Posted by Andy on January 9, 2010
OK, so in Part 1, we looked at the number of innings pitched for every start in the majors in 2009. Here, I want to break down that data by league.
The graph is as follows:
You'll notice that the y-axis is now percentage of starts instead of just total number of starts. This is because there are two more NL teams than AL teams, so comparing raw numbers is deceptive. By changing the numbers to percentages we can easily compare them on the same graph.
So what do we see? Here are some thoughts and conclusions:
- First a point of clarification. I did these searches simply by looking at starts by AL pitchers and starts by NL pitchers. My conclusions below have a lot to do with whether the pitchers are batting or not (which is of course the difference between the two leagues) but I did not make any consideration of interleague games. Therefore among the AL starts are a handful where the pitchers were subjected to NL rules, and the opposite is true for a handful of NL starts.
- Anyway, let's look at what happens in innings 5, 6, and 7. We can see that starts lasting those exact number of innings, there are more such starts in the NL than in the AL, and fewer starts in the NL that include fractions of innings. (I mean we see more NL starts that go 5, 6, or 7 innings, but fewer that go 5.1, 5.2, 6.1, 6.2, 7.1, or 7.2 innings.) This has got to be the influence of the DH. In the NL, a pitcher is a little bit more likely to be taken out between innings if his spot in the batting order comes up in a key situation. Therefore a pitcher is more likely in the NL to be taken out of a game a bit prematurely since his performance on the mound is not the only factor the manager has to consider. It may also be the case that the NL pitchers are more likely to be left in during an inning if his spot is due up the following inning because the manager doesn't want to waste a relief pitcher, i.e. bring in a guy who he'll have to immediately pinch-hit for. I think this pretty well explains what we see in innings 5 through 7.
- In the NL in 2009, starters pitched 15067.2 innings over 2590 starts, or 5.817 IP/start. In the AL, it was 13189.2 innings over 2270 starts, or 5.810 IP/start. So essentially the numbers were identical. This suggests to me that the DH effect is pretty neatly balanced. In other words, some NL starts are shortened by the pitcher coming out for a pinch-hitter while some are lengthened by the manager not wanting to have to waste a reliever. It may also be the case that NL pitchers can go a little bit longer in some games when they come out a bit earlier in other games. In any event, these effects seem to even out over time.
Keep in mind that my analysis is only for 2009. I would guess that results for other years are similar but I haven't checked yet.