Computer modeling at the University of Waterloo shows that professional baseball pitchers could make mechanical changes to avoid a common, career-threatening elbow injury without necessarily sacrificing competitive velocity.
"Our simulation found solutions that suggest there's untapped efficiency out there,” said Cedric Attias, who led the study as a graduate student in mechanical engineering at Waterloo. “Our goal isn't to tell pitchers to throw softer. It's to help them throw smarter."
Researchers built a detailed digital skeleton with muscles, ligaments, and joints to examine the extreme twisting forces exerted during the throwing motion on the UCL, a small band of tissue on the inside of the elbow that helps hold it together.
Their study, the first of its kind, revealed two main factors — a high arm slot, or angle and tilting the torso away from the pitching arm during delivery of the ball — that put the most demand on the UCL.
Here is an exclusive Tech Briefs interview, edited for length and clarity, with Attias.
Tech Briefs: What was the biggest technical challenge you faced while developing this digital skeleton?
Attias: Getting the simulation to actually complete a motion that resembles a realistic pitch was harder than it sounds. You're essentially asking a computer to figure out, from scratch, how to coordinate dozens of muscles and joints simultaneously to throw a ball at a specific speed, all while staying balanced and not falling over. The optimizer kept finding creative ways to "cheat" and produce motions that technically satisfied the math but looked nothing like a real pitcher. A lot of the work was setting the right constraints to the motion and musculoskeletal model so the simulation stayed within the bounds of human possibility. It was a bit like teaching someone to walk by only showing them the destination and letting them figure out the rest, within reason.
Tech Briefs: Can you explain in simple terms how the whole process works?
Attias: We started with real pitching data with actual MLB pitches recorded using markerless motion capture, where cameras track the movement of every body segment without the pitcher needing to wear any sensors. From that data we built a digital skeleton scaled to the proportions of a professional pitcher, with the relevant muscles, joints, and a baseball attached to the hand. We then handed that skeleton over to a mathematical optimizer that treated the muscles and joint motion patterns as equation and said: throw this ball at X MPH, stay balanced, and use as little effort as possible so that we can protect the elbow. The optimizer ran many of iterations until it finds a solution that satisfies all those conditions. Then we look at what that solution is doing to the UCL.
Tech Briefs: What could this mean for the future of MLB? Can the physics be applied to position players too?
Attias: At the professional level, the hope is that this kind of modelling eventually becomes part of how teams evaluate and develop pitchers. That means not just tracking what a guy is doing, but understanding why certain mechanics are putting him at risk before an injury actually happens. That's a meaningful shift from reactive to preventive.
As for position players, I absolutely believe this work can be extended there, but given the randomness of their positions, the problem becomes more difficult to solve . The throwing motion of an outfielder or a third baseman involves many of the same mechanical demands as pitching, just without the same repetition and consistencies of being on the mound. The framework we developed could in principle be adapted to study those motions too. And beyond throwing, the same modelling approach could be applied to hitting; understanding how swing mechanics affect injury risk or performance is an equally rich problem. The lowest hanging fruit would be to begin with catchers or batters given that they are largely stationary, allowing us to minimize the randomness of play outcomes.
Tech Briefs: Do you have plans for further research? Where do you go from here?
Attias: There are a few natural next steps. One is expanding the dataset since this study was built around a single pitcher, and we'd like to know how well the findings generalize across different body types, arm slots, and mechanics. Another is validation which means collecting synchronized force plate and EMG data to compare against what the simulation predicts. And longer term, the goal would be to make this kind of analysis faster and more accessible, so it could eventually be used in a real-time or near-real-time setting rather than requiring hours of computation.
Tech Briefs: Any advice for researchers aiming to bring their ideas to fruition?
Attias: Find a problem that genuinely matters to people outside your field, and talk to those people early and often. A lot of research stays trapped in academic journals because it was never translated into something actionable. The fact that this work used real MLB data and was done in collaboration with people inside a professional organization forced us to stay grounded and focused. The questions had to be answerable in a way that was useful, not just technically interesting.
Tech Briefs: Is there anything else you'd like to add?
Attias: Just that the injury problem in baseball pitching is genuinely serious and still growing. Tommy John surgery is increasingly common even among teenage pitchers, which suggests something systemic is going wrong well before players reach the professional level. We're still early in understanding the full picture, but having tools that can simulate and optimize human movement without putting additional strain on athletes feels like an important direction. The goal was never just to publish a paper but to contribute something that could eventually help keep pitchers healthy.
