The science that shows Hawk-Eye isn't as accurate as it seems

Telegraph

12-07-2025

At the Madrid Masters in April, top seed Alexander Zverev was given a warning for unsportsmanlike behaviour.
His offence was to take out his phone and snap a picture of a clear ball mark in the clay which was more than an inch outside the line. The electronic system Hawk-Eye had called it in, and he lost the point.
'The ball is not just a little bit – a millimetre – in or out, it was like four, five centimetres,' the German said in a post-match press conference. World No 21 Donna Vekic branded the line call as 'insanity'.
It is one of several recent cases in which players have questioned the accuracy of Hawk-Eye, which this year replaced line judges at Wimbledon after 147 years.
World No 1 Aryna Sabalenka also took pictures of a disputed ball mark during her quarter-final victory over Elise Mertens at the Stuttgart Open in April.
At this year's Wimbledon, Emma Raducanu said it was 'disappointing' that the 'calls can be so wrong' while Jack Draper argued after his second-round defeat: 'I don't think it's 100 per cent accurate.'
Scientific experts increasingly believe the players have a point.
In 2009, Professor Harry Collins of the University of Cardiff wrote a paper highlighting the problems with Hawk-Eye, and he believes little has changed.
'Hawk-Eye is useful for big errors but for very small errors everybody knows that the thing just cannot be exact,' he said.
How Hawk-Eye really works
Tennis viewers may not realise that Hawk-Eye is not a true video-replay of the shot, but a virtual reality simulation based on the speed and trajectory of the ball.
It relies on a network of 10 high-speed cameras which capture the ball from different angles at a fast frame rate, feeding the movement into a computer algorithm which then triangulates the position, creating a reconstructed image of the flight path.
It means that the final line call is not where the ball landed, but where it 'should' have landed based on how it was moving.
The distinction is important. Even Hawk-Eye Innovations, which developed the technology, admits that the system has a mean error rating of 2.2mm, which may not sound much, but can have big implications for players hitting shots with pinpoint accuracy.
Critics also point out that if 2.2mm is the mean error rate then, on occasions, it must be far less accurate than that.
The company has never released information on just how inaccurate Hawk-Eye can be, but if standard calculations are used to estimate the expected distribution of errors, it suggests accuracy could be out by more than 5mm in one in 20 occasions, and more than 7mm in one in 100 shots.
It is also unknown if the accuracy rate falls for certain types of shots – such as killer forehands or super-fast serves, where the camera frame rate struggles to keep up with balls flying at 150mph.
Likewise, it is unclear how well the system accounts for spin which can cause a ball to suddenly drop like a stone in mid-air or swerve unexpectedly to the left or right.
The system is also known to be impacted by both sunlight, shadow and floodlights as well as wind which can cause the cameras to wobble. The lines on the tennis court can sometimes be obscured by the players themselves.
In the Zverev example in Madrid, it was speculated that the cameras had misread a clay grooming line to the left of the court, believing the mark to be the real sideline.
Professor Collins added: 'A painted line on a tennis court cannot be a sharp edge, and when a tennis ball bounces it distorts, and you have a hairy, distorted tennis ball next to the ragged edge of a line, so it can't be exact.
'The way it ought to work is when the ball is close to the line, within a couple of millimetres, then there should be a let or the umpire's decisions should count.
'It's dangerous to rely on this technology alone because it is misleading the public into thinking that virtual reality is real reality.'
Electronic line calling has been used instead of line judges at the US Open and Australian Open since 2021 and the International Tennis Federation told The Telegraph that it has been 'rigorously assessed' against a wide range of criteria including accuracy, reliability, suitability and practicality.
But there have been numerous incidents when players and viewers claim the technology has gone wrong.
A glitch in the system
During a first round match of last year's Cincinnati Open, American Brandon Nakashima hit out a forehand which should have seen him facing a break point against compatriot Taylor Fritz.
But the electronic line call stayed silent until later on in the point when it finally corrected the error, disrupting the game and forcing a replay.
At the 2024 Miami Open, during the match between Australian Daria Kasatkina and Romania's Sorana Cirstea, a ball that was clearly outside the line was called in, with Hawk-Eye footage showing it moving in the wrong direction.
Defending champion Carlos Alcaraz was overheard complaining about some of the calls during his third round win at Wimbledon this year.
Last year, a paper by Nigerian researchers warned that there can be gaps in data captured leading to 'less accurate predictions or incomplete information about the ball's path' and said the system could be stymied by 'unorthodox deliveries' or 'poor calibration of equipment'.
Fans of electronic line calling claim that even with the errors Hawk-Eye is still more accurate than line judges, and argue that contested shots often look different on TV because of the 60 frames per second rate of broadcast cameras, which causes motion blur.
In comparison, testing of the technology is done using cameras of 1000 frames per second.
A question of trust
All England Club chair Debbie Jevans argues that players have been asking for electronic line calling for years because they thought it was more accurate, while Tim Henman has branded this year's Wimbledon complaints as 'garbage' claiming Hawk-Eye was '100 per cent accurate'.
However, critics say that if tournaments continue to use the technology, it should be made clear that the Hawk-Eye replays are not true footage and that they can be prone to errors.
Professor Robert Evans, of Cardiff University, who co-authored the 2008 Hawk-Eye paper with Professor Collins, said: 'It was more the way it was presented as completely accurate, with no error and no mistakes, that was the focus of our critique.
'Viewers should be told they are not watching a replay of the real event.'
The Telegraph contacted Hawk-Eye Innovations on three occasions to learn more about testing, accuracy and error distribution but did not receive a reply.