The Data Gap in the VAR Room: 523 Matches, 47 Seconds and a Lesson for V.League

On 16 June 2026 in Kazan, referee Andrés Cunha stepped away from the pitch-side monitor and pointed to the penalty spot. It was the first penalty ever awarded through VAR in World Cup history, in France versus Australia. Antoine Griezmann converted it in the 58th minute. In a Valencia radio studio, I told more than four million listeners that the decision was wrong, that the ball had struck Josh Risdon's armpit area and therefore could not be handball. I was relying on a law I had learned in 2026. Three minutes later, the colleague beside me handed over the 2026/17 Laws of the Game, in which the boundary of the hand had been redefined: measured from the bottom of the armpit. The editorial team had to issue a correction during the interval. After thirty years in the trade, it was the first time I had been contradicted live on air.

What I learned did not come from that incident. A wrong VAR decision rarely originates in poor eyesight; it usually originates in a data file that was never updated. Eight years later, as football entered the 2026 World Cup cycle with 48 teams and 104 matches, I recognised the Kazan story repeating at a larger scale: VAR rooms are running on uneven databases, and where the data is empty, people are forced to guess.

Three law revisions, three times I had to relearn

The handball law has changed at least three times in a decade, and each change left behind a layer of referees and commentators who were stranded. The 2026/17 Laws, issued by the International Football Association Board (IFAB), were the first text to define the boundary between shoulder and arm by a specific anatomical marker: the bottom of the armpit. Before that, the concept of the hand was understood through professional habit rather than through text. The 2026/20 edition tightened the rules on handball leading directly to a goal while restating the armpit marker as the dividing standard.

I built a personal reference table, classifying every regulation into three colours. Green means it is in force and I hold the original text. Amber means it has been approved but is not yet applied in the competition I am analysing. Red means I once read it somewhere but cannot find the source document. My rule is simple: every public statement may only rest on a green cell.

More recently, IFAB changed the framework for match management again. At its meeting on 1 March 2026, the board approved an eight-second limit for goalkeepers holding the ball, punishable by a corner kick, and standardised the principle that only the captain may approach the referee to question a contested decision. Both changes share one logic: reduce dead time and reduce the number of people intervening in a decision already made.

That is why I never answer a question about the laws immediately. Law does not live in memory; it lives in data. The memory of a man thirty years into the job is a beautiful library with outdated spines.

VAR reached V.League late, but pressure arrived early

In July 2026, VAR appeared for the first time in V.League 1 during the closing rounds of the 2026 season, under the rollout plan of the Vietnam Professional Football Joint Stock Company (VPF), after the Vietnam Football Federation (VFF) completed the conditions approved by FIFA. Since then the number of VAR-equipped matches has grown season by season, and by the 2026/26 campaign the technology has become a familiar part of the league.

A league newly deploying VAR has one characteristic I recognise clearly when comparing it against my own data: technical infrastructure is usually funded faster than intellectual infrastructure. Cameras can be installed in weeks. A system of laws, precedents and application criteria takes several seasons to accumulate. Inside that gap, referees still have to blow the whistle, and the stands still have to react.

Based on my experience observing matches and logging decisions, the pressure in a newly VAR-equipped league does not come from referees being more wrong. It comes from viewers having already seen the slow-motion image, already having faith in the technology, without simultaneously being given a framework of criteria for interpreting that image. A phase of play at 25 per cent frame rate looks entirely different from the same phase at real speed. And the criteria for reading the two versions are not the same.

523 matches, 47 seconds, and a file nobody opens

After the Kazan error, I began a personal project in August 2026: logging every VAR-involved decision in La Liga and the Champions League, with error codes, match timings, the distance between the referee and the monitor, and ball speed at the moment of contact. When the COVID-19 pandemic halted European football in March 2026, I had 523 matches in my file.

The results forced me to rewrite my entire working method. Among offside decisions that were overturned, 74 per cent had a review time beyond 30 seconds, with the average landing at roughly 47 seconds. I did not publish that figure as evidence about refereeing competence. I published it as an indicator of process design: the longer a review runs, the higher the chance that a factor beyond the image enters the decision.

The 48-page report I published on my personal blog proposed a 30-second cap on each offside review, with the requirement that if the time expires without conclusive data, the on-field decision must stand. The Valencia football federation then invited me to advise their reform process. I accepted on one condition: I would be allowed to publish the entire dataset so anyone could verify it.

One match is only a story. Five hundred matches are law. A shocking decision is not reckless if it is built on five hundred foundations. Conversely, a decision that looks reasonable can still be a systemic error if it matches no precedent in the file.

What haunted me most across those 523 matches was a detail unrelated to football. I noticed that whenever my dataset contained an empty cell, I tended to fill it with intuition. And intuition, in my trade, always wears the coat of certainty.

A data gap is more dangerous than bad data

There is a paradox in how we assess the quality of refereeing decisions. Bad data gets detected, argued over, corrected. An empty dataset triggers no argument, because nobody knows it is empty.

In an analytical workflow, a blank field usually raises no warning. It creates a silence. And silence, in a VAR room as in a newsroom, is always filled with the nearest available material: memory, a feeling about the match, or the authority of whoever spoke last.

I lived through exactly that mechanism. In 2026, the cell in my head about the hand boundary was outdated, but it was not empty, so I had no reason to look it up. Had it been empty, I would have opened the law book. Certainty was what stopped me.

Applying this to leagues newly adopting VAR, I would argue the biggest risk is not referees who do not know the laws. The biggest risk is referees and pundits alike believing they know the laws while the version they are using expired one or two seasons earlier. There is a way to see that risk in the data itself: if a league publishes no record of the criteria applied to each error category, every VAR argument in that league will end in emotion rather than standard. Fans do not lack images. They lack a yardstick.

The four-step process I apply to every analysis

After the 48-page report, I standardised my workflow into four fixed steps, and I complete all four before writing a single sentence about a contested decision. The first is to establish the subject: which team, which player, which referee, which competition, which season. It sounds redundant, yet most online arguments fail here, because the two sides are discussing two different incidents in two different matches.

The second is to establish which version of the law was in force on the date of the match, with the publication date and the start of application. This is the step I once skipped, and the cost was thirty years of credibility questioned within three minutes. The third is to find precedent. An incident with no precedent is one I am not yet qualified to rule on. I do not say it was right or wrong. I say my dataset does not yet cover it. The fourth is to state the reliability level of the source: primary documented source, secondary source with citation, or unverified information. The third level may never appear in a conclusion, only in the open-questions section.

Referees do not need protecting. They need to be understood through correct data. I once wrote a piece defending a referee using faulty statistics. That was the moment I betrayed my own principle, and I had to publish a public apology with a corrected data table.

One whistle, two levels of pressure

There is a hypothesis I have pursued for years: referees treat big clubs and small clubs differently. It is usually dismissed as a conspiracy theory of the losers. But placed inside data, it stops being a conspiracy and becomes an environmental variable. Stadium pressure is real and partly measurable. A fuller ground produces more noise. More noise produces longer reaction latency. Longer latency in a phase of play can determine whether a referee blows in the first second, and whether VAR is called in.

Alongside the stands sits the media. A decision against a big club is dissected for forty-eight hours. A decision against a small club gets fifteen minutes. Referees read the papers. Referees live inside the same information stream as the rest of us. My point is not that there is a secret directive. My point is that there is a different environment, and environment always affects people. If a league cannot measure that environment, it will never improve the consistency of its decisions.

The measurement is concrete. Count VAR interventions in matches involving big clubs. Count them in matches that do not. Count average review time per group. If average review time in the big-club group is significantly longer, that is a signal requiring explanation, and it cannot be explained by emotion. For a league like V.League, where the media and attendance gap between a few leading clubs and the rest is enormous, this would be the first dataset I would build.

Injury, return, and the cost of demanding too soon

Another area where data is systematically misread is injury and return to play. For decades, a player returning from a long injury faced a test written in no document: the first match had to prove he was still intact. That demand does not come from the coaching staff or the team doctor. It comes from the stands and from public opinion.

Sports medicine data shows reinjury risk concentrates heavily in the early phase after return, when a player has not yet regained decision-making speed in tight spaces. A player structurally recovered is not necessarily match-cognitively recovered. Those two require different timelines. Yet our assessment usually has only one benchmark: the minute he scored, the minute he was substituted. From that single data point we draw conclusions about a rehabilitation lasting months.

From a regulatory standpoint, this is where substitution rules deserve another look. Allowing more substitutions, and assessing players returning from head injuries through a dedicated protocol, are reforms grounded in data rather than emotional concessions. A good law must reflect real risk, not the expectations of the stands.

What I want to see next season

The 2026 World Cup cycle, with 48 teams and 104 matches, will generate a volume of decisions unprecedented in football history. Semi-automated offside technology has been in place since the 2026 World Cup, and referees announcing decisions over the stadium PA were trialled at club-level competitions in 2026. Those tools only have value if accompanied by a standardised, public dataset.

For V.League, what I want is not more cameras. What I want is a public dataset, updated round by round, recording which error category each VAR intervention belongs to, how many seconds it took, and which clause of the version in force it relied on. When that file exists, arguments shift from competing feelings to competing standards.

At 67, I do not need to remember everything. I need to know how to find what is right. I still attach the data file to every piece I write, so readers can check rather than trust me. That is the only way I know to keep a trap-laden profession honest.

My 2026 error was not forgetting the law. It was daring to be certain while my dataset was empty. If a league learns that before it learns how to add more cameras, it will save itself years of argument.