Johns Hopkins students develop technology to help Baltimore Orioles build better baseball bats

By Grace Grill

Click here for updates on this story

    BALTIMORE (WJZ) — Baseball is a game of numbers, and statistically, hitting a baseball is one of the hardest things to do in all of professional sports.

A Major League batter gets a base hit less than 30% of the time, and that percentage is much lower when it comes to hitting a home run.

But what if that statistic could change through AI innovation? That was the challenge the Baltimore Orioles presented to the Sports Analytics Research Group at Johns Hopkins.

“Over the years, I’ve just thrown out some projects that could be interesting,” said Baltimore Orioles Assistant General Manager Sig Mejdal.

Torpedo bat piques interest The Orioles do not fund any of the research and do not expect a return on any of the findings, but when a Major League Baseball team pitches a unique project, it’s sure to pique some interest.

“About a year ago, I think, [the Orioles] suggested that we find a way to measure bats with high precision,” said Professor Anton Dahbura, who heads the Sports Analytics group at Johns Hopkins University.

Dahbura has a background in baseball, and his interest only heightened when the Yankees went viral for the “Torpedo bat” at the start of the 2025 season.

“Like a lot of people, I’m saying ‘Why didn’t I think of that?’ because I played baseball for a long time and it’s just a given that the barrel of the bat is cylindrical and you know it just opens up a lot of possibilities,” Dahbura said.

The torpedo bat has been used for the past couple of years but blew up this season when the Yankees opened up the 2025 season, hitting 15 home runs in their opening series, with nine of those home runs coming from players who were using the non-traditional bat.

Developed by Yankees minor league hitting coordinator and former MIT physicist, Aaron Leanhardt, the torpedo bat redistributes weight to where the player is most likely to make contact, essentially optimizing its power or “sweet spot.”

However, each player is unique when it comes to making contact, which is why to find a player’s “sweet spot,” a bat must be measured with high precision.

“So right now the Orioles are taking a manual approach in measuring their bats,” explained Johns Hopkins computer science major Kevin Wu. “They are measuring a 34-inch bat at one-inch intervals.”

That’s a tedious process for human hands when you consider how many bats that would need to be measured for a major league baseball team.

Using AI to measure baseball bats Wu, and classmate Xiaojian (Jason) Sun, found a solution to speed up this process using AI computer vision.

“Computer vision is having the computer analyze what us humans see,” Sun said.

With the technology created by Wu and Sun, they can measure a bat down to .01 inches and get the quantified data within minutes with the touch of a button.

“All they would need to do is take a picture of the bat and hit enter using our system, and you get the results,” Sun said.

It sounds simple, but Wu and Sun said the biggest challenge was finding a surface for AI to get an accurate 3D reading of the bat. What they came to in conclusion was a matte green screen with the bat dispensed in the air in front of the screen using fishing wire. A process that took hours now takes minutes through their innovation.

“At the end of the day, we need to speak the language of the lathe that makes the bat,” said Mejdal. “The lathe just cares about the measurements throughout, and so ultimately, that’s what describes the bat, and until you measure it, you’re really unable to describe the design of the bat for whatever purpose you have.”

Please note: This content carries a strict local market embargo. If you share the same market as the contributor of this article, you may not use it on any platform.