VAR's Silence and the Trap of an Empty Check

On 16 June 2026, at the Kazan Arena, Antoine Griezmann stood over the penalty spot after referee Andrés Cunha left his position in the middle of the pitch, walked to the touchline, bent over a small monitor, then returned and pointed at the 11-metre mark. It was the first penalty awarded through VAR in World Cup history, following Joshua Risdon's challenge on Griezmann inside the box. The Kazan crowd knew the outcome within two minutes.

What the crowd did not know was that in that same match, and across the tournament, the VAR room ran hundreds of other checks and published not a single line about them. No announcement. No record. Nothing to argue about, and nothing to cross-check. I was 27 that year, working as a data editor for a football outlet in England, and I spent 12 matches simply logging every occasion a referee left the field to look at a monitor. By the end of the tournament I understood that what I was logging was not VAR. It was the visible tip of a process whose submerged body had never been published.

Football has no VAR, only blind spots waiting to be exposed. The largest blind spot VAR created is its own silence.

Four categories and a threshold with no unit of measurement

The VAR protocol issued by IFAB limits intervention to four categories: goal or no goal, penalty or no penalty, direct red card, and mistaken identity. The intervention threshold is written as clear and obvious error, or serious missed incident. The governing principle is minimum interference, maximum benefit.

The division of authority is tighter than viewers assume. The VAR team has no power to decide. It has only the power to recommend. The on-field referee is the only person permitted to request a review and the only person who signs the final decision. The process has three layers: the silent check, the recommendation to review, and the on-field review.

The first layer matters most and is also the most invisible. A silent check takes anywhere from a few seconds to a few dozen seconds, emits no signal onto the pitch, and produces no announcement to the stands. It ends one of two ways: the VAR team concludes there is no clear error, or it recommends the referee to the monitor. Only the second path leaves a trace.

According to FIFA's post-tournament report on the 2026 World Cup, the competition recorded more than 450 checks and roughly 20 on-field reviews. That ratio means one check in more than twenty left a public trace. The rest vanished from the public record.

At the 2026 World Cup in Qatar, semi-automated offside technology entered operation, and from the 2026-25 season the Premier League adopted it too. The technology solves part of the problem: it turns a geometric judgement into a publishable calculation. But it only touches the offside category. The other three remain entirely inside the zone of silence. From May 2026, PGMOL began releasing audio of conversations between referees and the VAR team in a programme involving Howard Webb. It was the first time a major league actively published the submerged part. But the programme selects only incidents that already caused controversy, which means it still publishes only checks that produced an outcome. Checks ending in silence appear in no release at all.

The intervention threshold is a judgement, not a number

In the series I wrote in 2026, I logged 47 VAR decisions and compared each incident against the protocol threshold. The result made me abandon my original approach. The same level of contact, the same position inside the box, the same speed of challenge, and two different referees can reach two different conclusions while neither breaches the protocol. The reason is simple: the intervention threshold has no unit of measurement. It is a judgement expressed as an adjective.

Every foul is a question about intent; the data only answers questions about consequence. This is the boundary every data system in football runs into. In the 2026-20 season, when the pandemic halted the Premier League and I had no matches left to report on, I rewatched all 380 games of the campaign and counted Liverpool's fouls under Jurgen Klopp. My average came out at 10.2 per match, distributed mainly in midfield. Mapped by location, the pattern was obvious: these were fouls used to cut off a counter-attack before it took shape, a way of controlling space through interruption.

But when I placed that data table next to refereeing decisions, it was useless. The data told me where the foul happened, in which minute, at what frequency. It did not tell me whether the player intended to play the ball. A referee must answer the question of intent in under a second, with the naked eye, at real speed, with no slow motion. The VAR team has slow motion but must still answer the same question, only with more time.

There is another arena in football where this problem has been solved. The Euro 2026 final took place on 11 July 2026 at Wembley, where Italy beat England 3-2 on penalties. Gianluigi Donnarumma saved from Jadon Sancho and Bukayo Saka, while Marcus Rashford hit the post. In my own notebook, I counted Donnarumma choosing the correct direction on 5 of 7 kicks. The distance from the penalty spot to the goal is 11 metres. A hard-struck penalty covers that distance in roughly 0.3 to 0.4 seconds, shorter than the human neural reaction threshold. Donnarumma was not guessing. He decided before the ball was struck, based on data about the taker's habits.

Goalkeepers hold pre-match data. Referees hold nothing equivalent, because what they need is not a probability model but a judgement about intent. The viewer sees the incident, the referee sees the moment, I see the whole process. And precisely because the process publishes only outcomes, the public is left with a single thing to judge: the outcome.

The blind spot lies elsewhere, not with the referee

The intuitive response to a controversial decision is to demand more disclosure. I used to think so. After PGMOL began releasing audio, I partly changed my mind.

Publishing more audio makes the process more transparent, but it simultaneously exposes that the intervention threshold is a subjective judgement. Viewers hear two referees discussing an incident and both hesitating, and what they take away is not that the process is working, but that the process is arbitrary. Controversy does not shrink. It shifts from arguing about outcomes to arguing about standards, and standards offer nothing to hold onto.

But the deeper blind spot lies elsewhere. The public is making a very basic logical error: reading the absence of an announcement as the absence of an error. A silent check ends with nobody knowing it happened, and the default conclusion becomes that there was no issue. Absence of evidence is read as evidence of absence. Yet an empty check could equally be the result of a check that was never properly performed, of inadequate camera angles, or of a miscommunication inside the VAR room. Nobody can distinguish those three possibilities, because all three leave the same trace: nothing.

This is why I never write about a referee's mistake as a personal verdict. The referee is the only person on the pitch not permitted to be led by emotion, yet the stands are permitted, and the stands write the record. VAR does not fix mistakes, it only changes who carries the responsibility. Before 2026, responsibility sat with one person in the middle of the pitch. After 2026, responsibility sits in a room nobody sees, operating on a threshold nobody can read, leaving traces in twenty cases out of more than four hundred checks.

The distance between an individual error and a process error is enormous. A referee misjudging an incident is the business of one person. A system that makes every correct decision unverifiable is the business of an entire football ecosystem.

What is needed is a record of what did not happen

My proposal is not more audio. It is to publish the checks that end in silence as well, in the same format as a match delegate report: which minute, which category, who initiated the check, at which layer the conclusion was reached, total duration. No need to publish images, no need to publish audio, only a raw data table.

Its value does not lie in re-litigating every incident. It lies in turning an empty space into a measurable fact. Once the volume of silent checks becomes public, there is finally a basis to ask why the same type of contact is treated differently, and that question can lead to standardising the intervention threshold with criteria more concrete than adjectives.

The laws never stand outside the match; they are the second match running in parallel. That second match is currently recorded in a ledger that notes only incidents producing an outcome, and reading from that ledger, nobody can say whether the system is working well or failing. A check that leaves no trace is a check that cannot be questioned, and cannot be trusted either.