The VAR system at Euro 2024 has been enhanced with AI, enabling it to track even the smallest movements of the ball and players with processing speeds double that of the 2022 World Cup.
Since its official implementation in 2016, Video Assistant Referee technology (VAR) has sparked lively debates. Inconsistencies in how referees utilize the technology and the time taken to make decisions have led to dissatisfaction. The Premier League even held a vote earlier this month to eliminate VAR, although the final support ratio was 19 to 1.
Inside a VAR room. (Image: UEFA).
At Euro 2024, VAR will be implemented in all matches. However, compared to the 2022 World Cup in Qatar, VAR has been significantly upgraded with enhanced accuracy through artificial intelligence.
“The AI’s ability to observe a match is becoming more advanced. It now feels like a sentient being working alongside humans,” said John Eric Goff, a sports physicist at Lynchburg University, Virginia (USA), in an interview with Nature.
According to him, the AI used in VAR for Euro 2024 combines algorithms and machinery capable of processing vast amounts of data that a standard referee might not recognize. This data is retrieved from 10 cameras strategically placed under the stadium’s roof, capable of monitoring 29 points on each player’s body.
Thus, with 22 players on the field, there are over 600 movement points tracked. Every 50 times per second, this data is fed into computers. “Essentially, all these cameras can provide real-time information about a player’s position on the pitch, the ball’s location, and the speed of the ball, players, and their body parts in motion,” Goff explained.
The ball used at Euro 2024 also features a special design, utilizing FussBallLiebe technology from Adidas. The AI sends data from the ball to referees in real time to assist in making quicker decisions, as well as helping VAR referees identify each ball touch.
“This is one of the most advanced and high-tech balls in football history,” commented Give Me Sport.
According to Goff, if one were to cut the ball open and look inside, they would find a small sensor at its center, connected by wires to the outer shell. This sensor records the ball’s position and its movements, transmitting data ten times faster than the cameras used in the stadium.
This data, combined with footage from the cameras, determines the ball’s position relative to the players’ bodies. The internal chip can accurately identify the timing and point of contact whenever the ball is struck or touched by a player’s foot or hand, even with the lightest touch. This is crucial for referees to determine who scored a goal or if the ball touched a hand.
One of the primary applications of VAR is semi-automated offside technology. The AI displays 29 data points collected from each player in three-dimensional space, combining algorithms for bone structure to assess the position of a particular body part on the plane that indicates an offside infraction. If one of the 29 points crosses the limit, the player is deemed offside.
With goalline technology, determining whether the ball has crossed the goal line is also aided by AI, with a 3D image of the ball displayed on the computer.
Finally, AI can make overall decisions quickly. VAR at the 2022 World Cup took an average of 70 seconds to determine offside infractions, but the new technology processes this in under half a minute. Therefore, referees will be able to make decisions faster.
However, Goff notes that the new technology still has some flaws. In reality, the cameras may not have a wide enough coverage to monitor all player actions. The algorithms used to create 3D images of players are also not entirely accurate, as the 29 points do not represent all points on a player’s body. Nevertheless, the errors are considered minimal, under 0.5 cm.
Additionally, another limitation is the point of contact with the ball. The chip in the ball can identify when it has been touched, but it cannot determine whether it was by a hand, head, or another part, requiring the referee to review the situation.
“While AI is helping to improve VAR, I don’t think we will have robot referees in the near future,” Goff remarked. “Tasks such as determining fouls, issuing cards, and other decisions still require human judgment.”
