clock menu more-arrow no yes

Filed under:

The Quest for the Perfect Pitching Machine

At least three parties are working to supply teams with the next generation of batting-practice tools—ones that can replicate any pitch in MLB—either this season or next. Could they level the playing field between hitters and pitchers?

Ringer illustration

On December 15, 1896, a crowd of undergraduates gathered in the Princeton University gymnasium to witness the first public demonstration of what one account described as a “discovery [that] will revolutionize the game of baseball.” The first pitching machine, or “mechanical pitcher,” was the creation of Charles Howard Hinton, an Oxford-educated math professor at Princeton. In addition to daydreaming about beings that he thought might exist in the fourth dimension and adopting an alias to take a secret second wife, Hinton was a “baseball enthusiast of the highest degree.” While watching Princeton play Harvard that spring, he had conceived of a device that would allow hitters to take as many swings as they wanted against game-speed deliveries without wearing out pitchers’ arms. The resulting concoction, which the Chicago Tribune called “a cross between a cannon, a rifle, and an electric battery,” used gunpowder to propel baseballs at variable speeds, with finger-like protrusions on the cannon’s muzzle that could make the ball curve in any desired direction.

The exhibition at Princeton, The Boston Globe reported, was “a complete success.” In practice, Hinton’s mechanical pitcher was as big a bust as Brien Taylor, despite its supposed “unerring accuracy” and repeated trials by multiple college clubs. For one thing, it took too long to load and discharge; the Tribune estimated that were the launcher to replace the pitcher, it would take “nearly six hours to play a nine-inning game,” which was considered unacceptable prior to the 21st-century Yankees. For another, as Popular Mechanics recounted in 1913, “the hot gases soon acted on the covers of the balls and made them as hard as bricks.” In a related drawback, hitters weren’t that eager to face a fricking cannon. When Hinton died in 1907, the New York Sun noted that “the famous ‘baseball gun’” had been discarded “on account of the fear it inspired in the batter.”

But the brilliant British bigamist had identified a genuine need. In any given game, Hinton pointed out in 1896, “the amount of batting which any one player gets is very small. Thus, in the art of striking the ball … there is a great lack of preparatory exercise.” The same dilemma persists today. As pitchers apply the latest training techniques, technological tools, and physics epiphanies to get better by the year, hitters are hard-pressed to prepare for the unprecedented speed and spin that they’re seeing solely by staring it down in increments of roughly four pitches per plate appearance.

To help hitters keep pace with their sworn pitching opponents, teams are pursuing innovative practice solutions ranging from the high tech (virtual reality) to the low (ex-pitchers employed to toss high-intensity batting practice). And driven by recent breakthroughs in tracking the behavior of baseballs in mid-flight, a few of Hinton’s intellectual descendants are closer than ever to completing the quest that their progenitor undertook 125 years ago: to develop a machine that can replicate pitches precisely and convincingly simulate game situations. They believe they’ve developed the power to perfectly mimic any major league pitch. And they’re planning to deploy their products—which might fulfill the mechanical pitcher’s potential to revolutionize the game—as soon as this season, with no gunpowder or cannons required.

The Pittsburgh Press, 1900

For hitters, help can’t come too soon. In 2020, MLB’s strikeout rate rose for a 15th consecutive season, to a record 23.4 percent of plate appearances, and the league batting average sank to .245, its lowest level of the DH era. The dinger-inflating effects of an extraordinarily lively ball kept scoring robust, but the shortage of balls in play provoked complaints by fans and media members about an increasingly static brand of baseball. Those trends have been exacerbated by anti-hitter trends in technology and analytics, including the burgeoning science of pitch design, improvements in pitch selection and location, advances in velocity development and biomechanics, an emphasis on seeking and creating spin and movement, aggressive pitcher usage, and data-driven positioning and pitch-framing.

Baseball’s recent player deployment and development overhauls have “definitely favored pitchers this decade, rather than the hitters,” says Colin Young, a former minor league pitcher who topped out in Double-A and the Atlantic League before he was hired to throw live BP to the 2019 Padres. Pitchers control the parameters of each pitch, whereas hitters have to react and adapt; although they’re ostensibly on offense, they’re always underdogs who have a limited influence over the terms of engagement, and strikeouts usually start edging up when pitchers’ power is left unchecked. One way to help even the pitcher-batter balance without moving the mound back or shrinking the strike zone could be by enabling and inculcating better batting practice. “I think the hitters are starting to come around,” Young continues. “More companies are gearing towards hitters on the tech. If you could get in there and have a machine that throws basically a certain pitcher’s repertoire … I don’t know if we can ever get Yu Darvish’s repertoire, all 10 pitches, but if you can simulate that stuff and see it beforehand and then in the game recognize it, that’s an advantage.”

These days, Darvish’s arsenal goes to 11. But even if he adds a dozenth pitch type, the next generation of pitching machines could be capable of recreating them all. Joshua Pope is the president and founder of Trajekt Sports Inc., a two-person (full-time), Toronto-based company that’s demoing its new-age pitching machine for major league teams this spring. Except Trajekt doesn’t call its product a pitching machine; the big black box—a blend of baseball aerodynamics, precision motor control, image processing, and machine learning—is the “T1 pitch replication robot.” The robot indirectly came to be because of Marcus Stroman, who debuted for the Blue Jays in 2014, when Pope was still in high school. Stroman’s arrival prompted Pope and his friends to discuss how many swings they would need to take to hit one of the phenom’s pitches. Starting in 2015, Pope noticed pitch data from MLB’s brand-new Statcast system popping up on baseball broadcasts, which gave him an idea. “I said, ‘Hey, if the data exists, and we can extract the launch conditions for any pitch thrown in a game, surely there can be a machine that can impart those launch conditions,’” Pope recalls.

After getting a degree in biomedical engineering, he set out to design such a machine, which would theoretically allow him to answer the question about swings against Stroman without facing Stroman himself. He discovered that the previous state of the art, while light-years beyond Hinton’s prototype, didn’t allow for that level of detail. “The existing machines weren’t really designed for high precision,” Pope says. “They were designed for many reps and a volume of training, as opposed to a very tailored approach to training … that used the ubiquitous pitch data collected by all these teams.” Humans, Pope says, are pattern-matching machines, which led him to conclude that “if you just had so many reps against your opponent pitcher before the game, surely you should see some increase or improvement in performance. … The more closely your practice replicates the game environment, potentially you’re going to see more transferable results to the game.”

Given that hitters improve against pitchers the more times they face them within a single game and across careers, that’s not a wild supposition. As the Mariners’ Mitch Haniger said earlier this month on an episode of FanGraphs Audio, “The more you see a guy, usually the better your eyes will get against him. You know how his pitches move. I feel like your brain’s like a computer, so it’s picking up on things, even subconsciously.”

Barton Smith, a professor of mechanical and aerospace engineering at Utah State University and a pioneering pitch-physics researcher, says, “Why is it that third time through the order everybody starts to suck? I think a lot of it is because the hitters start to see the ball, they see the pattern better. And that’s pretty unique to each pitcher, and I’m sure they have to recalibrate every day.” With a pitching machine that could accurately replicate pitch data on an individual level, he adds, hitters might be able to bypass part of the in-game learning curve, which would be especially beneficial in an era when teams are pulling pitchers earlier to minimize the familiarity effect. “Now you’re going to go face a certain pitcher, and you’ve got his book. And … you could totally match what they’re going to do to you. And you could hit them like that in the first inning.” If companies can offer that capability, a hitter could visit an indoor cage during a game and tune up against the same stuff a pitcher threw on his nastiest day right before facing him for real. “That’s what they’re aiming at, which is a really amazingly exciting capability,” Smith concludes. “Just to say, ‘I want to repeat yesterday’s game.’ It’s a player piano.”

Trajekt is one of at least three parties hoping to supply next-gen pitching machines to a team or teams this season, along with a Latvian company called Winmill and Jeff Kensrud, the assistant director of the Washington State University Sports Science Laboratory. Kensrud, who says his machine can clone “the exact spin and speed of the ball so that it follows the exact trajectory that is observed in play,” asks, “If in the end we can produce a machine that is lifelike and is easy to use, and can simulate any pitch in MLB, why wouldn’t you practice on that? Why would you go to BP when you know it changes your swing, when you know it’s not exactly the way it is in the game? … You should be taking a thousand swings on the most lifelike device out there that can produce what you’re going to see in the game.”


One possible reason you wouldn’t is the weight of decades of dubious batting-practice tradition. Gripes about ineffective batting practice date back to before Hinton’s first mechanical-pitcher prototype: As the St. Louis Post-Dispatch lamented in 1887, “The right sort of batting practice is altogether too much neglected and not properly conducted. No wonder the pitchers always ultimately come out ahead. They play to outwit the batsmen, while the latter generally trust to luck and brawny muscles.” Although those muscles are much brawnier than they were when lifting weights was considered counterproductive, batting practice is still less sophisticated and productive than pitching practice. “Some see it as trying to get better or trying to get a good look at something,” Young says. “Some guys, they just want meatballs thrown up there.”

Some of that tendency toward meatballs wasn’t cemented until the past half century or so. “I suspect for a lot of history, say at least through World War II, BP fastballs were surprisingly close to game velocity,” says historian and sabermetrician Craig R. Wright. In those days, Wright explains, pitchers didn’t throw nearly as hard as they do now, and they typically paced themselves to go deep into games. Batting practice was often thrown by active (or recently retired) pitchers or hard-throwing position players, and “there was a mindset unaccepting of soft-toss BP.” (Which isn’t to say that practice was rigorous.) Wright remarks, “I’ve come across a few stories of old-time batters being injured by a batting-practice pitch. When is the last time you heard that about a modern BP pitch?” (Max Scherzer’s nose-breaking BP bunt doesn’t count.)

Fewer injuries would be a benefit. On the downside, BP arguably became a waste of time from the standpoint of skill development. By the early ’70s, BP was widely understood to be a rote exercise that presented scant difficulty. In 1973, pitcher Phil Hennigan said that if he got the chance to serve up Henry Aaron’s 715th homer, “I’ll tell him what’s coming. It will be a half-speed fastball down the middle. It will be like batting practice.” The same year, A’s manager Dick Williams let his high-school-senior son throw BP, which prompted Orlando Cepeda to say, “His fastball is a changeup.” By 1977, “batting-practice fastball” was an insult that Milwaukee manager Alex Grammas could direct at soft-tossing Tigers starter Dave Rozema (who soon had his revenge). Aged coaches were intentionally laying slow pitches over the plate from well in front of the mound, and BP was synonymous with ease, not challenge—the kind of practice that doesn’t make perfect.

Smacking moonshots in easy-peasy BP might help restore a slumping hitter’s confidence (and be fun for fans), but to get better, “you need to get a bit more variable and chaotic in practice,” says Arizona State associate professor Rob Gray, an expert in skill acquisition and motor learning who has served as a consultant to teams on improving practice. Even when hitters took BP against higher-velocity pitching machines, there was “a long history of setting the speed the same, over and over, and not varying things very much … I used to go to practices and I would call it a dance recital, like everyone knows where they should be, where the ball’s going. That’s not how you learn. You learn when you make mistakes and you get challenged.”

Gray says most teams are moving more in the direction of variable, deliberate practice, using differing locations, pitch types, and drills to develop skills (such as power and plate discipline) that were once thought to be largely predetermined. As part of that process, players who once shunned machines are learning to live with or love them. “With pro guys, a lot of them used to just hate hitting off machines, hate it,” Young says. Part of the shift away from that fixed, mistrustful mindset is attributable to a rising generation of players who’ve been immersed in machinery from the start. “Once everything gets normalized, then you don’t know any different,” Young says. “You see the younger guys use it, and all of a sudden you see some of the old dogs getting on [board], like, ‘All right, I’m going to jump in and try it. They’re coming for my job.’” But Young adds that increased adoption depends on improvements in technology, too. “It seems to be much more accepted now because of technology, we’re able to create that spin, create that simulation as to what they’re going to see that night. I think guys like to really get in there and see that and prepare.”

That’s where Trajekt, Winmill, and Kensrud come in. (Another advanced machine made by Shinobu Sakai of Japan’s Komatsu University was proceeding along a similar path last year, but Sakai says the coronavirus pandemic has postponed field testing.) Their machines are riding the crest of a wave of new insights into how and why pitches move that could separate the batting-practice pretenders from those that can truly replicate any offering down to the last detail. At the SABR Virtual Analytics Conference last Saturday, four consecutive presentations dwelt on data captured by the Hawk-Eye optical-camera-based tracking system that was installed in MLB parks in 2019 and introduced to the public in 2020. Hawk-Eye’s enhanced sensitivity relative to the TrackMan system that preceded it made it possible for Statcast to track the orientation and spin axis of the ball in flight—the angle at which it’s tilted as it rotates—rather than imperfectly inferring the axis from the pitch’s movement.

That new piece of information enabled a groundbreaking realization: Pitches didn’t move the way most baseball brainiacs thought they did. Before Hawk-Eye, analysts assumed that the movement of pitches—relative to a spin-less pitch acted on only by gravity—resulted solely from the Magnus effect, which pushes the ball in the direction that its front is traveling. But because Hawk-Eye measured the spin and spin axis in 3D, analysts could calculate the direction and magnitude of the movement associated with the Magnus effect. When they did that, they discovered that it often failed to account for the full movement of the ball. In fact, there’s another, non-Magnus effect that stems from the position of the seams. This force, which has come to be called the “seam-shifted wake,” occurs when “seams destroy the symmetry of the wake,” says Smith. But “this only happens if they’re in certain locations and there isn’t a seam on the opposite side doing the same thing.”

The seam-shifted wake is “not an isolated effect in a few pitches,” physicist Alan Nathan said during his SABR address on Saturday. “It is … very, very widespread.” Nathan went on to declare that “a new era has begun,” and although he acknowledged that “only the tip of the iceberg has been found,” he expects “seam-shifted wake” to enter the lexicon like “exit velocity” and “launch angle” did in the early days of Statcast. In sabermetric circles, it already has, and studies are revealing how much it may contribute to movement and effectiveness. On Wednesday’s MLB Innovation and Technology Showcase stream, the league’s chief operations and strategy officer, Chris Marinak, identified seam-shifted wake as “probably our most interesting innovation and advancement this year from a tracking standpoint.”

Baseballs can possess three types of spin: backspin; sidespin; and gyroscopic spin, or “gyro” (that’s “football spiral” spin). Gyro spin seems to enhance the seam-shifted wake effect. Current pitching machines can’t impart gyro spin, which means, among other things, that they can’t produce seam-shifted wake, the new baseball buzzword. “Batters recognize spin, at least to some extent,” Smith says. “You hear about seeing the red dot on the slider and things like that—[current] pitching machines can’t do any of that. So all the cues that a batter could take visually are totally messed up on a machine. And then you have the fact that the flight might actually be different.” In other words, even if current pitching machines can mimic a particular pitch’s movement and speed, the pitch still may not look quite like the real thing. And in some cases, the movement may be different.

Exactly how much this matters is still unknown; “We’re still trying to understand what exactly batters can pick up about spin,” says Gray. But Gray thinks the new tech can help hitters. When prepping to face normie major league arms, Gray says, a less sophisticated machine may suffice, but “when you get this weird pitcher that doesn’t follow the rules—they have tons of spin and the ball doesn’t drop as much as you expect—then you’re in trouble.” Seam-shifted wake may help explain how outliers like Cubs starter Kyle Hendricks, who consistently beats both his FIP and his preseason projections, baffle batters and predictive algorithms alike.

Martins Marenis, the CEO of Winmill (which built a hockey trainer before branching out into baseball), says his company partnered with an unspecified MLB team during development to get a grasp of what teams’ wish lists look like. “The team was actually very specific on what we need to achieve, and that is if you can do gyro degree on it, then you’re golden in a way, because that’s something none of the [existing] machines are able to do,” Marenis says. The team conveyed that gyro spin “gives that natural look” and “opens up the door to having the capability to design the pitch you want.” Marenis adds that during those conversations, seam-shifted wake is “something we’ve heard a lot of times, because just in the last year or so, there’s so much information that’s come out relating to this.” Kensrud confirms that “that’s the big talk now,” and that it keeps coming up when he speaks to teams. Pope says his T1 can control the orientation of the ball (and its seams) with a degree of precision that could help unravel the seam-shifted wake’s mysteries.

One MLB front-office official says his club is evaluating potential pitching machines based on three criteria: whether the machine accurately and repeatably replicates movement, whether it’s easy to set up in a way that releases the pitch at the pitcher’s true horizontal and vertical release point, and whether it creates an environment that stimulates the hitter’s perceptions in a manner that resembles a real-world setting. All three, in theory, should do well in the accuracy department, though they vary in a number of ways. The Trajekt and Winmill machines use wheels to accelerate and spin pitches, which means they may make more noise (and do more damage to the balls) than Kensrud’s solution, which uses a high-powered linear actuator with a “special and simple” device on the end. The current T1 is designed for a fixed installation in a covered or indoor pitching lane, whereas Kensrud’s creation can at least be wheeled out of the way on a pallet jack. The Winmill machine is the most portable and can be wheeled by one person. All three are designed to mimic most or almost all MLB release points, though those hoping to hit against robot Tyler Rogers are probably out of luck.

The third criterion has plagued pitching machines from the start. One report about Hinton’s contraption observed that “The great fault with the gun is that the batter is unable to see when the ball is coming and hence loses the practice of observing the lifting of the pitcher’s foot, the throw of his arm and other actions connected with the delivery of a ball,” which Hinton said he hoped to address by adding a mechanical arm. Without the visual cues of the pitcher’s appearance, some of the value of the pitch simulation is lost. The three batting-practice purveyors have all at least entertained the idea of pairing their pitch-flingers with video screens that would pair the ball-launching with well-timed footage of a pitcher.

Kensrud has his robot synced up with an 86-inch LCD screen that displays life-size footage of the deliveries of pitchers who’ve consented to be filmed at the lab. It sounds a little like a ProBatter machine display with a far fancier mechanical wizard hiding behind the video curtain. “When you stand at home plate and you look down at the large screen on ours, it’s weird,” Kensrud says. “It looks like a guy’s just standing in a black box throwing to you. So I feel like that lifelike nature of the device is what would finally cause someone to pull the trigger and say, ‘This isn’t really a pitching machine.’”

The three competitors are still validating, tinkering with, or revamping their products, which are bound to be costly. All three clam up when asked about price tags, but the custom-designed units are intended for elite-level training and will presumably be priced accordingly. (We’re a long way away from gyro robots at your local cage.) “If the Yankees have one of these, everybody will want one,” Smith says. Pope says he’s heard from at least half of MLB teams, but being intrigued isn’t the same as being sold. Members of two front offices say they don’t see the current model of the machine as a cost-effective solution, factoring in some skepticism about its control over seam-shifted wake (considering the damage wheeled machines tend to do to BP balls) and uncertainty about how big the batter benefit over non-gyro alternatives would be.

To solicit further feedback, Winmill has shared its trainer with some analytically inclined college programs, including the University of Iowa. Ryan Gorman, Iowa’s head student baseball manager, calls the gyro component “a huge plus” and says “The machine’s ability to replicate pitches is the best I’ve seen in my time here.” He praises its capacity to save programmed pitches in a playlist for a practice session; other machines can do that, he says, but the time they take to adjust between pitches (10 to 12 seconds instead of 5 to 7) and their lack of protective coverings that obscure their wheel alignments sometimes telegraph what’s coming next.

Ironically, one of the gyro machines’ selling points is that they may hurt hitters, too. Smith and Pope point out that these Trojan horse robots, ostensibly there to help hitters, could also help pitchers refine their repertoires by showing them what one of their pitches would look like if they subtly altered a grip. They could also help catchers—overburdened by having to work with more pitchers per game and per season than ever before—familiarize themselves with their own batterymates. Or, Kensrud says, the machines could cycle through pitch sequences against batters in practice to determine which is most likely to lead to strikeouts in games. “You may only need two pitches in your arsenal, but if you throw the two pitches the right way, and in the right sequence, you can strike everybody out,” Kensrud says. Just what MLB needs.

Hinton had the right idea, but he didn’t have the right product. The three gyro rivals are hoping they aren’t ahead of their time too. Marenis says Winmill is planning to roll out its product this summer; Kensrud says he could have his machine, which was completed within the past month, in one club’s cage in as soon as four months; Pope says he’s targeting delivery of up to five units by next spring training. For now, at least, baseball may be big enough for the three of them.


As they try to polish off their products, Pope, Kensrud, and Marenis are also jockeying for batting-practice position with two other rivals, one futuristic and one retro: virtual reality, and old-fashioned flesh and blood.

In a spring training game earlier this month, Oakland’s Gold Glove Matts, Olson and Chapman, took Cleveland starter Adam Plutko deep on the first pitch each of them saw from the right-hander. If nothing else, the back-to-back bombs were a fantastic piece of PR for WIN Reality VR, the company that makes the VR simulation in which Olson and Chapman had been swinging against a virtual Plutko about half an hour earlier. The sight of Plutko getting torched so soon after his teammates advance-scouted him via Oculus Quest was enough to make A’s reliever Jake Diekman fear for the future of pitching.

WIN Reality was cofounded by former Rockies GM Dan O’Dowd and his son, former minor league catcher Chris. The junior O’Dowd says the elder O’Dowd long envisioned VR as a possible treatment for the Coors Field hangover effect that seems to plague Rockies hitters when they travel back and forth between altitude and sea level. But the technology wasn’t really ready for its baseball closeup until recently. When the Rays started experimenting with VR five years ago, practice sessions took place inside virtual batting cages that couldn’t make road trips. Back then, the headsets on the market couldn’t deliver the visual fidelity and refresh rate necessary to see the ball. Even a year ago, the product required room sensors, calibration, and a connection to a laptop. Now, it’s a self-contained $300 headset running cloud-based software, which has significantly lowered the barrier for entry.

O’Dowd says WIN Reality is working with hitters from every major league team, who can face off against VR versions of any major league pitcher, complete with representations of their real release point and pitch characteristics (so far, seemingly sans gyro spin and seam-shifted wake). More than two-thirds of teams, he says, are using the system across both their big league and minor league levels.

O’Dowd says the goal is to be ubiquitous: “your glove, your bat, and VR.” And even though the company is aiming to be more of a complement to physical batting practice than a substitute for it, WIN Reality is still competing for players’ time, which means that “as pitching machines continue to take leaps, the pressure on VR to keep up with that is just as high.”

So how well does it work? Nobody knows for sure. “We’ve seen transferability effects of when you see 12 pitches before you face an opposing pitcher,” O’Dowd says, adding, “It’s like, skip one game of cards and see a few pitches, and you’ll probably play better.” Some evidence seems to support his contention that facing a VR version of a pitcher can improve swing decisions against the real deal, but the link between virtual training and in-game improvement isn’t well established among high-level players yet, aside from anecdotal testimonials. Then again, that goes for old-school batting practice, too.

Nick Wan, a former cognitive neuroscience researcher and data scientist for the Reds, says there’s potential for VR practice in baseball, like previous VR applications, to prove its worth as its physics and graphics grow more immersive. (In December, WIN Reality partnered with Diamond Kinetics to improve its bat-tracking, and sensor-based body-tracking may be next.) If you have “the right weighted bat and there’s good headphones for stadium-like noise and if you’re in a room that’s temperature controlled and has some sort of breeze simulation for outdoor effects, it’s not out of the question to get near-replication of BP or even the feeling of live at-bats without any of the physical costs,” Wan says. “But the simulation of all of this versus the transfer of it is the bridge you would have to gap, and right now that bridge barely has foundation.” As with other brain-training apps, Wan concludes, teams are “buying on the upside vs. the efficacy.”

That upside, at least, includes little injury risk, no wheeling or moving machinery, and the potential for unlimited reps against a recognizable pitcher. (It’s also easy to take BP in VR during a pandemic.) Unlike a pitching machine or a VR avatar, a human can’t keep throwing forever, especially if he’s trying to uncork big-league-quality pitches from a mound instead of a shorter distance. But for as long as his arm lasts, “having a real pitcher there that’s getting closer to the real speeds you’re going to face is obviously the best scenario,” says Gray.

It’s also the oldest scenario, and one that modern teams neglected too long. In 1927, Cubs manager Joe McCarthy promoted a rubber-armed minor league pitcher named Henry Grampp to the big league roster as a dedicated right-handed batting-practice pitcher. McCarthy instructed Grampp not to save his arm for games but to bear down in BP, a task he took so seriously that he studied and aped opposing pitchers’ styles and even appearances. The Jazz Age pitch replication non-robot, who pitched in only three games during his three years with the Cubs, became a folk hero and celebrity of sorts as Chicago won the 1929 pennant in his final season with the team.

The Coleridge Blade, 1929

NPB teams have long followed the Grampp plan and employed pitchers to throw purposeful BP, but it’s still the exception in the States. Last year, Reds rookie Shogo Akiyama attributed his early-season struggles in part to the team’s “monotone” approach to BP, explaining that his teammates liked to “hit clean hits that go far,” but that he preferred to “hit different pitches.” After Akiyama asked the Reds’ coaches to throw him more off-speed stuff, his bat rebounded.

In June 2019, the Blue Jays emulated McCarthy when they hired the righty-lefty tandem of Zach Stewart and Matt Tracy, two former major leaguers who’d been toiling in Triple-A and indy ball, to throw BP to the big club during homestands and occasionally to prospects in the minors and at instructs. “It was brand new to everybody,” Tracy says. At first, requests for his service were slow. “Not everybody’s really comfortable with making practice much harder and failing way more in practice,” he says. “So that’s a leap you’ve got to make there. You’ve got to go into it knowing that ‘I’m going to struggle, and that’s going to be good for me in the long run.’”

Gradually, batters bought in, and Tracy threw game-like pitches, sometimes from the mound and sometimes from 40 or 50 feet in the cage. “Some guys would say, ‘Mix in fastball curveball,’ or ‘Fastball changeup,’” Tracy says. “Some guys would ask for certain locations, too. Which at some point, I’m like, ‘Guys, if I could do all that, I probably wouldn’t be coaching.’” But the occasional misplaced pitch only added to the constructive chaos of the practice session.

Tracy is now Toronto’s pitching development coordinator, but Stewart is still throwing BP in the minors. Under Ben Cherington, who helped put Toronto’s live BP plan in place, recent retiree Jeremy Bleich has sometimes thrown BP in Pittsburgh. And in San Diego, Young was part of a system of providing live looks from former pro pitchers whom he describes as “definitely very different from when I was in the game before.” Another difference: “I made more money in one year throwing batting practice than I did pitching in the minors in like six,” he says. “I should have just been throwing batting practice the whole time.” And maybe San Diego should have been having him do it.

The two front-office members who expressed doubts about the value of the T1 robot noted that it might be cheaper to hire multiple ex-pitchers to throw live BP to each of an organization’s affiliates. At least one club this season is seriously exploring that possibility. Maybe the pitching robots’ toughest competitors are the arms that Hinton constructed his mechanical marvel to spare.

“The more knowledge we gain, the more things we can try and do to make players better,” Young says. The question, as always, is which type of players will benefit faster. Pitchers have had their day (or decade) in the sun. Now, perhaps, it’s hitters’ turn. “It’s definitely trickier for the hitting side, but they’re going to catch up, and this technology is going to keep getting better,” Tracy says. Either batters will embrace the latest tools and techniques and learn to counter hurlers’ latest tricks, or they’ll be left to flail in their seam-shifted wakes.