Baseball Tech That Created the Most Buzz

SAN DIEGO — This week’s MLB Winter Meetings were about Aaron Judge, the Brinks truck the Yankees rolled out for him…and the robot on the fourth floor.

One by one, curious MLB teams took hotel escalator rides to demo the Trajekt Arc, a pitching machine that simulates every big-league pitcher, every big-league pitch and can even throw otherworldly 120 mph fastballs just for kicks.

Most of the execs who witnessed the so-called robot — straight down to its projection tool that can clone Gerrit Cole’s 6-foot-4 body — left shaking their heads at the thought of what this can do for baseball, or hitters in particular. Last season, the Chicago Cubs planted a Ferry Arc in the bowels of Wrigley Field, downloaded a Spotify-style playlist of every MLB pitcher known to man and let machine learning do the rest.

As a result, Cubs hitters were able to face that night’s starting pitcher at 3 o’clock in the afternoon. Literally up to 15 minutes before the first pitch, a hitter could step in against the robot and see exact replicas of the opposing pitcher’s release points, his repertoire and his velocity. Or they could recreate any of their previous at-bats against that pitcher in real time, as if it was déjà vu all over again.

Six other MLB clubs, purportedly including the New York Mets, also deployed the futuristic batting cage last season, and the company that brainstormed it, designed it and built it expects that number to grow to possibly 12 after these Winter Meetings.

“We do think we’re changing baseball with this,” says Jordan Rapps, Trajekt Sports’ VP of operations.

It was clear then at this week’s Baseball Technology Innovation Exhibition — the first held since the pandemic — that the sport’s latest inventions are designed to help hitters catch up to pitchers. Whether it was the virtual batting product Win Reality or the pitch recognition tool gameSense, tech is starting to favor the poor souls who have to make contact with 99 mph four-seamers.

“Look, I’ve been involved in these types of technology products, tracking, for a long time,” says Chris Marinak, MLB’s chief operations and strategy officer. “And what I’ve noticed is that the technologies historically have started with pitching. Ball tracking is historically the easiest thing to do and then pitcher tracking, too, because you don’t have a bat. You don’t have any other equipment to worry about.”

“But I think what you’re seeing right now is sort of a resurgence of technology to help hitters. The Trajekt system is a good example of that… So I think it’ll be interesting to see over the next couple of years how that translates back into hitter performance. And if it gives an edge to hitters maybe we haven’t seen in a long time.”

The science behind Trajekt is in its pitch design. From inception, its developers knew they would have to generate gyrospin, or riflespin, on the baseball in order to recreate Major League-quality sliders and slurves. “No machine had ever been able to do that before,” Rapps says. In fact, when the company’s Toronto-based founder Joshua Pope ran his original thesis by his engineering professor at the University of Waterloo, the teacher told him, “If you apply that force to the ball, you’ll actually tear the seams.”

Pope then collaborated with his future CTO and co-founder Rowan Ferrabee to discover that if they imparted equal force, or transverse force, to the baseball, they could impart gyrospin without simultaneously imploding the ball. The professor backed off, and Trajekt Sports was born.

Building the actual machine/robot was the next intricate step. Pope and Ferrabee developed a system that can hold 120 balls at once and uses automated slots to set the seam orientation of the pitch, whether it’s a Josh Hader fastball, a Jacob deGrom slider, an Adam Wainwright curveball or a Devin Williams changeup. Micro-cameras check the seam orientation or re-orient the seams if necessary, and then double-check it. If it’s an off-speed pitch, the whole machine will tilt up to produce the proper release point and location, and a conveyer belt will eventually usher the ball out to match the accompanying 3D video of the pitcher throwing — a life-sized human projection onto a green screen.

The end product, according to Rapps, is the ultimate replication of a live at-bat. The Major League teams that lease the robot upload Statcast’s in-game data to the Trajekt Arc software, allowing them to replicate the spin axis, seam orientation, spin rate and shape of every individual big leaguer’s pitch — to within 0 to 1% of pinpoint accuracy.

In other words, if a team is about to face Gerrit Cole, it feeds metrics from all of Cole’s pitches into the software, spin rate included, and has the robot throw each pitch four times. As of now, that’s all it takes for the robot to machine-learn how to be Gerrit Cole — four pitches.

Teams can also program the ball to land in the dirt or to break six inches, whatever their preference. And to verify a pitch’s accuracy, all of the Trajekt Arc’s are paired with Rapsodo systems that can validate the spin rates and velocities of the Major League pitchers they are simulating.

The use cases are endless. Five of the Trajkekt’s seven MLB clients set up the batting cage robot in their home stadium tunnels, another deployed one at its spring training facility and another at a minor league affiliate. That means that catchers, umpires and players rehabbing from injuries can all use it to sharpen skills.

Of course, some MLB teams are just as enthralled with Win Reality, a more established VR version of Trajekt that is activated through a Meta Quest 2 headset and is obviously more portable. There are multiple MLB teams and hundreds of players who simulate pre-game at-bats with Win Reality and often prefer it because they don’t need to run over to a batting cage.

“You can see any pitch, anywhere, anytime,” says Win Reality’s head of data, Chris Fiaschetti. “We know that the whole thing right now with baseball is don’t let the pitcher go through the lineup three times. Because numbers go up. So this is almost a way to expedite seeing the pitcher three times before getting in the batter’s box. “

Still another alternative tech tool is the gameSense product that, on a computer or app, shows video of a pitcher releasing a pitch and then occludes so the user has to guess whether it’s a ball or strike, fastball or off-speed. Through quizzes and repetition, players learn to identify the shapes of different pitches early in ball flight.

“What’s the most effective instructional design method or instructional technology ever invented?” asks Peter Fadde, gameSense’s chief science officer. “People want to say VR, AR, XR. No, it’s quizzing. That’s all we’re doing. The sound of learning on VR is ‘Wow, that’s so realistic.’ The sound of learning on what we do is, ‘Huh? Because you got it wrong, and your brain doesn’t like to get it wrong. And then you see the replay, and you go, ‘Oh.'”

Still, at these Winter Meetings, most MLB eyes were on the Trajekt Arc robot, which has invented pitches no human has even figured out how to throw yet.

“We’ve fired a 120-mph fastball,” says Ferrabee. “We’ve fired pitches that have so much speed and backspin, they rise. We’ve thrown 100 mph knuckleballs, pitches with 3,000 rpm spin that can break like you wouldn’t believe. We’ve created some pitches with up to eight inches of seam-shifted wake break. Inhuman.”

And worth the escalator ride to the fourth floor.