Wimbledon's AI line judges: 380 of 1,535 player challenges were right

Two incidents defined the tournament's technology debate. The Guardian reported on 15 July 2025 that Sonay Kartal's ball was not called long against Anastasia Pavlyuchenkova, and that Taylor Fritz had a forehand called out despite landing four feet inside the baseline.

Neither was a system failure in the ordinary sense. A ballboy was still on court during the Fritz serve, and an official had switched the system off before the Kartal call. Sean Ingle, writing in The Guardian, noted that the system in use was a souped-up version of Hawk-Eye, which Wimbledon has used since 2007. The two errors were real. The conclusion drawn from them was not.

What the challenge data shows

The most useful number in the debate comes from IBM. Ingle asked the company how often players got it right when they challenged a line call at Wimbledon in 2024. Across the men's and women's singles, players made 1,535 challenges. Just 380 were overturned. That is less than 25%. Put plainly: when a player thought the ball was out and challenged, they were wrong three times out of four. Researchers had earlier estimated that line judges get around 8% of close calls wrong, a figure Ingle cites as long-standing. Players perform several times worse than the humans they were disputing.

"No system is 100% perfect, but they are demonstrably more accurate than relying purely on human decision-making," says Matt Drew, who founded the integrity department at StatsPerform. "Sports believe that technology helps them get more decisions right and protects officials and players from abuse."

The abuse point is not rhetorical. At the last Rugby World Cup, referee Wayne Barnes described receiving threats of sexual violence against his wife and threats of violence against his children. Ingle's argument is that sporting bodies adopt technology partly because it shields officials from that pressure. The alternative, a disputed call with no review, now produces social media pile-ons rather than shrugged shoulders.

Crowd noise and unconscious bias

Human judgment has a documented problem that machines do not share. One study cited by The Guardian asked 40 qualified football referees to judge 47 incidents from a Liverpool versus Leicester match. Half watched with crowd noise; the control group watched in silence. Those exposed to the noise awarded significantly fewer fouls against Liverpool, 15.5% fewer. A separate Norwegian study found that successful teams were more likely to receive favourable penalty decisions. Psychologists call the effect conformity.

Ingle frames the question that critics of sports technology rarely answer directly: is the technology better and more accurate than what it replaced, and is there room to improve it? By that standard, Hawk-Eye has a reasonable record. It is more accurate now than when it was introduced in 2007, and the trend has been one direction.

The same logic is spreading. From September, according to The Guardian, the NFL will replace its chain gang of officials who walk onto the pitch to mark first downs with Hawk-Eye technology. Ingle writes that he will miss them. Then he adds that having someone guess where the ball should be placed feels closer to the 18th century than the 21st.

What comes next

At the International Olympic Committee's artificial intelligence conference last year, a diver was shown in real time with a screen telling judges the height of the jump, the number of rotations in the air, and how close his legs were to his torso during the spin. Each element was split into sequences and analysed in less than a tenth of a second. The stated aim was to give every judge a better basis for a fairer score.

None of this settles the argument about tradition, and Ingle does not pretend it does. He concedes that more traditions will slip away as the machines spread. His case rests on a narrower claim: that the outrage at Wimbledon rested on two incidents, one caused by a switched-off system and one by a ballboy on court, neither of which says much about whether computers can replace human judgment. The challenge data says they already do.