October 23, 2006
World Series Prospectus
Game TwoKenny Rogers is scuffing his pitches, but he only bothers to do so when he pitches at Comerica Park. That's why Jim Leyland has engineered his rotation throughout the postseason to ensure that Rogers pitches at home.
Think this through for a moment. Suppose that Rogers is cheating. Why would he only cheat at home? The usual ingredients that one uses to scuff a pitch are dirt, rosin, and spit. Maybe pine tar. Last I checked, those items are as readily available in Yankee Stadium or Network Associates Coliseum as they are in Detroit.
Put that aside and give the theory its day in court. If a pitcher were doctoring his pitches at home--but not on the road--what evidence would we expect to see? I'd think we'd expect to see two things: more strikeouts at home, and less good contact when the batter does manage to put the ball into play. Let's examine the evidence on each of those things.
Rogers' walk rate is slightly worse at home than on the road (8.1% versus 6.6%). His groundball-to-flyball ratio is almost exactly the same at home and on the road. He gave up about the same number of singles at home and on the road. It's the strikeout rate and the "good contact" categories where we see the disparities in his splits.
These differences are right on the verge of what we data geeks like to call "statistically significant." In other words, we can conclude that Rogers' strikeout rate, in all probability, really and truly is higher at Comerica Park, and his "good contact" rate really and truly is lower at Comerica Park. It isn't just a random fluke.
What we don't know is why Rogers' numbers are better at Comerica. Park effects are almost certainly part of it. Comerica reduces home runs by about 20%, and doubles by about 8%. However, the differences in Rogers' rates remain statistically significant even after adjusting for these effects.
Meanwhile, some pitchers are simply more comfortable pitching at home. Rogers had a better ERA at home than on the road in 2005, 2004 and 2002, when he was pitching in the hitter-friendly Ballpark at Arlington. The same was true in 2003, when he pitched for the Twins. So these differences aren't just something that surfaced overnight, and they don't seem confined to Comerica Park. A conspiracy theorist, of course, could spin this evidence his own way: not only is Rogers cheating, but he's been doing so for a long time. And that cheating somehow requires the help of a clubhouse attendant, groundskeeper, batboy, whoever, who can aid and abet the cheating at home, but not on the road.
Let me put this as carefully as I can. We cannot reject, on the basis of the statistical evidence, the claim that Rogers is cheating at home. But the statistical evidence does not prove that he's cheating, either. You wouldn't hang a man with this stuff.
There was an outtake circulating around the press box tonight that showed Rogers with similar dirt patterns on his pitching hand in his ALCS start against Oakland. That, coupled with the statistics, makes Rogers look pretty guilty. But it's still circumstantial evidence.
So what I'd propose is this. Hire a crack team of physicists, from MIT or Stanford, and give them the FoxTrax data. Have them see if Rogers' pitches are actually moving in ways that it is physically impossible for a normal baseball to move. Perhaps in 30 years' time, we'll have the science all taken care of, and there will be three lights behind home plate: red for strike, green for ball, yellow for spitball.
If the Tigers were a species, they would have very low genetic diversity. In other words, most Tigers hitters are remarkably similar to one another. Virtually all of them can be placed in one of two groups:
So when one Tiger hitter likes to hit against a certain type of pitcher, chances are that a number of other Tigers are going to like to hit against him as well. Conversely, when a pitcher like Anthony Reyes seems to have the Tigers' number, the problems are likely to run throughout the whole lineup.
The Tigers, for having a pretty good offense overall, have had quite a number of occasions on which the offense has been effectively shut down. The team scored two or fewer runs on 38 occasions during the regular season, and three or fewer on 63 occasions. As we noted, these numbers are higher than usual for a team with this amount of run production.
There's not much that the Tigers can do about this for now. But in planning for next season, the Tigers ought to be targeting players who break type.