Just yesterday, we were talking about Bartolo Colon's "radical" surgery that may have helped get his career back on track. We were talking about it because Bartolo Colon's career is back on track. Nobody would care had Colon had the procedure and failed. Nobody would care had Colon had the procedure and stayed in the Dominican. But Colon had the procedure and wound up with the Yankees, for whom he currently owns a 3.86 ERA, pitching well out of a rotation hurt by Phil Hughes' mysterious arm death. Colon's comeback hasn't been a miracle, but it's been about as close as you can get, and it's taken place on a big big stage.
Colon started out by pitching well out of the bullpen. As if that weren't good enough, though, he subsequently made the switch to the rotation, and here's what he's done over his first four starts:
6.2 innings, 2 runs, 7 strikeouts
8 innings, 1 run, 6 strikeouts
7 innings, 3 runs, 7 strikeouts
4.1 innings, 5 runs, 4 strikeouts
26 innings, 11 runs, and 24 strikeouts, with only four walks to top it all off. Despite the most recent clunker, Colon's numbers look very promising, and suggest an improbable return to a high level of ability.
But there's one thing about his numbers that catches my eye. It's also caught the eye of Dave Cameron and Matthew Carruth. Over those four starts, Colon has struck out 23% of the batters he's faced. But he's also posted a contact rate of 84%, which is worse than the league average. Those numbers don't seem to mesh.
Indeed, contact rate and strikeout rate are very closely related. There were 210 starting pitchers who threw at least 400 innings between 2002-2010. I plotted their strikeout rates versus their contact rates, and here's what I got:
There's a clear, strong correlation between the two. As contact rate increases, strikeout rate decreases rather sharply. It's an obvious point to make, but hey, a picture! People like pictures.
We can use this data to calculate an expected strikeout rate based on a given pitcher's contact rate. Not every pitcher will fall right on the best-fit line, of course, but they don't fall very far away. The biggest positive difference belongs to Mark Prior, who posted a strikeout rate of 27% against an expected strikeout rate of 22%. The biggest negative difference belongs to Matt Clement, at 21% and 26%, respectively. At the extremes, we see differences of about 5%.
Bartolo Colon's current strikeout rate as a starter is about 23%. Bartolo Colon's expected strikeout rate, based on his contact, is about 14%.
It's evident that Colon's strikeout numbers and contact numbers don't match up, and given that the latter tends to stabilize much faster than the former, we're left with the indication that Colon's strikeout numbers are unsustainably inflated.
The key? Of Colon's 24 strikeouts as a starter, 14 have been called, and ten have been swinging. The league average is that more than 70% of all strikeouts are swinging. Over Colon's career, 61% of his strikeouts have been swinging, so he's proven himself to be a little different than usual, but not to this degree. Colon's strikeouts are inflated because he's been generating an unsustainable number of called strikeouts.
And so we can expect his strikeout rate to regress over time for as long as he remains in the rotation. Will it regress to the point at which he becomes a problem? I doubt it. Even a regressed version of Bartolo Colon is still a big league-caliber starting pitcher. But one notes that, between 2007-2009, he whiffed 141 batters over 200 innings and posted a below-average ERA+. A healthy Colon can be useful, and the current healthy Colon might be a little more than that, but early success aside, Bartolo Colon is not a guy the Yankees will want to be leaning on down the stretch and in the playoffs. Overachievers make for heartwarming stories, but math doesn't lie, and bumps lie ahead.