Manufactured Runs: Does the Rockies’ Four-Man Rotation Make Sense?

Well here’s something you don’t see every day—the Rockies are going to a four-man rotation. And what’s more, they’re going to put their four starters on a 75-pitch limit. Jim Tracy explained his decision like so:

"I felt we had to do something non-conventional," said Tracy of his beleaguered pitching staff that includes a reliever Josh Roenicke who has thrown more innings than one of the team's starters. "I was given the opportunity to tweak this. We are going to see what transpires as we move forward."

A four-man rotation was last regularly used in the major-leagues in the 1960s and early '70s. Asked if a starter would be pulled at 75 pitches with no runs allowed, Tracy insisted he was committed to this experiment.

"He has got to come out because he has to pitch four days later," Tracy said. "But if he goes five innings, he has pitched you to the point where you can go to a bullpen with some very significant people."

Jeremy Guthrie moves to the bullpen as part of this “non-conventional” move. On one hand, it’s easy to see why a manager might be tired of rolling Guthrie out every day, given his sparkling 7.02 ERA, 6.79 FIP, and 6.43 Fair RA—he’s been bad pretty much any way you want to slice it. But one wonders if the cure isn’t worse than the disease, especially given Doctor Tracy’s diagnosis:

"Very similar to Jason Hammel last year, maybe coming out of the bullpen for even a brief period of time will help get Jeremy back to where he needs to be," Tracy said. "Because right now where he’s at it's not benefitting him or our club. We are throwing entirely too many bullpen innings out of the same guys on a daily basis."

It’s true, the Rockies are throwing too many bullpen innings. There’s really only one cause of this—not enough innings coming from the starting rotation. Going into last night’s action, the Rockies had the second-fewest innings pitched per game started of any team in the majors, with only the Royals getting fewer innings out of their starters.

That’s a problem for the Rockies, aside from the fact that the starters aren’t very effective even when they do pitch. Looking at teams from 1993 through 2010, each additional inning pitched by a team’s starters per game resulted in a 0.072 increase in win percentage, even after controlling for the Fair RA of a team’s starting pitchers as a unit. This is of a piece with past research I’ve done showing that starting pitchers who go deeper into games get better bullpen support. Typically, when a team has to use its relievers more frequently, those additional innings do not go to a team’s closer or any of its top relievers, but instead are soaked up by the team’s worst relievers—so the more innings a team’s pen has to throw, the worse it can be expected to perform.

In terms of getting innings out of starters, Guthrie has actually been one of the team’s better pitchers:

It doesn’t need to be said that it’s not a good thing if the guy who has pitched more of your team’s innings than anyone else so far has an ERA north of 7. But the one nice thing you can say about Guthrie’s performance for the Rockies so far this season is that he is soaking up more innings per start than the average Rockies starter.

So here’s how it shakes out. Up until now, Guthrie was making 17 percent of the Rockies’ starts. Those starts are getting replaced by pitchers on a 75-pitch count. Let’s assume that Tracy stays true to his word and holds to those counts. So far this season, Rockies starters have recorded .1666 outs per pitch (Guthrie himself was responsible for .1659 outs per pitch, so in that respect he was a typically average Rockies starter). At 75 pitches per game, that’s 4.2 innings per game, not the five innings Tracy spoke of. If the Rockies’ starters manage to improve to league average, which is .1826, that’s still only 4.6 innings per start.

Is it possible that the remaining starters could adapt their pitching styles to try and stretch to five innings? Pitching to contact, for instance, doesn’t seem to reduce pitch counts the way some suppose that it does—throwing fewer pitches to each batter doesn’t reduce your pitch count if you’re letting those batters get on base. The way to economize your pitches boils down to three things—strike out more batters, allow fewer hits on contact, and cut down on walks. If that also sounds to you like three ways to just pitch better, you’re not wrong. The Rockies, realistically, have to get an above-average starting pitching staff to get five innings out of 75 pitches, and if they had that kind of pitching staff, they probably wouldn’t be trying this idea to begin with. PECOTA, for instance, sees a collective ERA of 5.07 in the future of the four pitchers (White, Friedrich, Outman, and Francis) left in the Rockies’ rotation, and while that would be an improvement over what the staff has done so far, it does not get them within spitting distance of average.

So the pitch count is going to force the Rockies to get between two-thirds of an inning to a whole inning more out of its bullpen than it was previously, which is not-so-coincidentally what Guthrie was pitching per team game so far this season (at .91 innings per team game). If Guthrie can manage to pitch the same number of innings out of the pen as he could out of the starting rotation, the Rockies can manage to tread water.

This does not seem especially likely, though. Guthrie would have to work at a pace to pitch 147 innings over a full season to match what he’s done so far out of the pen. The last time there was a pitcher with over 130 innings of relief work in a season was Mark Eichorn in 1986. No relief pitcher has even thrown over 100 innings since Scott Proctor did it in 2006, and even that sort of effort would still require the other members of the Rockies bullpen to increase (not decrease) the number of innings they throw.

Without knowing if the Rockies are even going to try to squeeze that kind of workload out of Guthrie, though, it’s hard to say what kind of performance they can expect from him. Clearly, he shouldn’t be expected to continue to give up runs at his current clip, because of regression to the mean if nothing else. (That said, it should be noted that the mean is a fickle mistress, and just because it’s unlikely doesn’t mean it’s impossible.) But if the Rockies are going to try and use Guthrie to soak up innings in relief, that will likely mitigate some (it’s difficult to say how much) of the normal effect we see from moving a pitcher into the bullpen.

Is there a silver lining to this move? Maybe, but it’s hard to see it. There is the times-through-the-order effect, so in the abstract we could expect to see slightly better performance from the starters than we would otherwise. But it’s not clear that this won’t be counterbalanced by the increased strain of starting games more often. And while the average reliever is likely to be better than a starter who has thrown 75 pitches, the Rockies are unlikely to be able to get average relievers to soak up the additional innings this will incur, especially since we already know that many of those additional innings will be soaked up by Guthrie in long relief.

The Rockies will get a few additional opportunities to pinch hit out of this, which may sound useful in the abstract and might even be useful in practice. But again—additional pinch-hitting opportunities are more likely to go to the bench players who are pinch hitting the least right now, and presumably they’re pinch hitting less because they’re not as well suited to it.

Given the significant problems with forcing your relievers to throw too many innings, there’s nothing in the Tracy plan that seems to help rather than hinder. Tracy should probably be forgiven for not doing anything to aid the pen, in light of the team's lack of options there. But just because one is already bleeding doesn’t mean self-inflicted wounds will do any less damage.