top of page

How Artificial Intelligence Can Fix VAR in Football

Football is struggling this season with controversial VAR calls that range from the confusing to the outright incorrect. This has seen coaches come out and claim the technology is not up to scratch, fan outrage, and even calls for games to be replayed. So what has gone wrong and how can artificial intelligence help?

Can AI help overcome the recent VAR controversies in football refereeing?

What is VAR?

VAR is an acronym for Video Assistant Referee and it is a suite of software used in football leagues around the world to help the on-pitch referee. VAR is managed on a game-by-game basis by additional support referees. It is used to overturn “clear and obvious errors” by the referee during the game, to review offside and handball decisions with the aide of slow-motion replays, and to notify the referee of any incident he may have missed.

How Does VAR Work?

Although there is still a significant human element to VAR, some parts are based on software particularly when analyzing offside decisions. The essence of an offside call is to determine if one player is more advanced than another at a given moment. In a fast-moving sport like football, this is obviously very difficult. VAR has implemented a process of “drawing lines” on the pitch to determine which player is more advanced.

VAR And The Offside Rule

A player is in an offside position if any part of the head, body or feet is in the opponents’ half (excluding the halfway line) and any part of the head, body or feet is nearer to the opponents’ goal line than both the ball and the second-last opponent. If both of these conditions are met then a player is offside. This becomes an offence when a player in an offside position at the moment the ball is played or touched by a team-mate and then becomes involved in active play.

If this rule sounds very technical, that’s because it is. Many aspects of the offside rule are regularly discussed and argued by football fans but two key components are relevant to VAR. Firstly, how is the exact moment of the ball being passed decided? And secondly, how is “nearer to the goal” measured? Unfortunately despite some significant technological advancements, these two steps are manually completed by hand.

Could AI Help With VAR Offside Decisions?

As you can see in the video, the lines are drawn at the single-pixel level, manually, by a human relying on their own eye. Although VAR gets the vast majority of these calls correct, the consequences of an incorrect decision can be season-defining.

Computer Vision and VAR

Computer Vision is a branch of artificial intelligence that converts camera feeds into analyzable data. The VAR system already relies on well-established Hawk-Eye computer vision technologies to calibrate the camerras prior to the game. However, despite this use of computer vision, offside decisions are still made by hand. Computer vision is already used extensively to analyse player movement in football and it should certainly be considered for use by VAR.

Wearable Devices and Audio AI

Wearables are a relatively new field of AI that could assist referees and VAR. The wearing of highly advanced football boots with tracking devices has recently been approved by the Premier League. The purpose is for more general data gathering and analysis but it is not much of a leap to see it incorporated into the offside check process.

Another area of AI that could assist VAR in making offside decisions is the advancement of audio signal processing algorithms. It is essentially the aural version of computer vision, and software such as SoundSee can ingest audio feeds, convert it to analyzable data, run machine learning algorithms and produce usable insights from the sounds.

Will Artificial Intelligence Fix VAR And Refereeing?

FIFA have already begun testing the use of AI in refereeing but there is a long way to go. As artificial intelligence becomes more mainstream over the coming years it will almost certainly be incorporated into football referring and VAR. However, that does not mean AI will replace human referees. There is a famous quote from Rob Thomas about AI (that you can read here), that can be easily adapted to the application of AI in football refereeing: AI is not going to replace referees, but referees who use AI will replace the referees who do not.

bottom of page