August 20, 2002

Doctoring The Numbers

The Five-Man Rotation, Part 2

by Rany Jazayerli

Today, we'll pick up last week's discussion where we left off. (Take a look at last week's article if you haven't already.)

To answer the question I ended last week's column with, it is very important to understand that my support of the four-man rotation is not, in any way, mutually exclusive with my belief in the importance of limiting pitch counts. It is fairly well-established (albeit not as well as I would like) that high pitch counts increase the risk of pitcher injury. But it does not follow that starting on three days' rest is more dangerous than starting on four days' rest. This is not a paradox at all, as both phenomena can be explained by one simple principle, which is important enough that it ought to be named and capitalized. Let's call it the Principle of Pitcher Fatigue:

Throwing is not dangerous to a pitcher's arm. Throwing while tired is dangerous to a pitcher's arm.

A gathering body of orthopedic evidence suggests that pitchers lose their mechanics as they become fatigued, and it is those imperfect mechanics that put undue strain on the arm, ultimately leading to injury. This is why a single 150-pitch outing can be more dangerous than a dozen 100-pitch outings. This explains why the PAP3 system works.

But this also explains why working on three days' rest may not be any more dangerous than working on four days' rest, because there is no evidence - absolutely none - that a pitcher needs a fourth day off to reach a maximum state of rest. I've searched the medical literature for any study that measures a pitchers' resting strength following a start, and such a study doesn't exist. No one has gotten around to measuring whether, to return to full strength, a starting pitcher ought to rest his arm for 3 days, or for 3 hours, or for 3 weeks. The move to a five-man rotation is simply not based on hard medical evidence.

Does a reliever need four days of rest? Of course not. Jesse Orosco has been pitching every other day since a Beatles reunion was still possible. If more rest were always beneficial, then it would follow that five days of rest is even better than four. No one believes, that, of course, which is why no team has ever seriously flirted with a six-man rotation.

Certainly, every pitcher is constructed differently, and it's possible that even four days of rest isn't enough for some pitchers. For decades prior to the standardization of regular pitching rotations, teams would routinely give veteran pitchers five or six days of rest. The tradition of the "Sunday starter" culminated when the Chicago White Sox gave Hall of Famer Ted Lyons, then 41 years old, twenty starts in 1942. Lyons completed all 20 starts, went 14-6, and led the league with a 2.10 ERA.

The practice of giving certain pitchers additional rest is complicated by having to keep other pitchers on rotation, but occasionally the situation warrants it--for example, with Pedro Martinez. From 1999 through last year, the Red Sox tried their damndest to give him five days of rest when they could. Pedro is not a large man at 5'11", 170 pounds, and the Red Sox were tacitly admitting that Pedro's small stature meant that they had to personalize their care of him to account for his particular situation.

Their method didn't keep Pedro from breaking down, though, because the Red Sox continued to let him throw 120 or 130 pitches when he did start. Despite making only 29 starts in both 1999 and 2000, Martinez eventually wilted from the abuse and missed half of 2001. Upon his return this year, the Red Sox have not only given him as much rest as possible, they've also limited his pitch counts--he hasn't thrown more than 117 pitches all season. After a spring in which it looked like he would never return to his dominant form, Pedro has only gotten stronger as the season has gone on.

The example of Martinez not only illustrates the limitations of additional rest, it also points out the importance of integrating a move to the four-man rotation into a comprehensive philosophy on monitoring pitcher workloads. The last time a team seriously experimented with the four-man rotation was in 1995, when Bob Boone went with four starters for the first half of the season. After getting off to a rousing start--Kevin Appier began the year 11-2 with a 1.89 ERA--the trial was scrapped after Appier and Chris Haney suffered arm problems. The downfall of the experiment was that Boone did not compensate for his starters' decreased rest by limiting their pitch counts. On the contrary; Appier threw 141 pitches in a start shortly before he went into the tank.

I am confident that the organization that is willing to return to the four-man rotation, in conjunction with strictly monitoring their starters' pitch counts, will gain tangible benefits without increased risk to their pitchers.

There is another substantial benefit that comes from the four-man rotation, which is that pitchers have better command on three days' rest. Almost every pitching coach--and most pitchers that have tried--will tell you that when you pitch on three days' rest, your stuff is a little sharper, and has a little more sink.

This sounds like mere anecdotal evidence, but in this case the anecdotes are fortified by the statistical data at hand. With the help of Keith Woolner and the wonderful people at Retrosheet, I was able to obtain start-by-start data for all pitchers back to 1978. In the past 24 years, no less than 160 pitchers have made at least 8 starts in a season on both three and four days' rest. By looking at how their performance on varying amounts of rest, we can determine whether being brought back one day earlier had any deleterious effects.

The combined numbers of those 160 starters:

Rest       IP      H    ER    BB     K    HR   ERA   H/9  BB/9   K/9  HR/9
3 days  13666  12986  5428  4270  7516  1087  3.57  8.55  2.81  4.95  0.72
4 days  16082  15658  6882  5063  9052  1389  3.85  8.76  2.83  5.07  0.78

Did the pitchers suffer from being afforded less rest? Hardly. They pitched better on shorter rest, with an ERA over a quarter run lower. In a combined sample size of nearly 30,000 innings, that's significant. 92 of the 160 pitchers (58%) had lower ERAs on three days' rest. Even if we lower the threshold to include all pitchers who made even 5 starts on both three and four days' rest (368 pitchers in total), the numbers don't change: 215 of the 368 (58%) had lower ERAs on three days' rest.

A closer breakdown of their performance lends credence to the idea that pitchers have better command, and more sink, on three days' rest. With less rest, the starters let their defense work for them, with fewer walks and fewer strikeouts. But perhaps the most relevant effect is that they allowed nearly 10% fewer home runs per inning.

It's not unequivocally clear that starters perform better on three days' rest than on four. But it is crystal clear that they don't perform any worse.

A lot has been made of the recent struggles of pitchers who were asked to return on three days rest. In 1999, Darryl Kile and Pedro Astacio each made five starts on short rest, and Kile had a 7.39 ERA, Astacio 9.83. However, their performances should not imply that pitchers today have more trouble coming back on three days' rest than pitchers of 20 years ago. The last starters to make at least 8 starts on short rest did so in 1995, when a trio of Royals (Appier, Mark Gubicza, and Tom Gordon) and a pair of Pirates (Denny Neagle and Esteban Loaiza) were called upon a combined 47 times on three days' rest. Their performance that season:

Rest       IP      H    ER    BB     K    HR   ERA   H/9  BB/9   K/9  HR/9
3 days    302    310   131    90   214    23  3.90  9.23  2.68  6.37  0.68
4 days    541    561   250   184   325    58  4.16  9.33  3.06  5.40  0.96

The sample size is extremely small, but the performance of this group from just seven years ago ought to squelch any talk that pitchers today simply can't come back on short rest the way pitchers of previous generations could.

Another issue that is frequently raised regarding the four-man rotation is that it would be difficult to expect pitchers in mid-career, who have been trained to pitch on four days rest their whole lives, to suddenly change their work habits.

Hogwash. At the beginning of spring training, few teams have settled on their fifth starter, and most teams put the established members of their rotation on the mound every fourth day, beginning with long-toss sessions before exhibition games have started, continuing with 2-3 inning outings in early March, and progressively lengthening them, before finally switching to outings every fifth day in the final few weeks of the spring. Keeping those pitchers on an every-fourth-day regimen would not require a change in their training habits. On the contrary, it's switching to an every-fifth-day cycle that requires an adjustment.

That's the smaller point. The bigger point is this: if pitchers really had difficulty adjusting from having to pitch every fifth day to pitching every fourth day, wouldn't they have even more difficulty switching from starting to relieving? John Smoltz had been starting in a five-man rotation for over a dozen years before he was moved to the bullpen, and he's taken to the closer role like a fish to water. Eric Gagne has been a starter throughout his pro career--did he struggle when moved to the bullpen? Of course not. Neither did Mariano Rivera, or Billy Wagner, or Billy Koch, or dozens of other pitchers who started their entire careers before they were moved to the bullpen and found immediate success.

Starters throw to full exertion, on a regular schedule, with plenty of notice before their next appearance. Relievers throw in short stints, may pitch in three straight games or not pitch for ten days straight, and don't know if they're going to pitch on any given night until five minutes before they're brought in. You couldn't create two roles that required more different physical and mental preparation--and yet pitchers go back and forth between the two all the time. Yet we're supposed to believe that a starter can't adjust to a usage pattern where he 1) stays on a regular schedule and 2) still knows in advance when he's going to pitch, just because he gets a different amount of rest? Please.

Next week, we'll wrap up this topic by looking at the other, ancillary benefits of the four-man rotation.

Rany Jazayerli is an author of Baseball Prospectus. You can contact him by clicking here.

Rany Jazayerli is an author of Baseball Prospectus. 
Click here to see Rany's other articles. You can contact Rany by clicking here