UTK Special 7/16/24

Deploying Assets

I’ve been working on this piece for more than a month and hope you enjoy. It’s long and some mail systems may not get the entire thing. If not, you can always go to undertheknife.substack.com and read this or any piece. UTK will be back on Friday as we get back into baseball. As always, thank you for reading and subscribing.

In 2024, not much about pitching is the same as it was “back in the day.” It may look like Nolan Ryan or Stephen Strasburg when Paul Skenes is out on the mound, but behind the scenes, it’s simply not. Everything from pitch counts to pitch design, from warmups to cooldowns, from side work to off-season programs are just different. Largely better, but there remains a troubling lack of evidence and a more troubling lack of results when it comes to pitching injuries. We’ve sacrificed a lot of arms on the altar of velocity and movement.

That’s not going to change, because it works. So what can change? Many decry the lack of complete games and the devaluing of starters, but I think there’s an opportunity to make the game better, at least in the sense of limiting injuries and keeping more of the best arms on the field. At one point in ‘24, eight of the highest paid pitchers were on the IL together, while five of the six Cy Young candidates from last season were shelved.

If everything has changed, why do pitching assets - fragile, expensive, and rare - keep getting deployed in much the same manner as they were fifty years ago? That question led me to talk to nearly 100 people inside and around MLB in order to find out what could be done and more importantly, what is likely to change in the near future. I’ll quote some of them here, but all are anonymous to protect sources I use during the season.

The simple answer about that last question is “not much.” Baseball has used the five man rotation for fifty years and that change only came about because a couple teams had five really good starters. Prior to that, the game had a slow adjustment from the 1920’s to the 1960’s. It went from three to four defined “starters” as pitchers had more risk of runs from a livelier and regularly replaced ball.*

The true five-man era can be traced from the pitching changes after 1968, with both the mound and expansion thinning things. Through the 1970s and 80s, hitting didn’t challenge the pitchers, but the so-called steroid era did. Hitters got bigger and stronger, steroids or not, and nowhere shows that more than the change at shortstop. Ozzie Smith, let alone someone like Mark Belanger, couldn’t exist in today’s game. As Cal Ripken, Derek Jeter, and Alex Rodriguez became the model - bigger, stronger, power hitting - those slap-hitting spots in the order were filled with more home run threats and every pitch had more emphasis.

A pitching coach today who has connections back to that era said that it was very clearly effort. “You say the last dial-it-back pitcher was Orlando Hernandez,” he said, “but I think it was Aroldis Chapman.” Chapman as a starter did pace himself, throwing in the mid-90s and dialing up in key situations. I argue that Hernandez lived much lower and could dial it up in the 90s, even late in his career, rather than being a freakish velocity talent like Chapman. “After that,” he continued, “it quickly became max effort and not just on the mound. Watch those high intent sessions at Driveline early on, where they were solving the lab problem, and now even pens had destructive possibilities.”

This isn’t a knock on Driveline, but the coach is right. For many years the “lab problem” faced by facilities as varied as Tom House, ASMI, and Motus was that pitchers didn’t throw as hard as they did in game conditions, often much less. My first day at Motus, I looked up a pitcher that touched the 100’s to see his mechanics, but his pitches were in the 80s. That didn’t even get to the transfer thresholds for the elbow.

The idea that the in-season side work is ‘touch and feel’ while the off-season can be more programmed isn’t always the case. Travel, demands, and even the pitchers own preferences can often work against it, let alone adjustments to recovery. The five-man rotation - or any rotation - is based on the starters recovering by a certain point and the relievers recovering enough to cover the outs the starter doesn’t get.

That often leads to discussions of the six-man rotation. More rest is better, right? Not necessarily. One AGM who has investigated it after a couple years of pitching injuries had his team going 10-deep plus, says that he’s not sure we’re already not there and beyond. “Find me a team that’s gone through a season with less than ten starters in the last decade. Spot starters, injury replacements, even trades. We’re not doing a normal five-man rotation now,” he explained. “A six-man just changes up the rest and takes starts away from your best pitchers.”

If pitchers are throwing less innings today - and they clearly are - and less pitches, they’re also usually throwing less outings. Greg Maddux made 36 starts in 1993. By 2023, Miles Mikolas made 35, but only a handful of others were above 30. Going to a defined six-man would reduce it even further. If you have a Maddux or even a Mikolas, why steal three starts from them and hand it to .. Kent Mercker, who was one of only six starters and 13 pitchers used in that ‘93 Braves season! Mercker’s far from the worst option, but I think even Mercker himself would rather hand Maddux the ball!

Another option many discuss is the “one per day” option. It’s a variant on what Japanese leagues do and with discussion of MLB adopting the week-per schedule of MiLB, it could work with a defined off-day. Colleges often work this way, though their shorter schedule and more off days lead to more changes. There are few true Friday and Saturday starters, though relieving on a throw day, especially in the tournament, is very common. (Interestingly, the Tommy John rate is very similar to the minor leagues despite very different usage patterns. Someone needs to do a full study, age matched, to see why this is the case.)

Setting pitchers to pitch one per day (ie, this guy every Tuesday, that guy every Wednesday and so on) could create a need for functionally seven starters. Going seven days in a row is uncommon, but not so uncommon that one wouldn’t need to plan for it. Moreover, almost every pitching coach or scout that I spoke to said that the change in routine would be very difficult for many pitchers and would face a lot of resistance. That said, most think the change could be made. “If we went to a four-man or a six-man, they’d adjust,” said one NL pitching coach. “They adjusted to the pitch clock, or are trying to, so it’s a matter of figuring out what’s best. Is it as simple as adding a [recovery day] to the mix where they’re actually resting and doing the stuff we want them to? If so, we need someone able to run that without asking [the medical staff] to do one more thing.”

Another point against the OPD system is that it limits even further the opportunities for starts without evidence that there’s an upside. 26 starts would be the baseline on this system, handing as many as nine starts off to the seventh best starter that could have been taken by the best. There’s no evidence that extra days of rest reduces injury either. “We took a really hard look at collegiate injuries as part of an internal review of the draft,” explained an MLB team doctor. “Looking at those Friday and Saturday starters, we noticed that even a decade or more ago, they were no less likely to have injuries, even when you expect there’d be some selection bias.” (One would assume players being looked at as draft picks would also be healthy. That’s hardly the case, especially recently as players with recent Tommy John surgery have been drafted in the first round. Just this year, Braden Montgomery limped to the stage with a walking boot on, after breaking his leg during the NCAA Tournament. He’s expected to make a complete recovery and play later this year at some level once he signs with the Red Sox.)

Looking at those two systems often leads people in the other direction. The four man rotation has been discussed by nearly all teams and attempted a few times, though almost all have been the “four man skip” system which is functionally a five man. An NL team recently discussed a four man with pen day system, based on the idea that having pitchers get an extra days rest regularly would allow the ‘short rest’ to work.

There’s little evidence of this. With the proper loading, an arm (or any human system) adjusts to the load. Going back to my 2006 article on pitching development, buildup for pitching is no different than adding weight to a bench press or a running progression for a marathon. Baseball acknowledges this on the low end, ramping pitchers up in the pre-season and with rehab assignments, but never going beyond a low threshold. We’re very likely missing out on a few pitchers that could go 120 regularly, or perhaps significantly more. Essentially, the system acts against freaks. There’s increasing evidence that adding load is predictable and programmable in almost all human systems and activities, as well as some interesting research about the need for variation.

That variation is one where a pitching researcher outside of MLB pointed as the most likely implementation. “I think instead of the expectation of every start being six innings and 100 pitches, we should be using the pitchers a bit differently,” they said. “Every once in a while, we get a signal that one of our [collegiate pitchers] needs a short day. They go out for 50 pitches and someone knows they’re coming in. That has to be happening in MLB, but I never see someone get a ‘short day’ or even a programmed ‘long day.’ Right now it’s just throw hard, spin hard, max effort, and we can see the damage that is doing.”

With the shuttle of pitchers between the majors and minors, plus the taxi squad, the average team will use twenty pitchers in a year. Given that, there’s no excuse for not being able to sub in a starter or doing a pen day if things just line up in a way that would put a fatigued pitcher on the mound as a starter. Guarding against that alone would be a huge step, but teams would need to know. “The resistance to anything new is strong,” said one NL AGM, “even when it exists. Even when it’s being used in colleges. Workload monitoring is tough, largely because the players won’t wear a wearable of any type, unless it’s theirs. The next CBA has to fix this.”

There is some acceptance of the use of the opener. However, there’s very little evidence that it works. “It’s an inning or two,” said one assistant pitching coach, “so we’re just shifting where the pen is throwing. Do I want the starter pitching in the seventh? I don’t know, is it that much more important and why can’t he get there from the start.” There are situations like heavily platooned lineups, where it makes sense, but Earl Weaver had that figured out in the 1970s.

The key of course is workload. While teams and coaches have always monitored it, it’s never been direct. For over a decade, that’s been possible, but it’s not widely used at any level in baseball. “It’s hard,” said one collegiate pitching coach who has previously worked for an organization. “You were at Motus. The hardest part was keeping them charged. Telling the kid to wear it. Collecting them after practice to upload the data. Programming the next day’s practice. There’s not enough interns in the world to keep up with it even when everything’s working.” Several others echoed this and hoped that camera-based systems could combine pitch count and biomechanics on a much more automatic basis.

How important is workload, so that cop-outs on managing a system shouldn’t matter? Rather than rewriting a bunch of things, I’m taking the easy route and excerpting myself. This is from the pitching chapter of The Science of Baseball, which you can buy to get the rest of the book:

One of the questions I'm often asked is how pitchers from previous eras put up seemingly huge workloads in terms of both innings and pitch counts, but today's pitchers break down at far lower numbers. The fact is, there's two major factors.

First, we largely remember the outliers - Nolan Ryan and Tom Seaver are Hall of Famers for a reason - but at the same time, look at the hitters of the era. Ozzie Smith hit his "go crazy folks!" home run that's famous, but few remember that was his only homer of the year. Smith, a no-doubt, inner circle Hall of Famer had 28 career homers. Dave Concepcion, the shortstop at the heart of the Big Red Machine never hit more than 16 in a year. There are two shortstops in 2021 that had more than that ahead of the All Star break! There's no batters to "ease up" on anymore. In 1905, Christy Mathewson wrote a book called "Pitching in a Pinch," where he espoused letting bad hitter put the ball in play, but saving one's best stuff for the best hitters and situations where runs could score. Even Ryan did that, regularly changing the speed of his fastball from the high 80s to mid 90s, but occasionally throwing that 100 mph heat that the hitter always had in mind.

Second, we're likely looking at the wrong numbers. Innings and pitches tell us next to nothing. At lower levels, like Little League, a pitch count is often the most accurate measure possible and we've seen an increase in injuries for travel teams and the like that don't have any limits. One better measure is called acute to chronic ratio and was developed by Dr. Ben Hansen for pitching specific monitoring. Originally developed in Australia by Dr. Tim Gabbett for sports like rugby, A:C Ratio is exceptionally useful and accurate in studies. We can even use it to look backwards.

Thirteen innings. Nineteen strikeouts. Ten walks. Nolan Ryan's outing on June 14th of 1974 was a stunner, especially given the pitch counts of modern baseball, but even in 1974, it was an extreme outlier. Pitchers simply did not go 200 pitches, even in the "single pitcher" era when complete games were the rule rather than the exception.

Ryan is, of course, an outlier in almost every way, but what lessons can we learn not only from this exceptional outing, but his 1974 season? Variously reported as 235 and 238 pitches, almost everyone can agree it was a lot.

According to several studies, the Acute to Chronic Ratio (ACR) is key to understanding and managing workload. Dr. Ben Hansen helped me dig into the numbers from 1974 to see if modern workload management can be used to explain Nolan Ryan and give us lessons on how to manage pitchers now.

First, there was some math. While 1974 isn't in the distant past, we don't have accurate pitch counts for players in this era. That forces us to use a pitch count estimator rather than accurate pitch counts. It's shown to be accurate enough that we feel using it is acceptable for a format like this. (Ryan's noted 238 pitch game was estimated at 242.)

Hansen then ran those estimated counts through some calculations and created this chart, which shows the season long ACR for Ryan:

[Chart of Ryan’s A:C Ratio created by Ben Hansen/Motus Global]

Surprised? I was as well. Ryan is seldom outside the "safe zone" between 0.8 and 1.2. He's regularly at the high end of this range. The spikes outside the range are often followed with short outings and those short outings often seem to work against Ryan's health, which is counterintuitive.

In fact, Hansen highlighted one part to me, showing how the rest period for the All Star Game worked against keeping Ryan's ACR ratio in the safe zone. There are also a couple of low-count outings of 73 and 28 pitches that had noted effects on his ACR. (That 28 pitch outing came on two days rest after a 158 pitch outing!)

Yes, that does indicate that if a pitcher has a quick outing, they should either "complete the workload" in the pen or be made available in relief during the following days. This is again counterintuitive, especially given modern protectionist methods, but the underlying science of workload management is well proven in many areas. It remains to be seen how well it will work in pitching, but early studies indicate a major reduction in arm injuries, even in counterintuitive scenarios.

There's likely some upper limit to this. Even Nolan Ryan couldn't go 200 pitches time and again, but a high workload can be sustained if a player can properly build up to it and get proper recovery between outings. This was also suggested in previous research I did back in 2005 for my book "Saving The Pitcher."

The 200 inning number was a need for any ace, but research showed that the more important number was 190 innings. When a pitcher went above that level, he tended to stay there for years, until he didn't. Few got back to that level once they dropped below it again. While ACR in terms of innings is a bad measure -- an inning is not a consistent measure of anything -- it is an indicator that the underlying real measure was being suggested in other ways, like inning limits.

There are a few caveats to this that, even as a pure mental exercise, should be pointed out. First, the pitch counts are estimates. While reasonably accurate, having actual throw counts would be significantly better. Second, these are *pitch* counts only. Every throw counts, so we don't know how Ryan's side work, which became very noted later in his career, would have affected these ratios. Third, the early season ACR being outside the safe zone should be ignored. Ryan had a spring training where he was throwing and building up his arm, which can't be included in this example.

Of course, we shouldn't discount how unusual Nolan Ryan was. I reached out to Ryan for comment, but he declined. Ryan did things few can do, but perhaps the durability was more than just being a genetic outlier. Few remember that Ryan did eventually break and had elbow issues throughout his career.

These noted, it's clear that Ryan's 1974 season indicates that heavy workloads – much heavier than what we see in 2018 – can be handled safely. Not everyone can be Nolan Ryan, but many pitchers and coaches could learn an interesting lesson from Ryan's workload and workload management.

Yes, this still surprises me, even when I read it now. It was one of those points in the book where I went, “ok, I did not know that before.” Which was why I wrote the book. It certainly wasn’t for the money.

So if we take this as reasonable, and I believe it is, could we work people higher? I still say we have to account for the freak factor. We remember Nolan Ryan, but not Dick Selma, his long time teammate, or Andy Hassler, who pitched with him in that ‘74 season. Frank Tanana will be remembered by more, but they’ll also likely remember that Tanana’s arm nearly fell off under the workloads that Ryan handled.

And of course, Ryan had a significant UCL tear in the 1980’s and decided to pitch through it. Here’s a fun fact - Sandy Koufax had Tommy John surgery. According to Dr. Frank Jobe, who I interviewed for a long, award-winning piece, Koufax said his elbow bothered him playing golf, so he repaired it. I’ve never been able to confirm this, so if anyone has Koufax’s number, please check for me.

For me, the best answer to the question of deploying pitching assets comes down to ‘all of the above.’ One assistant pitching coach I spoke with compared it to The Bear. “Chaos menu. Always changing. New menu every day,” he said with a laugh. “Ideally, I’d do that with pitchers. Who’s up, who’s down, who’s full-go, who’s half. Who’s on a throw day. I have 27 outs to get and why do we care who’s where? Get outs.”

The problem is that the pitchers are humans. They like roles and routines and even with the best workload monitoring tools, we don’t have a gauge on their arm that works like a battery monitor on a phone. Some teams are using Statcast and other biomechanical systems more like a warning sign than a monitor. “We watch our system and we have triggers,” said one NL team’s analyst in charge of systems. “The things we all know, like release point and extension, but we have several measures that we’ve developed internally. To me, the most exciting is simply being able to detect change. Is this pitcher the same today as he was last week? Is he the same in the sixth inning as he was in the second? If not, what changed?”

These “change detections systems” are becoming more common, but as with sabermetricians, the question is often how they’re listened to and integrated into the baseball system. The analyst worried much more about that. “I can’t send something to the dugout under current rules, but there are ways to get a message to the pitching coach or manager. The problem is, that pitching coach is getting more direct feedback from the pitcher. ‘How do you feel’ shouldn’t overrule the objective data, but it still often does, even with a good coach who understands this kind of thing.”

If a team could give up on roles, or used available tools to decide how to piece together a rotation that didn’t have a purely defined rotation or relief roles, it could work, but I don’t think it’s viable. It would require a significant amount of fan education, player buy-in, and it would have to work out of the gate. A losing streak or an injury would start up the outrage machine.

Another former GM, now a special assistant, questioned whether this isn’t the exact area AI might be better at than humans. “If we had all this data - workload, lineups, fatigue levels, upcoming games - isn’t this the 4D chess stuff that [AI] is supposed to be good at? Couldn’t you just let the machine set it up and then manage in game?” he asked. “Even so, will we be able to keep the machines out of the dugout at that point? It’s one thing having the manager and the bench coach going on gut and another having Google Cloud doing the calculations and predictions in near real time.”

That adds a new element to the already complex system, but one that could make such complexity manageable. It could be that we’re just at the start of coming to an answer to the question that has vexed baseball for most of its existence, which would make the last twenty years a lost period, wasted time and ruined careers. I asked one question to everyone I spoke to as a close, greeted most often by silence:

What if we did more harm than good over the last twenty years?

Special thank you to everyone that took the time to speak with me. Even if not directly quoted here, the conversations colored the piece and each one was valuable in its own way. Some of the best aren’t quoted here because they were so insightful but complex, or because examples given were identifiable to those in the know.

*I’ve been wondering how the old style game where you didn’t get to keep foul balls worked. Did they just toss it back down and put it in the bin with a handful of others, or did the pitcher and batter have to wait for that one ball to make its way back? I asked John Thorn and the answer is … they waited. Wow!