The Shrinking Starting Pitcher Workload: Prospect Edition

Throughout baseball history, starters have thrown fewer pitches and innings than the generation of pitchers who preceded them. The trend dates to at least the early 20th century and has continued almost unabated ever since. Individual throwbacks will occasionally buck league trends — remember when James Shields tossed 11 complete games in 2011? — but history steadily marches on.

While the trend is clear, the curve isn’t linear. Throughout the game’s history, there have been a few accelerating events that have reduced workloads usage much faster than normal. The addition of the designated hitter and new definition of the save rule sent innings per start tumbling in the early 70s, for example.

It’s too early to tell definitively, but we may be on the precipice of another acceleration. We’re only two months into the major league season, and the pandemic and its related fallout are relevant and difficult variables to account for. But while major league starters are approximating the workloads they carried two years ago, minor league starters are not. On any given day in the minor leagues, someone might throw 100 pitches — but probably only one or two pitchers, and they almost certainly don’t reach 110. Remember when 100 pitches was a sort of threshold point for starters? This season, 85 is the new 100.

I went back and looked at every minor league game from May 19 to 24 (these days weren’t cherry picked; it, uh, took me a while to finish this article). In those five days, only four starters — all of whom were Triple-A vets — reached the 100 pitch mark. In that time, there were 12 entire organizations that didn’t have a single pitcher reach even 90 pitches. For the Astros, an organization that won’t let minor leaguers crest 30 pitches in an inning, the high-water mark was 79. For the Phillies, it was 67. In fact, Philadelphia has only had a minor leaguer top 80 pitches once all season.

The pandemic is to blame for some of what we’re seeing. At best, minor league hurlers had a jittery 2020, with intermittent bullpen work and perhaps a bit of summer ball at the alt site, but otherwise nothing between spring training and instructs. After such a weird season, teams are being understandably cautious with their personnel as they ramp back up.

Ross Seaton, Arizona’s Director of Pitching, is trying to balance the twin challenge of prioritizing development and keeping everyone healthy after a lengthy layoff. The Diamondbacks are limiting their farmhands to 85 pitches in the early stages of the season, which is slightly lower than the usual ceiling. “We know that in order to develop and get better, it requires reps,” Seaton says. “At the same time, we don’t put to guys in risky situations. It’s in our best interests, and the best interests of the player, to make sure they’re in safe environments.”

For Arizona that means limiting workloads. “Most of these guys didn’t get a chance to play last year,” Seaton said. “Hopefully this is a once in a lifetime experience. But you can’t replicate the season, no matter how much you practice.”

Oakland is following a similar model. Steve Connelly, the pitching coach for Oakland’s Double-A affiliate in Midland, said they’ve lowered their pitch count limit this year, not only because of last year’s truncated campaign but also because this season’s abbreviated spring training put pitchers a bit behind schedule.

“Usually, leaving spring training, our guys would be built up to at least five innings and 75 pitches,” Connelly said. “This year, with the way spring training went, we didn’t have anyone built up past 60.”

This season, most minor leaguers reported to camp more than a month later than usual. A few were brought into big league camp, but everyone else had to wait until late March before reporting. Spring training was also shorter than normal and for Oakland, that meant one less session of live BP and fewer games prior to leaving Arizona.

Innings limits are also a factor for Oakland. The A’s, like many other clubs, have innings targets for their pitchers, and those goals are going to be low for guys who didn’t throw last season. “Most of our guys have built up to 90-100 pitches” Connelly said. “But most of them are on innings limits because they missed all of last year. We’re keeping our guys to five or six innings so that they’re not limited to three or four inning starts in September.”

A few teams are treating 2021 like a normal minor league season. Seattle regularly has even their low-minors arms work past 90 pitches, as does Cincinnati. The Reds are particularly notable, because even their top farmhands get pushed: Nick Lodolo threw 90 pitches or more in two of his last three outings, and Hunter Greene reached triple digits in his last start.

In part, that’s because the Reds use pitch counts as only a very general guide. Kyle Boddy, the Director of Pitching for Cincinnati, prefers to use them as a ballpark figure for the low minor leagues. These limits are also about effectiveness, not health: “There are very few pitchers who can hold their stuff and velocity into the low-100s,” Boddy said, while also noting that the club will stretch pitchers more in Triple-A as they reach the cusp of the majors.

But the Reds are among the league’s exceptions. Even before the pandemic, pitch counts and innings totals had been falling throughout the minors in recent years. And as this year’s pandemic-fueled workload reductions have tightened innings and pitch limits even further, it’s worth taking a step back and tracing the route to how we got to a point that a six-inning start became such an irregularity.

For decades, it was effectiveness, not stress that dictated how long pitchers stayed in the game. Over time, as the relationship between workload and injury became clearer, teams began looking for ways to measure stress. They started tracking pitch counts in the 1970s and ’80s, and that metric widely became the instrument of choice for measuring stress around the turn of the last century, thanks in part to the pitcher abuse points (PAP) statistic created by Rany Jazayerli and Keith Woolner.

When Jazayerli and Woolner first wrote about PAP, 125-pitch outings were a daily occurrence: there were 200 of them in the year 2000, two years after Jazayerli and Woolner published their research. But the frequency of such starts soon plummeted. A decade later, there were only 25, and only three of those starts lasted more than 130 pitches. Since 2016, we’ve had only 14 such outings combined (Trevor Bauer has four of them).

Pitch counts are well intentioned, of course. But as workloads dip it’s worth asking: Are they doing any good? And is it possible that such tight restrictions may actually be harmful?

Let’s start with the obvious. Whether for tracking effectiveness or risk of injury, it’s better to know the pitch count than to have no data at all. And in youth baseball at least, studies have shown a relationship between elbow pain and a high pitch count. Pitch limits, at least for kids, are a good idea.

In the professional game, things get more complicated. If pitch counts are meant to measure stress, they’re a very blunt instrument for doing so. The stress of one pitch is not identical from pitcher to pitcher, nor even from pitch to pitch from the same pitcher. It’s also not clear whether the core idea behind PAP — that there’s a threshold at which point the next pitch becomes much more dangerous than the previous one — is correct, nor at which pitch that principle kicks in if it is. Nobody thinks 180-pitch outings are a good idea, but if there’s convincing evidence that, all else being equal, the 100th pitch is significantly riskier than the 92nd or 108th, I haven’t seen it.

All sorts of variables — how stressful a pitcher’s outing has been in particular but also his workload in recent days or weeks, his velocity, his pitch mix, the amount of sleep he got the night before, etc. — preclude a clear link between pitch counts and health. Jazayerli himself readily acknowledged the shortcomings of using PAP theory to dictate workloads: “For two years, I have tried to use PAP as a framework in which to center the ongoing discussion of pitcher usage. In the process, though, PAP became more than a framework for measurement; it became the standard for measurement. Which it was never intended to do.”

But even as PAP have faded from the discussion, pitch counts shrink ever more. In today’s game, it’s normal for minor leaguers and even many big league starters to throw fewer pitches per outing than they did before they became professionals. It’s taken as a given that prospects will be kept on tight leashes, and that teams will limit the throwing they ask their personnel to do, all for the sake of keeping players healthy. Again though: Are these restrictions helpful for development?

Public medical literature on this topic remains somewhat sparse. A 2016 study in The American Journal of Sports Medicine found no meaningful relationship between pitch counts and injuries among pitchers returning from Tommy John surgery, though the authors were careful to note that their conclusions should not be applied to healthy pitchers. Jon Roegele keeps a record of Tommy John surgeries, and his list is the gold standard for tracking when minor or major leaguers go under the knife. The data isn’t complete — finding news about minor leaguers getting surgery was very challenging 15 years ago, and so cases from back then are probably underreported — but at the very least it seems clear that shrinking workloads have done little to slow elbow injury rates:

It’s also worth mentioning that, per Derek Rhoads and Rob Mains at Baseball Prospectus, the injury rate for big league pitchers (as measured by days on the non-COVID IL) is up meaningfully so far this year compared to 2018 and ’19, though down relative to last year.

Other research suggests that it may not be the volume of pitches that matters most, but the effort expended in tossing them. Ian McMahon wrote a fascinating piece on this subject at The Athletic last year. The whole thing is worth your time, but for our purposes, this part of the essay particularly resonates:

“Chasing velocity has a price. Recent biomechanics research from Fleisig and ASMI indicates that, while it’s difficult to predict elbow stress between two pitchers based solely on velocity, quantifying elbow stress is much simpler when considering one elbow at a time. Quite simply, for a pitcher, higher velocity equals more torque and increased risk of injury. Max effort equals more risk of shredded ligaments. To lessen elbow torque, the study recommended pitchers vary the velocity of pitches.”

And yet, pitch counts continue shrinking. Today’s rookies have progressed through a minor league system that rarely asks them to reach 100 pitches or work through the order a third time. And despite how well-intentioned a pitch count may be, there are real developmental ramifications.

First and foremost is simply a lack of reps. For the most part, minor leaguers are where they are because they’re not ready to contribute at a higher level. They need to work on their slider, their command, or whatever their particular goals may be. Obviously, a pitcher who throws 80 pitches has 20 fewer tosses to work with than someone who throws 100. That may not seem like a lot, but over the course of a season, you’d imagine that most players would benefit from 20% more opportunities to hone their craft.

Inherently, they’re also not being tested when the game gets harder. We all know how stark the third-time-through-the-order penalty is, and how difficult it is to navigate a lineup for a third time. Logically, it can only be more difficult for pitchers who have had limited experience facing that particular challenge, which is the case for many younger pitchers.

Those reasons help explain why the Reds haven’t reduced their pitch count goals this year. “All evidence I’ve dug up shows absolutely no value to them. Limiting pitch counts isn’t cautious,” Boddy says. “It’s not preparing them for reality.”

At the core of Boddy’s philosophy here is the idea that building up a starter to handle a normal big league workload should be part of the development. He used Greene’s development path as an example.

“If Hunter is limited all year, then what exactly is he going to do in the show in 2022? He threw 80 innings in sim games last year, plus alt site, and he sits at 100-plus mph in the sixth inning. Cutting him to 65 pitches does… what, exactly?”

That isn’t to say Cincinnati uses a one-size-fits-all approach. “We have quick hooks for some guys,” Boddy says, adding that there are also players who aren’t properly built up still back at the club’s complex in Goodyear, Arizona. But for measuring fatigue, the Reds use other tools, including Motus and velocity/strain curves. These aren’t perfect instruments either, but they do provide a more player-specific reference point for club officials trying to monitor workloads and prevent injuries.

For those whose pitch counts are restricted, it’s fair to wonder whether low pitch volumes are making it harder for them to develop the arm strength needed to pitch deep into games. For a look at how this plays out in real life, let’s briefly examine Spencer Howard. Howard is the Phillies top prospect, and the 31st best in all of baseball. After averaging 85 pitches per start in 2018, Howard never topped 85 tosses the following year. Often, he was asked to throw much less than that, and he averaged just a hair over 70 pitches across 15 outings. That’s mostly held over the last two years as well: He threw 91 pitches once last August, but has otherwise generally been kept under 80. His highest count this season is 68.

There may be good reasons for that. Howard dealt with shoulder discomfort last year and he had a shoulder scare in 2019 as well. Between that and weird scheduling from the pandemic, you can understand why the Phillies are handling him very carefully. However sound that plan is though, it seems to have hindered his ability to develop as a starter. In his five big league outings this season he’s had trouble holding his velocity over the course of the game and hasn’t yet thrown more than four innings (graphic courtesy of Brooks Baseball):

For Howard, the long-term consequences are unclear. Perhaps the Phillies are wisely easing him back into a full-season workload after a weird year. Maybe, with his history of arm trouble and this pattern of dipping velocity, a future in the bullpen would be a better fit anyway. As it is though, Howard is somewhat stuck in the middle: Philadelphia is developing him as a starter, but without letting him build the endurance needed to thrive in the role.

Ultimately, we don’t know what effect the pandemic will have on injuries and starter workloads going forward. Only time will tell whether the diminished pitch counts we’re seeing this season are an aberration or a preview of things to come. Regardless, I can’t help but wonder if the workload reductions we’ve seen in recent years have gone a step too far. Pitchers may be less effective the longer they throw, but there must surely be some advantage in pushing them deeper into games, if just for the additional repetitions. And if injury rates don’t rise linearly with pitch counts, then trimming pitch counts seems counter productive. In a league where both workloads and competitive advantages are shrinking, an enterprising team might just find one of the latter by taking a different approach on the former.