IBM’s tennis AI measures player performance and selects highlights

Tennis and AI are like peas in a pod, if you ask IBM. The folks at the Armonk, New York-based company today detailed Coach Advisor, a cloud-hosted solution developed in collaboration with the U.S. Tennis Association (USTA) that taps AI to deliver insights to professional coaches. It introduces a metric that attempts to quantify a player’s physical exertion and endurance and a model that makes correlations to match performance.

“Over the nearly 30 years we have partnered with the U.S. Open, we have seen first-hand the enormous amount of tennis data generated over the course of a two-week tournament,” said IBM sports and entertainment partnerships vice president Noah Syken. “With Coach Advisor, we’re proud to now bring these advanced technologies to USTA coaches directly to tap into new data sources so they can drive a deeper analysis of the game and support the growth and development of U.S. tennis players.”

Coach Advisor uses what IBM calls the Energy System, a measure of physiological load and mechanical intensity meant to evaluate a tennis player’s exertion over time. It tracks the overall work performed in a match, taking into account the person’s height, weight, average speed, and distance traveled. And it keeps tabs on the cumulative weighted acceleration and deceleration throughout, enabling it to gauge the physical demand on a player’s body.

Within Coach Advisor, Energy System metrics are presented alongside traditional statistic and match video for coaches to review. Algorithms trained to detect the sound of ball strikes, umpire calls, crowd noise, and the sound of sneakers on the court determine the clip’s start and end points, allowing coaches to quickly skip to critical moments post-match.

IBM says that Coach Advisor is currently being piloted with a select group of players and their coaches and that it plans to roll the suite out over time to “the broader population” of U.S. players and coaches.

“The USTA is dedicated to developing the next generation of world-class American tennis players, and Coach Advisor offers our coaches access to all-new data to drive performance,” wrote USTA general manager Martin Blackman. “We are excited to continue our work with IBM to pair analytics with our coaching expertise and believe IBM Coach Advisor will give us a competitive advantage to help change the paradigm of tennis coaching and development in the U.S.”

In related news, IBM today revealed that it will work with the USTA to process hundreds of hours of live video from tournaments with AI. The company says its improved sound analysis models, delivered via the IBM Cloud and IBM Watson Open Scale, now allow for tighter cropping of highlight clips by recognizing when the ball has been struck. Additionally, it says they’re able to recognize players’ noise and excitement levels, minimizing bias when searching for highlights from particularly popular or animated players.

IBM deployed the same solution at the Wimbledon 2019 tennis championships earlier this year.