The Referee's Eye and the Limits of Evidence: How VAR Changed the Way Football Defines a Mistake

The Referee's Eye and the Limits of Evidence: How VAR Changed the Way Football Defines a Mistake

Part 1 — The moment

On 1 December 2026, at Khalifa International Stadium in Doha, in the 51st minute of Japan against Spain, Kaoru Mitoma chased a ball that appeared to have crossed the byline, cut it back, and Ao Tanaka swept it into the net. The referee awarded the goal. Then the VAR buzzer sounded in his earpiece.

For the next two minutes the stadium went quiet in a very particular way — not the silence of waiting, but the silence of a crowd trying to persuade itself that its own eyes had been right. The reconstructed image was released. The ball still touched the line. Japan won 2-1, Spain advanced, Germany went out.

I have watched that clip more than twenty times. What struck me was not the decision. It was where the argument relocated to. Nobody said "the referee did not see it" anymore. People said "the law is not written clearly enough." That is a far bigger turning point than any goal.

VAR does not fix a match's mistakes — it exposes how we define a mistake. I first wrote that line in 2026, in a long piece about Liverpool against Atlético Madrid at an empty Anfield. I have not found a better one since.

Part 2 — A decade of evidence

On 27 June 2026, in Bloemfontein, Frank Lampard struck the bar and the ball bounced clearly over the goal line against Germany. Uruguayan referee Jorge Larrionda did not award the goal. Germany led 2-1 at the time and won 4-1. The political fallout was enormous: in July 2026 the International Football Association Board approved goal-line technology, and it debuted at a World Cup in Brazil in 2026. For the first time in history, part of the law of football was handed to a machine.

Six years later, VAR was written into the Laws of the Game. At the 2026 World Cup in Russia, a finals tournament used it for the first time. France against Australia on 16 June 2026 marked the first VAR-awarded penalty at a World Cup; Antoine Griezmann converted it. The final on 15 July 2026 between France and Croatia was the first World Cup final with a direct VAR intervention: Ivan Perišić handled in the box and France were given a penalty.

Four years later, in Qatar, semi-automated offside arrived. FIFA published the specification: twelve dedicated cameras under the stadium roof, tracking twenty-nine data points per player at fifty samples per second, with an inertial measurement unit inside the ball running at five hundred hertz to establish the exact kick point. On 22 November 2026, in Argentina against Saudi Arabia, three Argentina goals were disallowed by that system.

Four dates: 2026, 2026, 2026, 2026. Read as a straight line, it is the story of technology catching up with the eye. Read more closely, it is the story of a contradiction: technology grows more precise, while the language of the law stands almost still.

Part 3 — What actually happened in those two minutes

Law 9 states that the ball is out of play when it has wholly crossed the boundary line, whether on the ground or in the air. The operative word is "wholly." For a ball twenty-two centimetres in diameter, that means the entire projection must lie outside the vertical plane of the line. In theory it is clean. In practice it depends on where you stand to look.

The technical problem lies elsewhere: the camera is not on the plane of the line. When a camera sits high and offset, what it records is a projection. Reconstructing whether the ball crossed requires building a three-dimensional model from several angles — and that introduces a chain of assumptions: ball diameter, surface curvature, frame jitter, signal latency. Every assumption is a place to argue.

In Vietnam I read a great many comments after that match. Most were very confident. Most were built on a compressed, rescaled two-dimensional still that had been shared online. This is what my old trade taught me: a compressed frame is not evidence; it is a photograph of evidence. People argue with what they have, not with what they need.

The same applies to offside. For decades offside was a judgement about the relative distance between two moving bodies, made in a hundredth of a second, by a person standing fifteen to twenty metres away. I have stood in that position. You do not watch the ball and the players at once; you hear the ball leave the foot, then sweep your eyes to the line, then decide. Nothing called "seeing" an offside happens at real speed.

Semi-automated offside solves exactly that part. It answers the geometric question: who is nearer the line, and at which moment. The answer can be accurate to a few millimetres. In Argentina against Saudi Arabia, people were startled that a goal was cancelled by a shoulder, a knee, a heel the naked eye could never catch. The system was not wrong. It answered the question the law asked.

The fault lies elsewhere: the law's language has fallen behind the precision technology can deliver.

The offside law exists to stop a forward camping in front of goal. It is an anti-advantage law. But a toe ahead of the line confers no advantage — it is simply a consequence of the human body not being a geometric solid. When technology pushes resolution to the millimetre, the law starts punishing things outside its original intent.

There is something an offside trap can never catch: the player's intent. And by pushing accuracy to that level, technology quietly turns intent — the only thing the law truly meant to protect — into a forgotten variable.

There is another layer few mention. A passage of play does not have a single moment. A long pass can involve two consecutive touches: one to control, one to release. The sensor array records both. So which touch is "the moment the ball left the foot"? On a short outside-of-the-boot pass, the gap between the two touches can be two hundredths of a second — and in two hundredths of a second a sprinting forward covers twenty centimetres.

In the 2026-2026 English league season I spent an entire night reconstructing the frame of Tottenham's goal against Manchester City in the Champions League quarter-final on 17 April 2026. Raheem Sterling's stoppage-time goal was ruled out by VAR for an offside earlier in the move. The Etihad erupted and then froze. I remember sitting in front of the screen at three in the morning, rewinding, and realising I was trying to answer a question no tool could answer: which frame was the one in which Christian Eriksen released the ball?

That is why I do not believe technology is making refereeing easier. It is making it different. It removes discretion from moments where nobody could exercise discretion, and replaces it with a new kind of tension: the tension of someone operating a procedure.

Part 4 — The eye outside the frame

Over more than a decade with the whistle, I learned that the best referees are not the ones who get the most calls right. They are the ones who stand in the right place most often. It is a strangely undervalued skill, precisely because it is invisible.

The referee's diagonal is a line from one corner to the opposite one, but in modern football that diagonal is rarely walked as designed. Players move in blocks, in clusters, and through zones no statistical table displays. The person with the whistle has to read where the match will break before it breaks.

I remember a game in which I ran against the direction of a counter-attack purely to keep a sightline into the challenge that would happen three seconds later. I arrived in the right place and nothing happened. The crowd saw nothing. The data recorded nothing. But had I been in the wrong place, and had that moment arrived, I would have lost the right to decide — and the whole match would have turned on the silence of a whistle that never blew.

Pedri does not run after the ball, Pedri runs towards where the ball will arrive — and that is the entire difference. At Euro 2026, as Spain reached the semi-finals, I spent most of my time not watching the ball but the gaps just before it was received. That is where an eighteen-year-old becomes the centre of a system, and where matches are actually settled.

Referees are the same. Most correct decisions are prepared before the incident occurs, by choosing a position. And this is where VAR cannot substitute: a camera only films what it was placed to film. If nobody stands in the right place, no camera angle compensates.

I once wrote about Liverpool against Atlético Madrid on 11 March 2026 at Anfield. It was one of the last nights in that stretch when European stadiums still admitted crowds, before football stopped for the pandemic. With the stands empty, I could hear the ball strike the boot — something ten years of refereeing had never let me hear. That sound tells you the force of the touch, and sometimes the intent behind it. No camera records it.

Liverpool led 2-0 through Georginio Wijnaldum on 43 minutes and Roberto Firmino on 94, then Marcos Llorente scored twice inside eight minutes of extra time before Álvaro Morata sealed a 3-2 Atlético win on 120 plus 1. The aggregate finished 4-2 to the visitors. But what I remember most is not the scoreline; it is a small detail: on one counter-attack I heard the Liverpool defender's studs tighten on the grass, and I knew immediately where the next tackle would come from.

No system supplies that. And it is not romantic. It is simply another data channel, cheaper and less accurate, but sometimes the only one present.

Part 5 — The paradox of precision

People often say VAR made football fairer. I am not sure that is true in the sense they mean. VAR does not distribute more fairness; it distributes more accuracy. Those are different things, and sometimes they oppose each other.

Consider an example. In a match with twelve offside cameras and a sensor inside the ball, a forward loses a goal because his heel is two centimetres ahead. In another match, on a pitch without VAR, the same act goes undetected. Both matches comply with the law, yet they are not fair to each other. Technology has created something football never had: inequality of resolution.

In the V-League this is starker. From the 2026 season, VAR was introduced into Vietnam's top division on a step-by-step roadmap, starting with a limited number of fixtures and expanding as infrastructure and referee capacity allowed. I have followed that process closely. The biggest problem is not the hardware. It is that when a league has VAR at only some of its matches, those matches with and without VAR become two different leagues, playing one law but living in two different evidentiary regimes.

For players, that creates a new kind of uncertainty. A defender in a VAR league can calculate the offside line to the centimetre. A defender in a non-VAR league must play by probability. The skills differ. The risks differ. More importantly: the two players do not learn the law the same way.

Conversely, I do not agree with the popular reaction that demands scrapping VAR. Removing it does not return football to an innocent age. That age never existed. It only returns football to an era in which a mistake could be erased from collective memory if it was never replayed.

What needs fixing is not the tool. It is the relationship between the tool and the language of the law. Today the offside law is written in the language of the twentieth century, while the technology measures in the language of the twenty-first. The gap between them is where every argument is born.

Part 6 — What never appears on screen

There is a part of the match that appears in no VAR room: game management.

A referee does not stop at making decisions. They also regulate tempo, allocate attention, and stop twenty-two players turning a match into a brawl. None of that is recorded in a frame. No system scores it. And no VAR overturns it.

I once watched a match where the two sides clashed three times inside ten minutes. The referee called the two captains over, spoke very briefly, and showed no cards. The match then flowed normally. Read the statistics and you see a dull game. Sit inside the stadium and you see a decision bigger than every penalty in it.

This is why I always draft the "game management" section before writing the "controversial incidents" section. That order matters. People usually start from the controversy and then hunt for context. In my old trade the order was reversed. You must understand the state of the match before judging a decision inside it.

I learned that expensively.

On 14 June 2026, the World Cup opener in Russia between the hosts and Saudi Arabia at Luzhniki. I was twenty-six, working on site for a digital sports platform for the first time. In the first half I mispronounced the striker Artem Dzyuba's name three times in a row. Afterwards, on air, a male colleague laughed. I did not argue. I went back to the hotel and started a notebook.

That notebook had three columns: player names transliterated with the International Phonetic Alphabet, match data, and notes on refereeing decisions. I archived the full footage of all sixty-four matches of that tournament, and every night I spent two hours reviewing my own work, cross-checking technical errors against rule details cameras usually miss.

That mistake in Russia did not teach me how to referee correctly — it taught me how to live with my own whistle. And it taught me something more concrete: credibility does not come from pronouncing a name right. It comes from being willing to say "I need to look again."

That notebook later became my analytical framework. Every piece I write carries sources and verifiable data, because I know what it feels like to be caught out in public. And because I know that in football, a number without a source is not data. It is an opinion wearing a shirt.

Part 7 — The counter-view: anger with nowhere to land

Fans do not hate VAR. They hate the feeling that a decision was taken out of a person's hands and placed in the hands of a process nobody answers for. In traditional football, when a referee errs, there is someone to be angry at. That person has a name, a face, can be criticised, can be suspended. In the VAR era, when a decision is disputed, there is nobody to be angry at. There is a system to be angry at.

A system that is blamed cannot apologise. And an error nobody owns cannot be forgiven. This is the psychological cost lawmakers rarely price in.

It is also why I distrust every call to return to the past. Nostalgia for the pre-VAR game is usually nostalgia for a feeling, not for a standard. People miss being able to shout at a person. That is entirely reasonable psychologically, and entirely insufficient for building a rule system.

On the other side, I do not believe those who say more cameras will settle it. More data does not reduce argument if the data is not tied to a definition. We have multiplied the volume of evidence several times over in twelve years, and the volume of argument has not fallen. That is a strong enough signal that the problem is not quantity.

Part 8 — What can be done

If I had to propose a direction, I would start here: separate the geometric question from the advantage question.

Technology can answer precisely who was where, when, and by how much. That is the data layer. But the law asks more. It asks whether the player gained an advantage from that position. That is a question of causation, and it still needs a human answer.

In recent years a "clear daylight" threshold has been proposed for offside: call it only when there is visible separation between the two players. I do not think the proposal is perfect — it moves the argument from a line to a threshold, and every threshold can be disputed. But in principle it heads the right way: it admits that technology can be more precise than the law needs.

That principle should be applied more widely. For the ball out of play, the question should be whether the ball left the field to a degree that affected the move, rather than whether it crossed the line. For handball, whether the arm created a clear advantage. In every case the test should be materiality — did it change the match — rather than pure geometry.

Each time a threshold is widened, someone will say football is licensing ambiguity. They are right. But that ambiguity has always existed in football. What technology does is push it into the light. Whether we use that light to judge or to understand better is a human choice, not a machine's.

Part 9 — For Vietnamese football

For Vietnamese football I have one specific worry.

When a league lags in adopting technology, the natural reaction is to import it as fast as possible. But VAR is not a device. It is a procedure, and procedures need people who are trained, assessed and protected. If a league has VAR at half its matches and none at the other half, the biggest cost is not the hardware bill. It is referees switching between two decision-making modes inside a single week.

In many conversations with colleagues in Vietnam, I noticed something: their greatest fear is not technology failing. It is technology running correctly while the law is interpreted wrongly. A precise system laid over uneven rules education produces unpredictable precision — and that is the worst kind of injustice, because it has no pattern.

I believe this is where Vietnam could outperform many bigger football nations. Training in the laws is cheaper than buying cameras. Building a unified interpretive framework is cheaper than building a VAR centre. If Vietnam invests in language first and hardware second, that order could produce a long-term advantage.

And it connects to a belief I have held throughout my writing life: a good judgement is not a fast judgement. It is a structured one.

Part 10 — Closing

Late on 1 December 2026, after the images from Doha were released, I sat for a long time in front of the screen thinking about something very small.

Not one of the people arguing that night could say with certainty what they had seen. Not because they were ignorant. Because at that distance, at that speed, from that angle, a football has no fixed shape. It is a blur, and everything we call truth is built from a blur that has been stopped.

That is the most beautiful and the most uncomfortable thing about this game. It forces us to decide before we have enough grounds, and then to live with the decision.

In a match in an empty stadium, when you can hear the ball strike the boot, you realise that what we call evidence has never been complete. It is only slightly more complete than last time. Each new technology does not end the argument. It moves us to a new boundary, where we must again decide on the basis of what we cannot see.

And I think that is why I am still in this trade. Not because I want to know the answer. Because I want to know how I will define the question next time.