Subscribe to the Sports Tech Feed

Complete your details below to get the latest news and content from the global sports & technology industry


The 2019 Wimbledon Championships have used AI technology to package match highlights of the 600+ games across the 13-day competition. With over 18 courts and hundreds of players, the task of creating ‘highlights’ footage is no simple feat. The six ‘main courts’ alone host an average of 4 matches each per day, with men’s and women’s singles, doubles, mixed doubles and more. This is where IBM’s Watson AI steps in, using automation to sift through hours of footage and curate exciting match highlights.

Wimbledon 2019

The partnership between IBM and Wimbldeon organizers isn’t new. IBM has been working with Wimbledon organizers for 30 years, covering a range of operations from yber-security, to consumer-facing websites and apps. In the last 6 years, IBMs Watson AI has been used to capture match-play and select exciting clips to package as match highlights. What was previously a manual task that required a team of video editors, can now be accomplished by Watson within 2 minutes of a match ending.

Simon Boyden, chief architect at IBM, speaking to VentureBeat said, “We’re taking an activity that traditionally would take a lot of manpower, create a lot of costs for the club, [and] take a lot of time and allowing Wimbledon to get those highlights out as quick as it can”

“What we’ve created is a system that is incredibly well-versed at understanding tennis from an excitement point of view. That’s ‘excitement’ defined in a way that an A.I. can understand it. We humans may appreciate tennis; the A.I. is looking at the same thing in terms of data.”

The AI development team had to find a way to train the system to automatically identify ‘exciting moments’ from matches, using the excitement of the crowd, on-court actions and player reactions as data-points. By analysing these data-points, AI can pinpoint ‘exciting moments’ from matches and automatically compile a highlights reel from each match. For example, crowd noise and player gestures are important elements that help AI to rank every point of a game in order of excitement, and ultimately select clips to create a highlights package.

Match highlights captured by Watson



The production process of Cognitive Highlights

However, the AI technology is not without problems. The issue of bias, as with most AI projects, is an issue that can distort results and affect accuracy. Player’s facial expressions can be misinterpreted and something like hand raises and handshakes can be mistaken for key data points. The system can confuse a player’s arm raise and request for a new ball, as an arm raised in celebration after winning a point. Another example is bias from crowd reactions. Local favourites may draw greater crowd reactions and more prolific players may attract greater crowd appreciation for winning points.

As the Wimbledon case study has shown, using AI systems in sport require extensive training to ensure results are not biased. “The point where we’re 100% convinced that Watson will get it right every single time? That’s probably going to be a few more years,” Boyden says.

Leave a Reply