MLB Just Tried a Bunch of Experimental Rules in the Minors. How Well Did They Work?

As the old IBM ads used to say, You Make the Call: Should this pitch have been a strike?

It’s clear from the batter’s reaction that he didn’t think so. For what it’s worth, the pitcher didn’t either. “Maybe that ball clipped the zone, but I missed my spot by 2 feet,” says Craig Stem, a 31-year-old right-hander, who delivered the pitch for the independent Atlantic League’s High Point Rockers on October 10. “No human ump would call this. I don’t think that’s good for baseball.”

A human ump put his hand up to signal that strike, but a computer made the call. The Atlantic League, the country’s highest-level indy league and the first “partner league” of Major League Baseball, used the automated ball-strike (ABS) system this season, just as it did when the system debuted there in 2019. The TrackMan technology that powers the system in the Atlantic League doesn’t care about missed spots. It considers location alone. One could, of course, argue that whether the pitcher hit his intended target and how the catcher caught the ball should both be irrelevant; all that should matter is whether the pitch crossed the plate. Stem, a former Dodgers and Marlins farmhand who has pitched all over the world, has heard that argument, though he doesn’t subscribe to it: “I think those people would change their mind if they saw it in action,” he says. “It’s pretty ugly to see.”

In this case, appearances (and offset broadcast camera angles) aren’t totally deceiving. The “Did it cross the plate?” purity test would also suggest that the pitch wasn’t a strike. Yet it did intersect the strike zone, as indicated by the location of pitch no. 1 in the graphic below. It’s just that the zone in this part of the Atlantic League’s season was set such that it extended slightly off the edges of the plate.

In other words, this wasn’t a glitch. The pitch was a strike because the zone had been widened—which, ironically, was partly a product of feedback from players. In 2019, the Atlantic League zone was the width of the plate (17 inches), and on Opening Day of this season, it was expanded to 20 inches. On July 10, the zone was stretched further, to 21 inches. Thus, some pitches off the edges of the plate were supposed to be strikes.

In fairness to the robo ump, an identical call on a different pitch in the same location might not have seemed so flagrant. “Human umps call that all the time, but typically only if that’s the spot the pitcher was going for and the catcher received it well,” Stem says. “The way it works with ABS is not the same.” The reaction from the fans was the same, though. “If an umpire called this pitch a strike, he would be booed out of the stadium,” Stem says. “People were even booing the TrackMan.”

The Atlantic League’s ABS system has produced plenty of divisive calls, though not necessarily more than, say, plate ump Laz Diaz’s scattershot performance in Tuesday’s ALCS Game 4 (the latest reminder that the postseason zone is often just as amorphous as the regular-season zone). Robo umps just make different divisive calls. Even without the traditional “human element,” the strike zone remains a living document. In effect, the old human element of umpires is being replaced by the new human element of MLB executives who are trying to determine which size and shape the strike zone should take. And it’s difficult to take a strict-constructionist stance on what is and isn’t a strike when the zone is repeatedly reconstructed.

The ABS system, which was also installed this season in the Low-A Southeast (a first for affiliated ball), was just one of a wide array of experimental rules implemented in pro ball below the big league level this year. Major League Baseball’s recent takeover of the minors hasn’t prevented minor leaguers from facing low pay and lousy living conditions (yet), but it did allow MLB to audition a number of new rules across several levels, yielding more and better data than it could have collected by trying out one rule per season or stacking multiple trials in the same league. These tests took place against the backdrop of MLB, which has seen longer games, falling batting averages, and rising strikeout rates (which finally stopped ascending—for now—thanks to sticky-stuff inspections). The league’s audience surveys, it has said, suggest that fans want more balls in play and more action on the bases. Doubles, triples, and steals are in demand, and strikeouts, pitching changes, and dead air between pitches—the hallmarks of notoriously slow-paced postseason play—have worn out their welcome.

Even if the league had the power to tailor the sport to preset specifications—if commissioner Rob Manfred, like an esport executive, could essentially patch the sport and set the strikeout rate, walk rate, home run rate, stolen base rate, and so forth at will—it wouldn’t know precisely where to place those rates to please the most people. But it does have preferences for the directions and degrees that the settings would move. “We do have some intuitive sense of roughly the magnitude that we’re going for,” says Morgan Sword, MLB’s executive vice president of baseball operations. This year’s tests were intended to help further refine those preferences and establish solutions.

“We ended the season with rules that we’re very excited about, rules that you’ll probably never hear from us again on, and some rules that need tweaking, but continue to have promise,” Sword says. Sword, who will soon help present recommendations concerning the various experiments to MLB’s competition committee, won’t specify which rules he believes belong in which buckets, but we can come to our own conclusions, based on stats and eyewitness testimony.

The following rules—some of which had gotten earlier looks in the Atlantic League—joined the ABS system on the list of tweaks designed to “increase action on the basepaths, create more balls in play, improve the pace and length of games, and reduce player injuries.”

• In the Atlantic League, the mound was moved back by a foot at midseason, to 61 feet, 6 inches. (The Atlantic League also tried out the starting-pitcher-promoting “Double Hook” rule, the effects of which are difficult to discern without access to game-level Atlantic League data from prior to 2021.)
• In Double-A, the infield shift was banned, first by requiring all infielders to position both feet on the infield dirt when the pitch was released and, later, by requiring them to be on the dirt when the pitcher was on the rubber and by dictating that two infielders be on each side of second base when the pitch was delivered.
• In Triple-A, bases were enlarged from 15 inches to 18 inches to a side; in High-A, pitchers had to step off the rubber to attempt a pickoff; and in Low-A, pitchers were limited to two step-offs or pickoff attempts per plate appearance. (A third unsuccessful attempt would be ruled a balk.)
• In Low-A West, a 15-second pitch clock was put in place.

Many of those changes have carried over to the prospect-rich Arizona Fall League, which started play last week. The AFL is using the ABS system (in games at Salt River Fields), the full shift restriction, the 15-second pitch clock, the pickoff-attempt limitation, and the 18-inch bases, which may hint at the rules that MLB considers keepers. Let’s examine each of the major 2021 experiments to see which ones worked and which ones still need work.

The Strike Zone

In theory, the strike zone is whatever the rule book says it is, although that’s subject to change over time. In practice, the zone has historically been whatever individual umps decided it was, which (depending on the ump and the era) may or may not have borne a great resemblance to its specified dimensions. Now that umps are evaluated and trained with pitch-tracking technology, the difference between the theoretical and actual zone is smaller than ever, especially among younger umps who debuted during the data-centric era. But the ABS system has the potential to close the gap between theory and practice almost entirely. The big question is what the computer-controlled zone should look like.

Even in the short history of the ABS system, the robo zone has gone through five permutations:

• 2019 Atlantic League (and 2019 AFL): The three-dimensional area over the 17-inch-wide plate, with upper and lower bounds at 56 percent and 28 percent, respectively, of batter height
• 2021 Atlantic League, until July 10: A two-dimensional pane, bisecting the plate horizontally 10 inches from the back tip, which was 20 inches wide and 51 percent and 28 percent of batter height
• 2021 Atlantic League, July 10 and later: A two-dimensional pane, still 10 inches from the tip of the plate, which was 21 inches wide and 53 percent and 28 percent of batter height
• 2021 Low-A Southeast, until July 20: A two-dimensional pane at the front of the plate (or 17 inches from the tip), which was 17 inches wide and 56 percent and 28 percent of batter height
• 2021 Low-A Southeast, July 20 and later (and 2021 AFL): A two-dimensional pane 10 inches from the tip of the plate, which was 21 inches wide and 51 percent and 27.5 percent of batter height

Pitch-tracking information from ABS-equipped leagues was publicly accessible this season, which means we can closely compare the robo zones to the MLB zone for the first time. The following GIF, based on TrackMan data from the Atlantic League and Hawk-Eye data from Low-A Southeast and MLB, shows the various zones, one after another. We’re displaying the early and late Low-A Southeast zones separately, but for simplicity’s sake, we’re conflating the early and late 2021 Atlantic League zones, which were almost indistinguishable from each other in partial-season samples.

Compared to the MLB zone, the early Low-A Southeast zone was tall and narrow, while the Atlantic League and late Low-A Southeast zones were short and wide. In terms of total size, the early Low-A Southeast zone was 7 percent larger than the MLB zone, while the Atlantic League and late Low-A Southeast zones were 5 percent larger, respectively.

One obvious way in which robo umps fail the Turing test is that the robo zones are rectangles, whereas the MLB zone—which is also rectangular on paper—is in practice rounded like a superellipse. Nature abhors a straight line, but robo umps don’t. That’s because human umps call pitches probabilistically rather than in the robo ump’s binary fashion. Human umps will call pitches in a particular location at the edge of the zone strikes sometimes and balls at other times, but a robo ump will (accounting for calibration differences) make the same call every time.

If MLB wanted the transition to the ABS system to be seamless, it could program it to mimic the behavior of human umps. But that would defeat the purpose of pursuing perfect consistency, and it could also come at a cost to comprehensibility. “In addition to the baseball considerations, there’s another thing we’re focused on, which is having a strike zone that is easily understood by fans—fans that are watching the game on television, fans that are following the game on the MLB App,” Sword says. MLB wants to ensure that “the call that ABS is producing matches the representation of the pitch location on all our different platforms,” so that a strike on the field also looks like a strike on K-Zone and Gameday (and vice versa). For that reason, Sword says, “We’ve been prioritizing rectangular, two-dimensional zones as we work through this process.”

The ABS system provoked player complaints in the Atlantic League and AFL in 2019, and the same was true to some extent in 2021. Stem says the combination of in-season adjustments and technology that’s “not perfected yet” was “incredibly frustrating” at times. In the Low-A Southeast, the system also drew mixed reviews. Broadcaster John Vittas, who did play-by-play for the Twins’ Low-A affiliate, the Mighty Mussels, says, “The ABS system was quite a roller coaster.” He heard more griping about the zone in the second half of the season, when he perceived more “egregious mistakes” and when the most common types of polarizing pitch calls migrated from the top and bottom of the zone to its horizontal borders.

“Those breaking balls at the top of the zone that were strikes had now dropped down to breaking balls at the knees that bent around the plate that were called strikes,” Vittas says. “It infuriated hitters.” Vittas, who was optimistic about the future of robo umps at the start of the season, concludes, “I’m in favor of change and technology in general, and so were our coaches. But by the end of the season, I think we were all very much in favor of human umpires and were about ready to completely write off ABS.”

Another broadcaster, Justin Rocke, called games for the Reds-affiliated Daytona Tortugas, the one team in the league that didn’t have the ABS system installed at its home park. The 50-50 split between human umps and robo umps was tough on the Tortugas. “Our coaching staff did not seem particularly fond of it,” Rocke says. “I know our pitching coach pointed out inconsistencies and glitches where they’d look at the strike zone overlay and a pitch that was a strike was called a ball. However, I did hear some pitchers say they liked the ‘black and white’ nature of the system.”

Some grousing is probably inevitable, for the same reasons that every website redesign seems jarring at first. If MLB had somehow used a robo zone from the start and was suddenly switching to a human one now, that would be off-putting too. But part of the problem—and part of the impetus for the midseason alteration to the zone—was the caliber of umpiring and pitching talent in Low-A, which lends itself to a more generous zone than that of MLB. “At that level, you don’t generally see umpires call a zone that tight, nor do you have pitchers who are generally consistent enough to pitch to a tight strike zone,” says former major league pitcher Joe Martinez, who was hired as MLB’s senior director of on-field strategy late last year. MLB opted to widen the zone in late July both to give wildness-prone pitchers more leeway and to encourage contact. “Some of the analysis that the team’s done … said that a zone that is a little bit shorter or compressed vertically and a little bit wider horizontally was more conducive to guys putting the ball in play rather than the swing and miss that you get north and south,” Martinez says.

Walks went down after the redefinition, but the strikeout rate stayed the same. Overall, offense increased, though that may have largely been because of warmer weather.

The cancellation of the 2020 minor league season, and the subsequent contraction and reorganization of the minors, make comparisons between 2019 and 2021 tricky, but the 2021 Low-A Southeast is typically compared to the former Florida State League, where all 10 of its squads used to play. The FSL was a High-A league, which makes the comp somewhat suspect, but offense was much more potent in the ABS-ified Low-A Southeast. The leaguewide slash line finished at .236/.344/.370, producing an OPS differential of almost 50 points relative to the 2019 FSL’s .242/.313/.353. The upticks in walks and strikeouts were also startling. Although those rates were up in all leagues compared to 2019, they increased more from the 2019 FSL to the 2021 LASE (especially prior to the zone redefinition) than they did anywhere else, which wasn’t the goal.

“I think we would need to get probably higher-level players to participate to really understand what the impact would be at the major league level,” Martinez says. (Hence the desire to put the system to work in the Atlantic League, the AFL, and probably beyond.) But even in Low-A, he says, it was valuable “to make sure the technology works, which it does really well.” That’s not to say that the robo umps never made mistakes: Of roughly 3,700 Low-A Southeast pitches with a modeled called-strike probability of greater than 99.9 percent, about 1 percent (41 calls) were balls, assuming the stored pitch coordinates match the ones that were used to make the real-time calls. (The system seems more prone to calling balls on should-be strikes than strikes on should-be balls.) The corresponding rate in MLB was about 0.5 percent, suggesting that ABS was roughly twice as likely as human umps to miss the most obvious strikes in the heart of the zone. This image of the center of the Low-A Southeast strike zone shows a smattering of balls in a region where any taken pitches should have been gimme strikes.

Another possible source of error: All of the ABS-powered leagues thus far have used team-provided batter heights, which are often inaccurate, to set the upper and lower strike-zone boundaries. But human umps make many howlers too, and they aren’t immune to height-related biases. It’s impossible to perfectly assess the accuracy of the ABS system using publicly available data, both because MLB updated some hitter heights during the season and because the rulings made in the moment may be based on data that differs slightly from the post-processed values MLB reports. According to the data we have, the ABS system called strikes on 8.5 percent and 6.5 percent of pitches outside the horizontal boundaries of the Low-A Southeast and Atlantic League zones, respectively, but because of slight differences between Statcast and the Atlantic League’s TrackMan setup (and possibly between the Statcast installations in MLB and Low-A), as well as the interleague divergence between two- and three-dimensional zones, it’s difficult to compare those figures fairly to MLB, where the same method suggests that 13.5 percent of such pitches are called incorrectly.

The ABS system has proved that it’s capable of making calls quickly, and there’s no reason to think that it can’t hew to a provided definition of the zone more closely than a human ump. There’s just no clear consensus on what the definition should be. Beyond that, robo umps would (mostly) eliminate manager-umpire dust-ups and render receiving skills irrelevant, which would be bad news for framing specialists and for fans of framing technique. But the ABS system could cause another unintended consequence: less competitive plate appearances. When human umps are calling pitches, the size of the zone fluctuates based on the count, as shown by this GIF of the 2021 MLB zone on 0-2 and 3-0.

Analysts have proposed multiple explanations for this persistent, sizable effect, but whatever the reason for umpires’ stinginess on 0-2 and largesse on 3-0, it helps even out the pitcher-batter balance within plate appearances. It’s not necessarily logical for the zone to morph from pitch to pitch, but it is dependable and, arguably, beneficial from an entertainment standpoint. Essentially, umps lend a helping hand to whichever party is currently down in the count, increasing the odds of a comeback.

This season, MLB hitters generated a .164/.198/.261 slash line after falling behind 0-2, eventually striking out in 47.3 percent of those plate appearances. After getting ahead 3-0, they walked 60.5 percent of the time and recorded a .723 OBP. Imagine how much more lopsided and predictable the results would be if umps weren’t putting their fingers on the scale. Actually, don’t imagine: Let’s look at the data. The table below shows batter performance after falling behind or getting ahead in the count in the Florida State League in 2018 and 2019 and in Low-A Southeast (with the robo zone) in 2021.

That seems significant! The OPS difference between batters being ahead and behind in the count went from 444 points in 2018 to 429 points in 2019 to 604 points this season. Batters were better when they went ahead (like Low-A equivalents of Bryce Harper) and worse when they fell behind (like Low-A equivalents of Evan White). All else being equal, that seems a little less fun.

Sword doesn’t dispute this potential pitfall, but he says MLB wouldn’t want to hardwire that existing subjectivity. “If you make the decision toward ABS, I think it’s unrealistic that you would preserve those smaller biases that currently exist in the strike zone as they relate to the count, and the game score, and the weather forecast, and all those things,” he says. “What you’re seeking when you go to that kind of a system is uniformity in all those different situations, so I think that we did not consider recreating any of those biases in the system.”

That’s understandable, but it’s another reason robot backers should be careful what they wish for. “The overarching thing I learned with the experiment was that the strike zone is a movable object,” Vittas says. “Hitters have different heights, stances, and swings. Pitchers throw from different angles. The score of the game matters. Pitchers make quality pitches and pitchers make bad pitches. All of that changes how the zone can be interpreted and how the players want the zone to be governed. In practice, the cookie-cutter ABS zone just didn’t work.”

The status quo is far from perfect, and your mileage may vary on whether its variability is a bug or a feature, but the AFL’s ABS demo hasn’t gone smoothly so far. The cookie-cutter zone could work, and it will likely make it to the majors, possibly before some of the players in the Low-A Southeast do. Maybe it’s time for robot checked-swing calls, too. As of now, though, there’s still some uncertainty about the ETA, the implementation, and the extent to which more mechanized baseball would be better baseball. For now, MLB will continue to test in search of the sport’s Goldilocks zone. As Sword says, “We’re certainly not at the point where we have a zone that we think is the one.”

The Mound

No experimental rule generated more trepidation this year than the Atlantic League’s MLB-driven decision to move the mound back by a foot from the strangely sacrosanct distance of 60 feet, 6 inches. There were real reasons to think this change would reduce strikeouts and that the decision was long overdue: It’s worked at various points in the past, in the majors and elsewhere, and pitchers are much bigger and harder-throwing than they were when the current pitching distance was set in 1893. There were also some suggestions that pitchers might benefit from increased pitch movement over a farther flight. Finally, there were fears that the change in distance would cause a spike in injuries, a concern that scuttled a plan to push the mound back by two feet in the Atlantic League in 2019, and nearly incited a revolt this season in the lead-up to the move on August 3.

“Initially, obviously, there was some negative reaction, just because it was a change and a change that some people viewed as kind of fundamental to the game,” Martinez says. “I would say the overall opinion after a little while with it was that it didn’t make a huge difference. You really didn’t hear a lot of grumbling, and most of the players and staff that we’ve been talking to have said that after a series or two, nobody even really talked about it anymore. I think it was a lot less disruptive than maybe people thought it was going to be. … Hitters said that timing was essentially the same.”

Stem backs up Martinez’s take, describing the mound move as “not a huge deal at all.” The veteran righty adds, “That’s the one that everybody was freaking out about, and once it was done, and after that first week when everybody quit talking about it, it was completely unnoticeable.” No pitch type’s usage rate shifted up or down by even 1.5 percentage points, and the average four-seamer release speed budged only slightly, increasing by about a quarter of a mile per hour.

On the plus side, one reason the mound move was unnoticeable is that it didn’t seem to spark an injury epidemic. “We do track injury rates in the Atlantic League for all of our rules tests, and there wasn’t any uptick in pitching injuries as a result of the move,” Sword says. That may make it easier for the league to convince future pitchers on non-standard mounds that they aren’t putting their careers at risk to be baseball guinea pigs. The disappointing news is that the move didn’t appear to have positive effects either, at least when it comes to strikeouts and walks. The league’s strikeout and walk rates rose slightly after the move, and overall offense was almost unchanged.

“We did not see the magnitude or even the directional impact that we would have expected,” Sword acknowledges. Martinez adds, “Given a larger sample size, maybe we’d start to see the behaviors that we thought we would come through and show up rather than what we’re seeing right now. I think we’re all just a little bit scratching our heads on this one, because there were a lot of people who were optimistic about it and thought it would have a little bit of a different impact.”

The encouraging news for mound-move proponents—including MLB consultant Theo Epstein—is that the percentage of swings that made contact was higher after the mound move than before (76.8 percent vs. 75.4 percent). The increase in K’s seemed to come from a corresponding increase in the percentage of taken pitches that were called strikes (28.3 percent vs. 27.5 percent). That could be because hitters and/or umpires needed time to adjust to slightly more movement. Then again, those first-half contact and taken-strike rates were somewhat skewed by anomalous early-season results from before the strike zone was adjusted.

All in all, the mound move wasn’t a clear-cut disaster or triumph. It may be that more time, a bigger backward shift, or a season without the confounding effect of a midstream strike-zone change would be needed to accurately assess its effects. MLB hasn’t decided whether it will try to extend the test, but as Sword says, “If your goal is to improve contact and reduce strikeouts, your options for trying to do that are fairly limited.”

The Pitch Clock

Here we have a rule whose impact is easy to see. Prior to this season, pitch timers had been tried temporarily or installed permanently in five affiliated leagues:

• AFL: Since 2014
• Double-A: Since 2015
• Triple-A: Since 2015
• Florida State League: 2016-19
• MLB spring training: 2019

In most cases, the time savings were tangible but transient. As Baseball America’s J.J. Cooper documented last month, average nine-inning game times have gradually climbed at almost all levels and in almost all leagues, despite sporadic successes in arresting or reversing their inexorable rise. “The first generation of pitch timer that’s currently used in Triple-A and Double-A was successful in reducing game time for a season, and then all of the gains were lost over the period of the next couple of years,” Sword says. “Players, coaches, etc., were able to learn under the prior version the way to get around those regulations and continue to play at their old pace.”

Inertia and entropy are difficult foes. Taking more time between pitches can help hitters and pitchers perform better, so it sort of makes sense for major leaguers to lollygag. It’s just not as much fun for fans. “It’s not the length, it’s the pace” is baseball’s version of “It’s not the heat, it’s the humidity,” but even a relatively brisk three-and-a-half- or four-hour game can be uncomfortably long, just as a dry 100-degree day can be uncomfortably warm. This year’s average of 3:10 per nine-inning MLB game is the highest ever, and the three-hour-plus 2021 averages at Triple-A and all full-season leagues combined are also unprecedented. Those records were set despite 20-second pitch clocks at Double-A and Triple-A (slashed to 15 seconds when bases are empty), shortened breaks between innings, greater vigilance about keeping batters in the box, and other, even more ineffectual efforts like limits on mound visits and the three-batter minimum rule.

So it was with only cautious optimism that some observers greeted the latest sally in the campaign against game bloat: a 15-second pitch clock (17 seconds with runners on base) in the Low-A West, a circuit composed mostly of teams from the former High-A California League. After the pitch clock started ticking on June 8, the league’s average time of game fell by roughly 20 minutes, from 178.2 minutes to 158.4.

That dramatic decrease in game time, fueled in part by a roughly two-second drop in the time between pitches, was achieved even though scoring increased considerably after the pitch clock began counting down. Average fastball speed didn’t go down despite pitchers’ reduced recovery time, so warmer temperatures—a dry heat—may have played a bigger part in the hitter-happy result. Offense spiked by 36 points of OPS after the same point in the 2019 Cal League season, compared to 61 points in the Low-A West this year.

So, will the speedier pace of play last? This time, MLB means business. Not only was there less time on the clock—and a requirement for batters to be in the box and ready to hit with eight seconds left, instead of seven at higher levels—but the league took steps to, as Sword says, “close all of the loopholes that existed in the prior version.” For one thing, Double-A and Triple-A pitchers could reset the clock without penalty by stepping off the mound, but Low-A West pitchers could do so only twice before a balk was called. The league is trying to be even more diligent about enforcement in the AFL. “We are more confident that this version of it … will be more durable,” Sword says.

MLB players have opposed the pitch clock in the past, but the rapid pace of play in Low-A won over doubters such as another former major leaguer and recent addition to the MLB C-suite, senior vice president of on-field operations Raúl Ibañez, who calls the stricter pitch clock the most exciting of the 2021 experiments. “I was not just surprised, but blown away by the pace, the quality of play, how crisp and fluid the game flowed, and not just the action that was involved in the game, but the frequency of action,” Ibañez says. “You had to keep your eyes focused on the game. It felt like a baseball game just with a really great tempo and rhythm to it. It felt like the game that I grew up watching in the 1980s.”

Joe Ritzo, broadcaster for the San Jose Giants, San Francisco’s Low-A West affiliate, says enforcement got a little lax late in the year but that the clock was a winner until then. “Our manager [67-year-old former major leaguer Lenn Sakata] seemed skeptical at first but definitely came around to the clock, and he is about as old school as you can get. We went almost two months without a three-hour nine-inning game at one point this season. Just expected 2:30-2:45 every night after a while.”

Not every ex-player or active veteran would take to the timer so readily, but seven years after the AFL broke the seal, the pitch clock is close to conventional. At this point, only seven—yes, seven—of the 1,508 players who appeared in the majors last year have had no exposure to the pitch clock in an affiliated, domestic league: Kyle Cody, Kris Bubic, Garrett Crochet, the Astros’ Luis Garcia, Shohei Ohtani, Andrew Vaughn, and Ha-Seong Kim. Exclude 2019 MLB spring training, and the count climbs to 37. Exclude that spring training and set a minimum of more than five pitching appearances or 15 non-pitching appearances—which rules out some veterans who had brief rehab assignments or assorted other stints in a pitch-clock-equipped league—and the tally tops out at 184, or only 12.2 percent of 2021 major leaguers. That count includes some of the most senior, influential players in the league and the union, but it probably won’t be long until the pitch clock no longer seems scary or transgressive and the league can persuade the players to accept it without trying to attach it as a rider to a lopsided labor proposal.

The Running Game

Thanks primarily to two factors—a run environment in which singles are historically rare and homers are historically common, and analytically oriented tactics that promote conservative, shrewd, and boring baserunning—stolen-base rates are down. Three minor league measures this season were wholly or partly intended to address that trend, and all three seem to have done the trick.

Stolen base attempts per nine innings and stolen bases per nine innings were roughly flat or slightly down from 2019 to 2021 in Double-A and the majors. But they were slightly up in Triple-A (where bigger bases were deployed, partly as a safety measure), significantly up in High-A (where the step-off rule was used), and way up in Low-A (where pickoff attempts were capped at two per plate appearance).

Success rates, meanwhile, were up everywhere, but nowhere more than at Low-A.

Most of the increases in attempts and swipes in Triple-A probably weren’t attributable to the slightly bigger bases: The larger models were used in only one half of each Triple-A season, and the combined figures for team stolen bases per nine innings (.76 vs. .75), steal attempts per nine (.99 vs. .99), and stolen-base success rate (76.2 percent vs. 75.3 percent) were hardly any higher in the halves with 18-inch bags. However, the images above actually undersell the effect of the High-A step-off requirement, because that rule was rescinded on July 13 (which Sword says was by design, in order to measure its impact). After the traditional pickoff motion was restored, the attempt rate per team per nine innings fell from 1.69 to 1.26, the stolen-base rate dropped from 1.30 to 0.93, and the success rate sank from 77.2 percent to 73.9 percent.

“What we were looking for on the running game changes was a meaningful uptick in stolen base attempts and the success rate on those stolen base attempts without creating something ... that was too dramatic,” Sword says. “Reasonable people could disagree about what [‘too dramatic’] is, but what we saw with the step-off rule [and] with the pickoff limitation was something on the order of a little under an extra steal per game. I think that is a great result. I think it’s really hard to argue that one extra attempt per game is dramatically changing the game, but both of those rules did create the effect of enhancing the running game and reversing a long trend of fewer steals.”

That’s difficult to deny, though both rules caused some (literal) missteps and engendered some blowback from players. “The pickoff rule took some adjusting for our pitchers,” Ritzo says, recalling a key early-season balk. “Baserunners were sometimes thrown off too—[they] would just take off after two pickoff [attempts], thinking they could just go, and were easily thrown out.” After about a month, though, “everyone had adjusted,” and the limit “absolutely moved the game along.”

The step-off experiment didn’t go so smoothly. “It was not well thought of in the slightest,” says Jesse Goldberg-Strassler, broadcaster for Oakland’s High-A affiliate, the Lansing Lugnuts. Goldberg-Strassler recounts a conversation with an umpire who told him that players and team personnel were bewildered about what was and wasn’t legal, which led to a lot of balks, ejections, and confused looks. The ump began creating and distributing diagrams and talking things through with teams to convey what was allowed. “Because of this widespread lack of understanding,” Goldberg-Strassler says, “there was a great feeling of relief when the rule was scrapped after the first half.”

Because the rule wasn’t paired with a pitch clock, it also led to long delays as pitchers held the ball to try to disrupt runners’ timing, usually in vain. Although steals were much more common, Goldberg-Strassler says, “players were frustrated and the larger reaction from spectators was that games/pace was slower than it had ever been.” The rule also posed a problem for pitchers who were promoted into and out of the league and forced to learn new habits on the fly. Still, the steal-promoting measures accomplished their primary mission. If MLB saw the same percentage rise in stolen bases that Low-A or High-A did, the league would undo approximately 20 years of declining larceny.

The Shift

Like the mound move, the Double-A shift ban produced inconclusive results. “We did not see a large effect of limiting shifting at the Double-A level,” Sword says. Compared to 2019, Double-A BABIP was up by about six points. That’s not nothing, but it’s not conclusively a byproduct of banning pronounced shifts. Sometimes BABIP bounces a certain way for inscrutable reasons, as MLB’s BABIP did when it climbed six points from 1993 to 1994 or from 2005 to 2006.

Two other factors make the connection between the experimental positioning rules and the modest BABIP boost seem tenuous. Sword says hitters who are traditionally shifted more often enjoyed bigger BABIP boosts, but right-handed batters’ BABIPs were up slightly more than left-handed hitters’, even though the latter tend to be shifted more frequently and effectively. And the Double-A BABIP actually dropped by a bit more than a point in the second half of the season, when infielders’ movements were restricted even more.

It’s possible that the same rule would make more of an impact in the majors, where shifts are more common and precise. “It’s actually not trivial to write a shift restriction,” Sword says, so “what we accomplished this year was getting to a rule that we think works in practice and that we could enforce at the major league level if we wanted to.” There is some evidence that optimized positioning has contributed to the eight-point drop in MLB BABIP over the past few years, but it’s not a given that banning infield shifts is the answer to a problem that may have more to do with changes in outfield alignments. Plus, banning the shift would in one sense be more boundary-breaking than moving the mound. The pitching distance has been moved back before, but the rules have never specified where defenders can or can’t stand in fair territory.

Nicholas Badders, the play-by-play voice of the Northwest Arkansas Naturals (the Royals’ Double-A affiliate), says the rule was well-conceived and unobtrusive. “Something about it felt right,” he says, “especially in the second half of the season. And after a certain point, I really didn’t notice it.” The absence of extreme shifting faded from his notice until after the season, when out-of-traditional-position players robbed his hits in MLB the Show and he thought, “OK, yeah, this is why they were testing the rule out this year, to prevent stuff like this from happening.”

Non-computer players Badders spoke to in the Double-A Central were largely blasé about the shift crackdown. “For the most part, nobody felt strongly one way or another,” he says. “Middle infielders I talked to didn’t seem to indicate it was much of an adjustment.” Even in the second half, shortstops and second basemen could still steal hits up the middle by shading toward the opposite side of second and standing a step away from the bag, then crossing into the pre-pitch no-man’s-land after the ball was put in play. The Naturals, Badders says, were “still shifting every play, just not to the extremes that they otherwise might.”

To return to Sword’s hierarchy of rules to be excited about, rules that need further tweaking and testing, and rules to, well, rule out, here’s how we would sort this summer’s experiments:

• Inspired: 15-second pitch clock; larger bases; limited pickoff attempts
• Wired: ABS system; mound move; shift ban
• Tired: Step-off rule

As the CBA nears its expiration date, the on-field future of the game could follow any number of branching paths. Although the process of throwing rules at the wall wasn’t always smooth and sometimes put young players in compromised positions, the results will help chart the course of the sport. On the MLB side, Martinez says he will work to “overlay that former player perspective” from him, Ibañez, and another new hire, Rajai Davis, on top of the analysis of Epstein, Sword, former Marlins president of baseball operations Michael Hill, and Reed MacPhail (bolstered by the number-crunching of baseball operations analyst Josh Keen). That group, in consultation with committees and figures from the union, will keep collecting, dissecting, and discussing the data. Eventually, they’ll decide which rules will live on, and which will find themselves on the scrap heap along with a discarded 2019 experiment from the Atlantic League that would have come in handy last week: “‘Check swing’ rule made more batter-friendly.”

Thanks to Lucas Apostoleris of Baseball Prospectus for research assistance.