Should MLB teams consider a return to the 4-man rotation?

In 2011, 92 pitchers qualified for the ERA title, meaning they pitched at least one inning per team game, that is 162 innings give or take. Just six years later in 2017, there were only 58 qualified pitchers, the fewest since the 1960 season. That was no aberration either, with 2018 coming in at 57, before an ever so modest bump up to 61 in 2019 (Dylan Bundy and Yusei Kikuchi both missed by one out, J.A. Happ by two).

Starting pitchers throwing fewer innings is nothing new of course, it’s been a very consistent trend for the last 50 years:

The two lines represent two group of pitchers. The first is the average innings thrown by the top 25 pitchers (by innings pitched) in a given season. I’ll refer to this group as the “workhorses pitchers” throughout. The second line is the average innings thrown by a number of pitchers that represents two times the number of pitchers per team. So today that’s the top 60 (30x2), whereas pre-1960 it was the top 32. The idea is to have a less exclusive group than the first, but one that changes with the league rather than stuck at a given number. I’ll refer to this group as “frontline pitchers”.

Post deadball era, the averages stablize in the mid-1920s, so that’s where I started the chart. From that point, things are pretty steady through to around 1960, when there’s actually a pretty big tick up. Some of that the adoption of the 162 game schedule, some of that is expansion diluting the pool and teams riding their best pitchers, but the magnitude — both group seeing their averages rise 40 innings in a decade — really surprised me and makes me curious. But that’s an issue for another day.

From those peaks in the early 1970s has been the consistent downward trend referenced above, as teams adopted the five man rotation, used fireman to close out games, and then specialist short relievers for single innings or even shorter stints. Pitch counts became fixture, with a particular fixation on the magic 100 level.

Initially, this was driven by attempts to keep pitchers healthy: it’s no co-incidence that the maximum workloads were at the dawn of free agency and the declines started right as player salaries exploded and multi-year contracts forced teams to treat pitchers as investments rather than horses to be ridden until their arms gave out.

Then there’s the even more precipitous decline in recent years I alluded to at the top. Even compared to just the beginning of the last decade in 2010, top 25 workhorse pitchers fell from averaging more than 220 innings per year down to just over 200 innings a year. That’s about the same decline as occurred from the late 1980s to 2010, over twice the amount of time. Likewise, frontline starters fell from averaging over 200 innings to just over 180 innings.

This step change over the last decade is about more than reducing workloads to preserve arms. One can debate how effective that has been as pitchers are missing as much time as ever with ever more major elbow injuries (Tommy John surgeries), though are also throwing harder than ever. The real question is whether they’d be getting hurt even more with higher workloads given the ever higher velocity (if that would even be possible). But again, that’s beyond the scope of this.

Instead, it’s more attributable to teams being reluctant to leave pitchers in to face batters three and four times, as opposed to bringing in a fresh reliever throwing darts for one inning. So it’s a story of understanding the “times through the order” (TTO) penalty, as well as the rise of “all out” one inning/short relievers. Thus increasingly, backend rotations types only go twice through the order, maybe four innings or into the fifth instead of 6+ innings, and the frontline starters increasingly only go six innings and maybe into the 7th in a good start.

In the near term, it probably can’t go much further. The decline in innings from starters has largely been accommodated by adding more relievers, with rosters going from nine or ten man pitching staffs when the Blue Jays came into the league 40 years ago to 12- and 13-man staffs common today. With rosters limited to 13 pitching slots and nine inning games, spreading out 1,450+ innings means each slot has to average over 100 innings even if options and injury stints are strategically manipulated to effectively create another slot or so. And few relievers even approach 100 innings. Someone has to throw the innings.

While many decry this development as a pretty radical departure from history and what they know or grew up with (and I sometimes gravitate to this reflexively), the more interesting question to me is whether this is truly optimal. By definition, frontline starters are the best of the best, with elite combinations of stuff and command. Mid-rotation starters aren’t exactly chopped liver either, they’re not far behind.

At the margin, you’re transferring innings from the best 100 on the planet, to something like the 400th or 500th best. Even if they’re fresh and can go-all out, the further down the list you go, the more marginal the talent. At some level, one wonders if the elite arms are being squandered unnecessarily.

It would be one thing if it was extending their careers such that they ultimately threw as many innings or more. But we’re not particularly close to cracking that nut and understand what does and doesn’t prevent injuries. Moreover, from a realpolitik point of view, even if that were the case, there’d still be an incentive to maximize the value during the time the player is contracted/controlled by the club. Getting fewer innings now in order to save a pitcher’s arm into his 30s to potentially pitch against you isn’t a great business move.

Which leads me to wonder if the solution is not in fact a return to the past: the four man rotation. If the driving force behind shorter outings at this point is limiting how many times a pitchers faces batters in a given game, the total workload could be repackaged to prevent dilution of the talent thrown innings.

For frontline starters who are capable of competently turning over a lineup three times, the idea would be something like making 40 starts averaging something like five innings or so (6-7 innings in good starts) to get to a total workload of 200 innings rather than 30 starts averaging closer to seven innings as was typical in the earlier part of this century.

For backend type starters whom you want to limit to more strictly limit to twice through the order (or third time through the bottom with an opener), it would be more like 40 starts of four innings a start to get to around 160 innings compared to 30 starts at 5-6 innings of yesteryear (and more like 130 innings these days).

Potentially, this could open the door to a wider variety of roles. Pitchers who aren’t quite starters but who rely on a broader arsenal as opposed to huge stuff that plays up in short bursts could be lined up as essentially piggyback starters, where they line up with a backend type starter and make 40 appearances at 2-4 innings a game. Aesthetically, this would be a nice change from the cavalcade of one inning relievers that happens almost every game when the starter departs.

Perhaps this is misguided. But as teams search for competitive advantages, I wonder if the next one isn’t squeezing more out the best pitchers as innings start to fall, and if this isn’t a beter way to balance workloads within games and across the season.