The author is the founder Sorted outFT-backed media company covering European start-ups
Cameroon striker Vincent Aboubacar thought he was offside when he lobbed the ball wastefully over the head of the Serbian keeper and into the back of the net. Most importantly, so do the assistant coaches. Aboubakar’s celebration began on Monday when the video assistant referee suggested otherwise after helping Cameroon pull off a thrilling comeback at the World Cup in Qatar.
Aboubakar’s goal would have been disallowed before VAR was introduced. Had it been working in Mexico in 1986, VAR would have disallowed Diego Maradona’s infamous “Hand of God” goal in England’s goal when the Argentinian flicked the ball over his shoulder. On such decisions will hang the fate of football, national mythology and the emotions of millions of sports fans around the world.
During the most recent World Cup, football gained interest. But these gigs should be of interest to anyone involved in designing and delivering technology services. Using video assist technology provides an interesting example of how difficult it can be to create optimal product-market fit in a high-pressure environment with millions of adoring followers.
Technology may be more accurate, but at what cost? Traditionalists complain that VAR wastes time, undermines the rights of on-field referees and puts the value of the sport at risk by adding a new and different dimension to controversy.
The use of VAR was first incorporated into official football laws in 2018, and the technology was introduced at the World Cup in Russia later that year. Off-field referees, who review numerous videos, are tasked with identifying “clear and obvious” errors and “missed serious offences” when awarding goals, penalties, issuing red cards and confirming the identity of penalized players.
Evidence suggests that VAR has indeed increased decision-making accuracy. During an international soccer match, a referee makes an average of 137 observable decisions, many of which are monitored in real time. During the World Cup in Russia, FIFA found that in 99.4 percent of 455 infractions investigated by the VAR system, the referees made the final call correctly, while 95.6 percent did not. One consequence of this was that at the previous World Cup in Brazil, the referees found more violations and awarded 13 penalties, compared to 29 penalties (9 as a result of VAR review). However, the use of VAR increased the duration of the match: the average time to review the incident was 82 seconds.
Since then, VAR has been introduced in many football leagues around the world. But critics continue to argue that it adds more confusion than clarity. For example, few would object to the objective of checking whether a football has crossed the goal line or whether a player is offside. But when the on-field referee is asked to reconsider his original decision, there is much more debate about subjective decisions such as a penalty or a red card. A higher standard for decision-making means that fans may be more outraged if they get it wrong.
There are perhaps two lessons to be learned from the use of VARs that can be applied to the implementation of many other decision-making systems. First, technology should never be used just for technology’s sake. It should only be used in specific and limited circumstances where it clearly enhances the process and does not replace it to inform human expert judgment. But efficiency is also important. Technology should not create new ones to solve one set of problems. The system should be constantly improved based on feedback.
For this, it is important that users and fans understand how the system works and trust the methodology. Black box systems are rarely a good idea. In this sense, video officials of cricket matches do a better job of presenting evidence to viewers and explaining how they make their decisions. Being able to interpret decisions is critical to VARs, as they are to AI systems now widely used in fields such as finance, healthcare, and law.
The “minimum intervention, maximum benefit” principle behind VAR is a good one. But experience shows how difficult it is to implement in real life. As math teachers say, show your actions when you solve a problem. The VAR decision-making process itself should be reviewed.