Revolutionizing baseball training with AI-simulated pitchers

Two University of Waterloo alumni are changing Major League Baseball (MLB) with a data-driven pitching simulator that replicates the real experience of batting against a professional pitcher.

Joshua Pope (BASc) and Rowan Ferrabee (BASc) founded their company Trajekt Sports in fourth year, supported by Dr. John McPhee, a professor in the Department of Systems Design Engineering. But the idea behind the tech dates to Pope’s time as president of the Athletic Council in high school.

“We were brainstorming fun sports activities when someone jokingly suggested taking swings against Marcus Stroman from the Blue Jays. This was around 2013, and we were big fans,” Pope said.

“I started wondering if there was a way for people to simulate hitting against their favorite pitchers. What tools were professional athletes using to train? Could I build a machine to replicate the pitching skills of real players?”

His curiosity led Pope to study biomedical engineering and gain the knowledge and network to realize his pitching simulator machine. The idea was too complex for his Capstone Design Project, but with support from Velocity, Waterloo’s startup incubator, Trajekt Sports was born. And then the world was thrown into chaos with the COVID-19 pandemic.

“It wasn’t exactly the best time to start a business,” Pope said. “But Rowan and I were determined and fortunate to have people who believed in us. We worked out of my parents’ garage with access to Velocity’s resources, had a grant from the Accelerator Center and raised more money in a ‘friends and family’ funding round to help get things off the ground.”

Traditional pitching machines are limited to basic speed and spin. Pope and Ferrabee focused on iterating their system to master variables such as velocity, spin axis and ball orientation to recreate any trajectory and pitch type. They also wanted to factor in optical stimulus, or what a batter sees when facing a live pitcher.

“Our approach is rooted in physics and first-principles thinking,” Pope said.

“Asking the basic questions like ‘What defines a flying ball?’ If we can control all the elements that determine its flight, we can accurately replicate a pitch. For us, it was a fundamental physics problem that needed a fresh take, so we built the solution from the ground up. Some might say we over-engineered it, but that last inch of precision is exactly what makes the replication so valuable for professional use.”

Their novel robot, the Trajekt Arc, is the only system that uses artificial intelligence (AI) to integrate ultra-realistic visuals of pitchers—complete with the exact arm angles, release mechanics and motion blur—to mimic the real experience of standing at home plate. This realistic training helps hitters prepare physically and cognitively for the game.

In December 2021, Pope and Ferrabee secured their first MLB client—the Chicago Cubs. With the Cubs on board, others soon joined and by the end of the following year, seven more MLB teams had signed on.

Today, 30 professional baseball teams train with Trajekt Arc machines—24 MLB teams, four from Japan’s Nippon Professional Baseball league, one from the Korea Baseball Organization and one from the Chinese Professional Baseball League. With 21 full-time employees and a co-op hiring pipeline that continues to tap into Waterloo’s talent, the company is exploring more ways to deliver a superior training experience.

Ideas include applications beyond professional baseball, like college training programs and recreational sports centers, and integrating advanced analytics, video tracking and personalized training plans for batters.

“At its core, the Trajekt Arc is about leveling the playing field, giving all athletes, from rookies to seasoned professionals, a way to improve their game,” Pope said. “I love sport, and I love that we’re changing how people play and train—for the better.”