The Kazan Lesson of 2026: VAR Needs Correct Data, Not Faith

On June 16, 2026, at Kazan Arena, in the 55th minute of France against Australia, referee Andrés Cunha stopped play and walked to the monitor at the side of the pitch. Josh Risdon had just lunged into Antoine Griezmann inside the penalty area. Cunha did not give the penalty at first; after consulting VAR, he reversed his decision and awarded it. Griezmann opened the scoring in the 58th minute, Mile Jedinak equalised from the spot in the 62nd, and Paul Pogba sealed a 2-1 win in the 81st, according to the FIFA match report.

More than four thousand kilometres from Kazan, I was sitting in a small studio in Valencia, invited to analyse the laws for a local radio station. I told more than four million listeners that the ball had struck the armpit area and therefore could not be handball. The colleague seated to my right leaned over and said quietly that since 2026 the armpit had fallen inside the definition of the arm. I was wrong live on air. The editorial team had to read a correction in the next bulletin, and that was the first time in thirty years of work that I was publicly contradicted.

I once got one sentence wrong and lost an entire reputation. If only I had known this back then: the law does not live in memory, it lives in data.

That error belonged less to one individual than to the way an entire industry remembers the laws. The definition of handball has been rewritten several times across three decades. The version I learned in 2026 required referees to determine a player's deliberate intent. In 2026, the concept of an unnatural arm position was added to reduce reliance on subjective judgement. In 2026, IFAB extended the boundary of the arm up to the armpit line, and in 2026 the text was rewritten to be shorter and clearer. Each revision carried a different effective date, and each time, a group of commentators somewhere kept citing the old version.

The 2026 World Cup was the first tournament to use VAR across every match, meaning that for the first time hundreds of millions of viewers watched a decision being overturned live. They were shown the pictures but not the legal text. That gap was filled with emotion: the team was robbed, the referee was biased, the tournament was fixed. I was in the studio at exactly that moment, and I filled that gap with belief instead of data.

My law index records the publication date and the effective date of every revision. If a reader finds a different version, I will correct it and publish the correction, because an error fixed through proper process is worth more than a correct claim without a source.

After Kazan, I built a yearly index recording the publication date, the effective date, and the version in force for each competition. Since then, every time I analyse an incident, I add a note underneath stating which revision was in force from which date, instead of saying what I remember. That habit sounds small, but it changed my entire way of working, and it led me to a project I never imagined doing at sixty.

In August 2026, I began logging every VAR decision in La Liga and the Champions League. Each incident was stored with an error code, the minute it occurred, the distance from the referee's position to the monitor, the number of players inside the penalty area, and the ball speed at the moment of contact. By March 2026, when the pandemic halted football, the database held 523 matches. Based on my experience watching matches in those two competitions, I no longer trust my own memory when judging an incident; I trust only a marked data cell.

Three findings stood out. Among offside reviews, 74 percent of decisions drew visible protest from the affected side. On timing, the average interval from the start of a review to the restart was 47 seconds, with some cases running beyond two minutes. On crowd context, matches in larger stadiums showed longer review times; I hold that finding at medium confidence because the sample is not yet dense enough to rule out other variables.

I compiled everything into a 48-page report, published it on my personal blog, and attached the data file so anyone could check every row. The central proposal was simple: a review should not exceed 30 seconds. A hard limit forces competitions to invest in wide-angle cameras, in pre-match error classification, and in training referees to read incidents faster. The Valencia football federation then invited me to advise on procedural reform, and that was the first time data collected by a man watching television entered a formal technical meeting.

What VAR still lacks is not technology but a shared data standard: unified error codes, clear time thresholds, and an obligation to publish.

One match is only a story. Five hundred matches are the law. When a controversial decision appears in Sunday night's bulletin, the viewer sees one camera angle, one slow-motion replay, one commentator raising his voice. Place that same incident among hundreds of similar ones and new variables emerge: player density, the referee's viewing angle, the distance from the point of contact to the touchline, and the time pressure of the match itself.

What is easily missed is that every VAR argument shares one feature: the audience demands instant certainty while refusing to pay for it. A review costs forty extra seconds, and ten minutes later thousands of posts declare that the game has been broken. Those two demands cannot be satisfied by the same procedure unless the procedure is designed with data rather than with reactions.

In my dataset, matches with fuller stadiums and concentrated media attention tended to show longer review times in incidents near the away team's penalty area. I do not conclude that this is deliberate bias, and I refuse any such conclusion without an adequate control sample. I simply record that crowd pressure is a measurable variable, and if it is measurable it must be managed like every other variable in refereeing.

There is another lesson I only learned after long enough in the job. When data is empty, people tend to write rather than stop. I once received an analysis containing not a single information point: no team, no player, no competition, no date, only blank fields presented as if they were conclusions. The correct response is to declare insufficient data and return the file to the collection stage. The easy response is to fill the gap with a plausible story. Football commentary makes this mistake every day; the difference is that the consequence does not appear immediately on the scoreboard.

Referees do not need to be protected. They need to be understood through correct numbers. I once wrote a piece defending referees using faulty statistics. That was the moment I betrayed my own principle.

My proposal for next season's competitions fits in three lines: cap review time, standardise error codes, and publish per-round data. None of them requires new technology. What is required is governing will, and a mechanism clear enough that referees know they will be judged by numbers rather than by the noise in the stands.

When I count every incident, I understand that the law judges no one. It simply waits to be applied correctly. At 67, I do not need to remember everything. I need to know how to find what is right. If this season adds one more competition to the dataset, I will start again from the first empty cell, and I will not write a line until that cell is filled.