In the last post, I mentioned some pitchers that had taken good advantage of their opportunities to face other pitchers in the batter’s box. I don’t have any information on how repeatable a skill, if you even want to call it that, dominating other pitchers is. It doesn’t strike me as an independent skill like getting ground balls. Rather, I imagine there’s a giant correlation between how good a pitcher is against all hitters and how good he is against pitchers.
So how about the other side of the coin? Who are the pitchers who performed poorly against other pitchers, relative to our expectations? Daniel Cabrera and Scott Olsen are two big outliers each only striking out three pitchers in 26 attempts, over five fewer Ks than average. Of course, both Olsen and Cabrera were mediocre in getting strikeouts in general, so it would be unwise to call them unlucky by the full five strikeouts.
In order to figure out roughly how many strikeouts lucky or unlucky a pitcher was, I devised a quick formula based on the following variables and stats:
KH = Percentage of non-pitcher hitters struck out
PF = Batters faced that were pitchers
KP = Number of pitchers struck out
17.5% = League average percentage of non-pitcher hitters struck out
32% = League average percentage of pitcher hitters struck out
The formula is:
( ( ( KH / 17.5% ) * 32% ) * PF ) – KP
Here’s an example sticking with Scott Olsen above. He struck out 14.8% of non-pitcher batters, which was 85% as good as the league rate. Therefore, I expect him to strike out pitchers at a rate that is 85% as good as the league rate, giving us a 27% expected K rate on pitchers. He faced 26 pitchers so we expect seven strikeouts (26 * 27%), but he actually only got three so in this case we would have expected Scott Olsen to record four more Ks than he did.
Is this perfect? Of course not. If I were constructing this for serious study, I would need to see if this “skill” is actually independent or not, regress the strikeout rates and do some other tweaks. I’m not proposing a modification to tERA or WAR though. I wanted an easy way to rank pitchers “luck” on this matter. Again, this is not perfect.
Using this as a guideline, the same four pitchers I mentioned before appear at the top of the “luckiest” rank with the addition of Roy Oswalt who killed opposing pitchers but was below average against all other hitters.
Looking at the other side, the most unlucky pitcher under this rank was Tim Lincecum. The Cy Young Award winner had a strikeout rate nearly double that of the league average but was only about average against other pitchers, suggesting that he missed 12 expected strikeouts. Other pitchers breaking the five strikeout barrier were Jason Hammel (9.5), John Johnson (9.5), Ricky Nolasco (8.2, boy is there anything this guy was lucky at last year?), Chris Carpenter (7.9), Carlos Zambrano (6.0) and Chad Gaudin (5.3).
Hard numbers? No, but fast ones and good enough to paint the broad strokes of a picture.
Matthew Carruth is a software engineer who has been fascinated with baseball statistics since age five. When not dissecting baseball, he is watching hockey or playing soccer.