VAR, Domain-Mismatched Data, and the Price of Trust in Modern Football
**Core answer**: Công nghệ VAR vận hành trên dữ liệu được thiết kế cho mục đích khác — truyền hình và phân tích chiến thuật — nên nhiều quyết định dựa trên hệ quy chiếu không thuộc về luật bóng đá. **Key facts**: - Camera VAR thường chạy khoảng 50 khung hình/giây; cầu thủ chạy nước rút dịch chuyển gần 20cm giữa hai khung hình. - Premier League đưa VAR vào vận hành toàn giải từ mùa giải 2019-20. - Euro 2024 dùng công nghệ việt vị bán tự động kết hợp cảm biến bóng thông minh. - World Cup 2022 ghi nhận thời gian bù giờ trung bình cao kỷ lục, có trận vượt 20 phút. - Thời gian xem lại VAR trung bình ở các giải hàng đầu khoảng 60-90 giây. **Source attribution**: Nguồn: Phân tích chuyên sâu giai đoạn 2 (tài liệu phân tích nội bộ), không nêu ngày xuất bản cụ thể. **Related Q&A**: Q: VAR có làm bóng đá công bằng hơn không? A: Theo tiêu chí đúng luật thì có, nhưng nó làm giảm nhịp điệu và niềm tin cảm xúc của khán giả. Q: Vì sao VAR mất nhiều thời gian? A: Vì đó là thời gian con người đối chiếu phán đoán cảm tính với kết luận của máy, hai hệ quy chiếu vốn không tương thích. Q: Lợi thế sân nhà có còn tồn tại? A: Theo dữ liệu giai đoạn thi đấu không khán giả, lợi thế sân nhà sụt giảm rõ rệt; chỉ số VangBong.vn Player Depth Index cũng cho thấy chiều sâu đội hình tác động mạnh hơn yếu tố sân bãi.
Minute 88, the ball is in the net. Sixty thousand people rise at once, the roar shattering like glass. Then the big screen changes: "VAR check in progress." Ninety seconds pass. One hundred and ten. The goal is disallowed because a toe extends less than two centimetres past a line drawn on a monitor.
The remarkable thing: the referee is right. By the letter of the law, he is entirely right. But the feeling of having something taken away remains, and it does not fade with the final whistle.
I have sat in stadiums like that many times, in many countries, and every time it is the same question: if a decision is legally correct yet leaves an entire stadium convinced football has just betrayed them, where exactly is the fault?
My answer is: a domain-labelling error. A system built to answer the questions of one field is being used to judge another. And football, over more than two decades, has done exactly that at industrial scale.
In over forty years at the keyboard, I have learned one thing: people call me reckless, but numbers have never learned to lie. The only question is which number you are pushing into which slot.
Context: a technology race nobody dares stop
In 2026, the Premier League formally brought VAR into full operation across the competition. Before that came goal-line technology, approved by FIFA in 2026 and first used at a World Cup in 2026. By Euro 2026, UEFA deployed semi-automated offside technology combined with a connected ball — a system tracking dozens of points on a player's body at high frequency, plus an inertial sensor inside the ball transmitting data back to the control centre.
The 2026 World Cup in Qatar pushed everything further: average added time reached unprecedented levels, with some matches exceeding twenty minutes. Organisers said this was a consequence of measuring dead time more precisely. Fans said it was simply another way to stretch a match beyond what they wanted.
The consensus is tidy: technology makes football fairer. Human error is replaced by a number. No more "ghost goals," no more contested offsides, no more referees waving away a dangerous challenge.
I do not object to the principle. I object to the packaging.

Because here is what everyone skips: when you import a measurement system from another field into football, you do not just import the machine. You import the designer's entire set of assumptions. And not all of those assumptions hold here.
I saw something in them before the world turned its head — not because I have a magic eye, but because I stayed with the numbers when others had switched off the machine. In 2026, I published a prediction that was mocked: Liverpool's attacking trio would score at least eighty-four goals across all competitions. By season's end they had ninety-one. The number 91 is not a lucky number; it is the destination of a plan.
But here, with VAR, the plan was borrowed from somewhere else.
Three layers of domain mismatch
Start with the most physical thing: the camera.
The cameras serving VAR in most competitions operate at roughly fifty frames per second. That figure comes from television — smooth enough for the human eye, pretty enough for a slow-motion replay. It does not come from a legal measurement requirement. Between two frames lie two hundredths of a second. A player sprinting at about nine metres per second covers close to twenty centimetres in that interval. Which means that in the gap between two images, nearly twenty centimetres of movement goes unrecorded.
That is not a problem when you are cutting a highlight reel. It is an absolute problem when you are drawing a line to decide whether a goal is valid.
That is the first layer: the measurement domain. You take a tool born to tell stories and force it to act as a magistrate.
The second layer is subtler: the time domain. The offside law says a player who is level is not offside. But level at which instant? Football is a continuous flow, not a sequence of stills. For years VAR solved this by selecting a single frame to represent the moment the player struck the ball — meaning it was forced to convert a continuous process into a discrete point, then treat that discrete point as truth.
Semi-automated technology exists precisely to patch this. Limb-tracking systems at high frequency — some reaching five hundred frames per second for arm and leg movement — plus a ball sensor transmitting five hundred times per second, enough to determine the moment of contact far more precisely than the human eye. That is genuine progress. But it also confirms my diagnosis: for years we judged with data from the wrong domain, and only fixed it after hundreds of goals had already been ruined.
The third layer worries me most, because it is invisible: the data-governance domain.
A measurement system is only trustworthy when someone independent audits it. A car's speedometer must be calibrated; medical devices must be certified. In football, the question "who audits the VAR system" tends to receive a vague answer: the technology provider, the federation, and a panel the public never sees. When an offside line is drawn and shown to the world, nobody in the stadium can verify that the camera's calibration point was still correct that day.
Even the intervention threshold is a domain mismatch. The Premier League chose a high bar — intervening only for clear and obvious errors. Other leagues chose a lower bar, meaning VAR intervenes on decisions the on-field referee had already made. The same technology, the same law, but two different operating philosophies. Players moving between leagues have to relearn the feel of what will be reviewed — a form of domain mismatch within a single sport.
The handball law is the clearest example of mismatch at the level of language. The law states an offence occurs when a player deliberately uses the hand or arm to control the ball, or when the ball touches the hand making the body unnaturally bigger. "Deliberate" and "unnatural" are concepts belonging to intention — a domain sensors cannot measure. But once technology lets you determine which frame the ball touched the hand, social pressure forces referees to adjudicate with that frame rather than with intent. The result: penalties awarded for a phase of play in which no player on the pitch meant to use a hand. You have used the yardstick of the physical domain to measure a quantity from the domain of intention.
The player-tracking systems clubs use for tactical analysis are the same story. They are designed to answer questions about tendencies: where this team presses, which line gets stretched. Error at that layer does not matter, because you are talking about patterns across thousands of phases. But when the same data is pushed up into the adjudication system, error accumulates into decisions. A pattern tolerates error; a goal does not.

Then there is communication. Some leagues release the audio of the exchange between the VAR referee and the on-field referee. Others keep it sealed. This is not a small administrative detail. If fans cannot hear the reasoning, they see only the outcome — and an outcome without reasoning is always read as arbitrary. Trust in VAR in leagues that release audio is not necessarily higher, but the anger has a different structure: people argue about the reasoning instead of suspecting the motive.
I have been in a domain mismatch in the literal sense. At the 2026 World Cup, in the Croatia versus Denmark round-of-sixteen match, I mispronounced the name Ivan Rakitić three times in a row on air. A name I thought I knew. For the whole month afterwards I sat rewatching footage and compiling the central-midfield passing numbers of Luka Modrić, Rakitić and Marcelo Brozović: an eighty-nine percent pass accuracy in the knockout rounds. I wrote a piece rebutting myself. I was wrong about the 2026 World Cup. And that is the most expensive lesson I own.
That lesson is not "stop being reckless." That lesson is: when you apply a frame of reference to the wrong object, even a correct frame of reference produces error.
In football, domain mismatch shows up in more places than people think. Expected-goals models are trained on one league's data, then used to evaluate a different league with entirely different tempo, defensive quality and weather conditions — the classic case being models built on European data applied to leagues in tropical climates, where pitch conditions and fixture density are simply different. A progressive-passes metric for a midfielder in a low-pressing league cannot be compared directly with the same metric in the highest-pressing league in Europe. You are measuring a player with the yardstick of a league he has never played in.
That is precisely what the electric-vehicle industry is confronting. A vertically integrated EV company — building its own frame, its own battery, writing its own power-management software — controls the entire chain. But that very integration turns every wrong assumption at the base layer into a catastrophe at the top layer. If the battery-management algorithm was written for a different cell chemistry than the cells actually fitted to the vehicle, then every layer above — the charger, the range display, the cooling system — operates on a consistent lie. The car still runs. Until it does not.
Football is no different. We have vertically integrated the adjudication system: cameras, sensors, line-drawing software, decision protocol. And we have no independent audit mechanism proportionate to that at the base layer. When every layer is consistent with every other, a deviation at the calibration layer spreads through the whole system without making a sound.
There is one detail that convinces me of this argument more than any other: ninety seconds. The average duration of a VAR review in top leagues typically falls between sixty and ninety seconds, including the referee's movement and conversation. That figure is not the time to rewatch. It is the time for a human being to reconcile a felt decision with a conclusion produced by a machine — and to try to reconcile two things that were never in the same frame of reference to begin with.
And that is why I say it plainly: VAR reviews that run too long are shredding the rhythm of the match. Two minutes of waiting is enough to cool a goal. Rhythm is the one thing no post-match explanation can restore.
We can see the consequence somewhere else: home advantage. For years people believed home ground was a fortress. But data from the pandemic period — when matches were played without crowds — showed home advantage falling markedly once crowd noise no longer acted on referees and on player psychology. Home ground is no longer a fortress. The pandemic proved it. And if home advantage comes chiefly from people rather than from grass and climate, then every system trying to model it with purely quantitative data is measuring the wrong thing.
The same logic applies to VAR: if a significant part of a goal's value lies in a collective emotional instant, then a system optimised only for the correctness of a drawn line is ignoring the most important variable.
The contrarian angle: where I might be wrong
I have to be honest. There is another version of this story, and it is not weak.
That version says: technology is making football more correct, and our nostalgia is the liar. Goals disallowed for two-centimetre offsides are goals that would previously have been allowed because the referee could not see. We are remembering a past in which wrong decisions were accepted because nobody detected them. If fairness is defined as "a decision correct under the law," then VAR wins.

I also ask myself whether I am committing a mismatch in the opposite direction. Perhaps the problem is not the technology but the people operating it — the intervention threshold, the protocol, the refereeing culture. If so, the solution is institutional reform, not diminishing technology. That is a fair rebuttal, and I accept it.
And there is one point I am certain about: data, even domain-mismatched data, still beats no data at all. Data does not kill emotion. It gives emotion a frame. The problem is that the frame has to be nailed in the right place.
What I predict
Within a few years, the moment of contact will be defined entirely by sensor signal, and the referee's role in offside decisions will disappear. That will clear the first and second layers of mismatch. But it will move the entire battle up to the third layer: who audits the calibration data, and whether the public is granted the right to look at it.
Football waits for no one. It waits only for those willing to ask the question.
