up 0 homers, induces 0 GIDP and has 8 of those other 16 batters get out, he'll have pitched 5 2/3 innings and struck out 8 for a ratio of 13.5 K/9. If however, he's a little luckier and retires 12 of the 16 other batters, he'll pitch 6 2/3 innings and have a ratio of 10.8 K/9. If you buy into DIPS, in which game did the pitcher pitch better? Neither, the differences are a result of defense and luck. So why would we want to represent his performance with the stat K/9 when it can fluctuate for no good reason? I sure wouldn't.

The worse the defense playing behind a pitcher, and the unluckier a pitcher is with hits falling in, the higher his BB/9, K/9, and HR/9 stats will appear, artificially improving K/9, and artificially worsening both BB/9 and HR/9. Instead, everyone should be using the raw DIPS ratios: $BB, $SO, and $HR, which are NOT dependant upon defense and luck.


Sure, they take a little getting used to, but well worth it. And you should be glad I didn't get started on H/9...

Comments: Post a Comment