Last week, on my favorite podcast in the world (seriously) Effectively Wild, Sam and Ben interviewed now-former Dodgers head trainer Stan Conte. Their interview covered a wide range of topics related to player injuries, both the ones that we get to hear about and the ones we don’t. At one point, Sam asked a question about a hypothetical team that was pretty much guaranteed to go to the postseason and had little to play for over the last two weeks of September. Should the team rest its starters down the stretch, saving them for the playoffs and letting the September call-up guys take over, or perhaps go to a six-man rotation to take some of the strain off the four guys who will actually be starting playoff games?

Conte’s response was counter to what I would have expected. He suggested that the team not rest their pitchers (within reason, of course) because pitchers are creatures of habit and their habits are based on a five-day schedule. Instead, Conte suggested that a better approach would be to give pitchers extra days here and there during the earlier months of the season. If a pitcher was taken care of then, he’d have plenty left in the tank if needed for a stretch run or for the eventual postseason games. Maybe the Royals should have given Edinson Volquez or Yordano Ventura an extra day in May?

In the 30 or so years since major-league baseball has moved functionally to a five-man rotation, it’s rare any more to see a pitcher go on three days rest. But what’s interesting is that baseball occasionally throws a team an off-day on a Monday or Thursday. In theory, if a pitcher is good to go every fifth day, a team could use those off-days as a way to skip the least productive of their starters, but still keep everyone on “regular” rest. And sure, sometimes they do skip no. 5, but often they let everyone take their turn and thus give everyone a sixth day of rest. It seems a strange thing to do. Why let the worst starter on your staff start if you can avoid it?

We know that pitchers pitching on five days of rest do no better (or worse) than guys on six days of rest in general, so there’s no advantage in the short term to giving a guy extra rest (and there’s a penalty to be paid in the fact that your fifth starter pitches when one of the better ones could be doing that instead.) But maybe there’s something to giving guys a little rest now that you gain down the road.

Warning! Gory Mathematical Details Ahead!

I used data from the regular season from 2010-2014, and only used pitchers who started 30 or more games during the year at the MLB level. This means we get pitchers who weren’t hurt (for the most part) and who spent all or almost all the year at the majors, so we know their rest patterns. I coded each start as taking place on either four days of rest or 5-9 days of rest. Of course, there were some three days of rest starts and the occasional 10-12 days off, and those were not counted. I then took a running count of how many times during the season a pitcher had gone on regular (four days) rest and long (five or so days) rest, so that by the time we got to September, the equation knows how many times he’s gotten an “extra” day.

I looked specifically at September (and early October) to see whether pitchers who had gotten more days off in the early part of the season were performing better (or worse) than their season stats would have indicated. If he was doing better (or he was holding the line while the guy who hadn’t gotten as many days off was doing worse than we might expect), it tells us something about the cumulative effects of rest.

I controlled for the pitcher’s overall quality and the quality of the batters he was facing. We know that in September, contrary to popular belief, everyone is still trying hard, so we don’t have to worry about that as a confound. I looked at several outcomes at the plate appearance level (strikeouts, walks, etc.) and at the pitch level (swing-and-miss, strike vs. ball). For the initiated, I created an expected probability for each outcome given the identity of the batter and pitcher, and put that as a control variable into a binary regression. I also entered the number of times the pitcher had gotten an “extra” day earlier in the year. In the data set, I only looked at September (and early October) appearances, so everyone had made a lot of starts already and so we assume that everyone was functionally equally fatigued.

The results? Nothing. Extra days of rest earlier in the season didn’t help a pitcher in September (or hurt him). Everyone pitched about how their seasonal stats would predict he would, at least in the aggregate.

The Four-and-a-Half Man Rotation

Now, before we over-interpret these findings, let me cop to one major methodological issue. There’s a big difference between doing epidemiology and being a clinician. (I’ve done both.) The epidemiologist sees patterns across the entire population and talks in terms of risk factors and probabilities. The clinician sees the person sitting in front of him. I wouldn’t say that it’s never a good idea to give the rotation the benefit of that extra day in May, but the burden of proof should be on the side of showing why they need it. It’s not that the burden can never be overcome. If a guy has a sore something or other in his arm, sure, rest him. It’s that there’s no reason to give guys rest on the thought that there are long-term benefits.

(There’s also a second methodological issue for the super-initiated: homogeneity of variance. Because I used data from 2010-14, and almost all teams do give their pitchers extra days here and there and liberally use a fifth starter, regardless of the schedule, it means that just about everyone in the data set had gotten some “extra” days off and so basically everyone was the same in the regression. It makes it hard for the regression to see degrees of separation. So maybe this is all a spurious finding.)

It brings up a strategic issue. Should teams be so willing to carry a fifth starter, especially if there’s an off-day on the schedule? They could start their top four guys and not rush anyone back onto the hill and squeeze more starts out of them, rather than the guy who is likely just a DFA waiting to happen. There will be some weeks without an off-day, but if you can juggle things around, it might be worth it for a team to really only use the fifth starter when they actually need him and free up the roster spot.

The one thing that I can’t directly speak to is the issue of whether pitchers who get a little more rest are less prone to injury. I do know that injury patterns tend to follow the era of pitcher usage. When four man rotations were common, pitching on three days rest didn’t seem to impact a pitcher’s injury risk as much. In the modern era, the effect seems to be pointing somewhat in the other direction. It’s entirely possible that giving a pitcher extra days gives him a little health boost, but it’s not clear whether that would be because there’s something inherent in the occasional extra day of rest that’s protective or because that’s what his body has grown accustomed to expecting.

Or maybe the extra day off is just to keep the starters happy. Everyone likes a day off here and there when they weren’t expecting it and maybe that has some positive effect that we can’t really measure yet.

Why does baseball have to be so complicated?

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
The premise to the article is great. The process to get the answer is great.

The entire payoff was this:

"The results? Nothing. Extra days of rest earlier in the season didn’t help a pitcher in September (or hurt him). Everyone pitched about how their seasonal stats would predict he would, at least in the aggregate."

Not a single piece of data to hang our hat on. If Verducci did this, we'd blast him for it. Baseball Prospectus? Hugely disappointed.

At a pure minimum, can't you just provide a link to the results of the data?
Tom, it would be a link to a bunch of regression output that showed non-significant finding after non-significant finding. There's only so many times someone can read p = .872 until they glaze over. There just wasn't anything there to show.
Russell, you can easily show the "before" data, in terms of IP, and K and BB and ERA, and you can show the "after" data in the same way, and split it based on 4 days rest and 5+ days rest.

Forgive me if I've missed this in my brief read-through, but were you able to answer the question of whether or not guys with extra days off earlier in the season were less injury prone? I gather that extra days off doesn't have an effect on results, but I wonder if it his an effect on a pitcher's ability to actually generate results.
I don't know the answer to that one yet.
There is another way to "rest" a pitcher, of course, which is to pay attention to the pitch counts. Has anyone tried a 4-man with strict pitch counts? Essentially that allows you to either make what would have been your 5th starter into a long reliever, or carry another guy whose limited repertoire makes him a non-starter. There would be other options of course for that roster spot. The main point is, especially with pitch counts becoming more prominent, why aren't we seeing more short rotations?
The Rockies did it a few years ago (4 man rotation, 75 pitch limit...) Evaluating anything to do with pitching in Colorado is dicey.

The Rays have had some success with a model emphasizing a starter going 18 batters (and then out!) this past year.
What is the value of rest or extra rest? As I watch Jacob deGrom get hammered, pitching on very long rest, my vote goes to a pitching routine that includes consistent intervals between starts. Extra rest throws off the finely tuned body clock that all professional athletes have. Feeling too strong is nearly as bad as fatigue.