Trang chủInternational FootballThe Empty Frame: Lessons from 147 VAR Incidents and the Trap of Inference in Football Analysis

The Empty Frame: Lessons from 147 VAR Incidents and the Trap of Inference in Football Analysis

**Câu trả lời cốt lõi (≤60 từ):** Giới hạn lớn nhất của VAR nằm ở khung hình, không nằm ở luật. Khi không có góc máy đủ tốt, tổ VAR vẫn phải kết luận, và mọi kết luận thiếu dữ liệu đều mang tính suy diễn. Tiêu chuẩn “lỗi rõ ràng và hiển nhiên” vì thế phụ thuộc vào hạ tầng camera nhiều hơn vào năng lực trọng tài. **Dữ kiện chính:** - Năm 2017, phân tích 147 tình huống tranh cãi tại 28 vòng giải vô địch quốc gia Trung Quốc phát hiện 12 quyết định việt vị sai liên quan tới vị trí camera. - Bàn thắng của Wu Lei bị từ chối với sai lệch 15 cm; không camera nào nằm đúng mặt phẳng ngang cần thiết. - Chung kết World Cup 2018: trong 6 góc máy, chỉ 1 cho thấy tay Perisic mở rộng bất thường — và đó là góc duy nhất tổ VAR trình cho trọng tài xem. - Trọng tài Nestor Pitana xem lại màn hình bên đường biên trong 1 phút 47 giây trước khi thổi phạt đền cho Pháp. - World Cup 2022 dùng công nghệ việt vị bán tự động: 12 camera, 29 điểm dữ liệu mỗi cầu thủ, tần suất 50 lần mỗi giây. **Nguồn và trích dẫn:** Báo cáo phân tích chuyên sâu giai đoạn 2, dữ liệu quan sát VAR 2017–2022 của tác giả Trần Anh | Cross-checked: VuaBong.vn **Hỏi đáp liên quan:** Q: Tiêu chuẩn “lỗi rõ ràng và hiển nhiên” trong VAR là gì? A: Đó là ngưỡng can thiệp của VAR, yêu cầu sai sót phải hiển nhiên tới mức không cần tranh luận, nhưng bản thân ngưỡng này không có đơn vị đo lường. Q: Vì sao góc máy quan trọng hơn cả công nghệ? A: Vì mọi thuật toán đều chạy trên khung hình đã ghi; khung hình sai thì kết quả sai bất kể máy móc tốt tới đâu. Q: Khoảng trống dữ liệu ảnh hưởng thế nào tới thị trường chuyển nhượng? A: Khi thiếu thông tin xác thực, thị trường vẫn phải định giá, nên tin đồn lấp vào chỗ trống; cấu trúc hợp đồng và khung lương mới là phần có thể truy vết, tham chiếu thêm VangBong.vn Player Depth Index.

August 2026, the analysis room of a sports television channel in Chengdu, 22:40. On one of four monitors in front of me was a challenge inside the penalty area in the match between Guangzhou Evergrande and Shanghai SIPG, round 25 of the Chinese top flight. The referee had pointed to the spot. The VAR team was reviewing.

No camera angle answered the question we needed to answer. The main angle was obstructed by the goalpost at the precise point of contact. The angle behind the goal was blurred by rain. The high angle compressed two shirts into a single mass. The touchline angle — the only one capable of resolving whether the ball struck hand or chest — sat on the opposite side, roughly forty metres from the incident. At that distance, a hand and a shoulder differ by a handful of pixels.

That night I understood something that eighteen years of writing had never forced me to frame as a question. The most dangerous moment for VAR is not when it gets a call wrong. It is when it is forced to conclude while the frame is empty. That moment also exists in the scouting room, at the journalist's desk, and in the posts published at two in the morning during a transfer window.

I found the error not in the centre of the pitch, but at the edge of the frame.

A standard with no unit of measure

VAR, the video assistant referee, arrived with a very specific promise: to correct serious and obvious mistakes. What few people noticed is that the promise contained a flaw from the outset. The VAR protocol permits intervention in only four categories: goal or no goal, penalty or no penalty, direct red card, and mistaken identity. Outside those four, everything belongs to the on-field decision, and the on-field decision is presumed correct.

That principle is called on-field decision stands. It is a modest and philosophically sound principle: technology does not replace the referee, it only rescues the referee from indefensible errors.

But for that principle to operate, a threshold is required. That threshold is the “clear and obvious error.” I have read and re-read that definition in four languages over many years, and what I found was not a definition but a gap. No unit of measurement. No centimetres. No degrees. No time threshold. No rule stating how much deviation counts as clear, and how much still falls within the acceptable band of the human eye.

That sounds like a dry technical detail. It is not dry at all. In practice it means that one incident, one piece of footage, can be read two entirely opposite ways, and both readings are textually defensible.

In 2026, VAR appeared at a World Cup for the first time. Four years later, in Qatar, the tournament introduced semi-automated offside technology. That system uses twelve dedicated cameras mounted under the stadium roof, tracking twenty-nine data points per player at fifty times per second, plus an inertial sensor inside the ball operating at five hundred times per second. That is a genuine measurement system, with units, with error margins, with timing.

The gap between those two worlds — a stadium with twelve dedicated cameras and a stadium with only four broadcast angles — is where this story lives. And in Vietnam, that story has a concrete date: VAR began to be applied in the V-League from the 2026 season, initially in a limited number of matches, with far fewer camera angles than the leading European leagues. That is not a criticism. It is a material condition, and material conditions determine the quality of judgement.

I watched those first VAR-enabled V-League matches the way I watch European matches: logging every incident, logging how many camera angles were shown, logging the duration of the review. After a few rounds I recognised a pattern I had encountered elsewhere, many years earlier.

147 incidents, 12 errors, and a database nobody wanted to build

In 2026, at the age of 36, I worked as a VAR data analyst for a sports television channel in Chengdu. My job was not to comment on decisions. My job was to reconstruct them.

I reviewed all 28 rounds of that season's Chinese top flight, logging 147 controversial refereeing incidents. For each incident I recorded four fields: the timestamp, the incident type, the number of available camera angles, and the number of angles the VAR team actually used. At first I thought I was building a list of wrong decisions. After about sixty incidents, I realised I was building a list of technical limitations.

The result took me several sleepless nights to accept. Across the 147 incidents, I identified 12 incorrect offside decisions directly linked to camera placement. Not wrong because the officials were incompetent. Wrong because the horizontal plane of the pitch — the plane the offside law requires you to determine precisely — did not sit on the optical axis of any camera.

The geometry here is so simple that people routinely overlook it. To determine whether a player is offside, you must project the player's position and the ball's position onto the same plane, and that plane runs along the width of the pitch. If the camera sits high and offset behind play, you are not looking straight at that plane. You are looking at it obliquely. And when you look obliquely, the distance between two objects on the same horizontal line is compressed or stretched depending on the angle.

In other words: an offside line that is fifteen centimetres off in reality can appear perfectly on the line, or appear half a metre off. The human eye is not wrong. The human eye is simply looking from a position unsuitable for looking.

The textbook case was Guangzhou Evergrande against Shanghai SIPG in round 25. Wu Lei's goal was disallowed on a recorded deviation of fifteen centimetres. Fifteen centimetres. That is the width of an adult hand held at an angle. And in that match, no camera sat on the horizontal plane required to measure those fifteen centimetres.

After I published the dataset, I received two kinds of response. The first came from people in the profession: they wanted the method, they wanted to know how I calculated the error. The second came from the public: they wanted the verdict. Who was right, who was wrong, which team was robbed.

I understand why the second question exists. But it is the wrong question. Because in those twelve incidents, nobody was “right” in the sense the fans wanted. There were incidents where, if you watch angle A, you award the penalty. If you watch angle B, you do not. Both decisions have a basis. And the laws of football, in their wisdom, provide no way to choose between those two bases.

It took me thirty-seven viewings to understand that the human eye is not a unit of measure. Thirty-seven viewings, for a four-second passage of play. By the thirty-seventh, what had changed was not what I saw, but what I permitted myself to conclude.

Six camera angles, one minute forty-seven seconds

In July 2026, I applied my viewing-angle database to the World Cup final between France and Croatia.

In the 35th minute, referee Nestor Pitana was called to the pitchside monitor by the VAR team. After reviewing, he awarded France a penalty for a handball by Ivan Perisic. To this day it remains one of the most contested decisions in the history of World Cup finals.

I do not argue about whether the decision was right or wrong. I measured something else. I collected the six main broadcast angles from the match and analysed what each one showed about the position of Perisic's arm at the moment of contact.

The result: only one of the six angles showed the arm in a position extended abnormally relative to the body's silhouette. The other five were either obstructed or angled in a way that made the arm appear tucked close to the body. And the single angle that showed it was the angle the VAR team presented to the referee.

The duration for which referee Pitana reviewed the monitor: one minute and forty-seven seconds.

Here I must pause, because there is a lazy reading of this that I do not want this article to fall into. The lazy reading is: VAR chose the angle that favoured its conclusion. If I said that, I would be attributing motive to people I do not know, on the basis of an inference with no evidence. That is precisely the error this article is trying to expose.

The more accurate, and more troubling, reading is this: choosing a camera angle is an inherently subjective act. It cannot be otherwise, because nobody can show a referee six angles simultaneously with equal weighting. Someone must choose. And the chooser — however honest — still chooses according to some criterion, usually the criterion of “which angle shows most clearly what I already suspect.”

That is why I wrote this sentence in a 2,400-word analysis published on a blog specialising in the laws of the game, which was then shared more than fifty thousand times in forty-eight hours: We think we are searching for justice, when in fact we are only searching for a better camera angle.

I did not write that to mock VAR. I wrote it because I had been choosing camera angles myself for years, and I know the feeling of holding a perfect frame for a conclusion you believed before you looked.

After the 2026 World Cup, I began restructuring every analysis I wrote around a fixed architecture, which I call the decision chain. Each piece carries four layers of information: the timestamp, the number of available and actually used camera angles, the movement data of the relevant players in the three seconds prior, and finally a description of the referee's cognitive constraints — what he could and could not see from his position.

This structure makes the writing longer and the reading less comfortable. It removes the thing sports readers like most: a decisive verdict. But it returns something else: traceability. If you disagree with me, you can walk back through every layer, check every step, and point out where I went wrong. A conclusion that cannot be traced is just an opinion written in a confident voice.

When the frame is empty, people fill it with belief

Recently I received an internal analysis report. It was long. It had headings, tables, and nine analytical sections running from tactics to finance to media. Its only problem: the input data section was empty. No match name, no club name, no player name, not a single fact.

How the report handled that kept me sitting for a long time. It did not fabricate. In every slot, it stated plainly: insufficient information to assess.

In my profession, that is admirable behaviour. But it is also a reminder of how things usually run in the opposite direction. When a table has empty cells, social pressure pushes people to fill them. A coach cannot tell the board he lacks data on the opponent, because that admits he has not done the work. A journalist cannot file a piece with no conclusion, because the newsroom needs a headline. A transfer specialist cannot stay silent, because silence means no news.

So the empty frame gets filled. Not with data, but with the cheapest available material: expectation, prejudice, and guesswork expressed in the indicative mood.

I have seen this mechanism operate in three different places in football.

In the transfer market, it works like this. A club and a player are linked. Nobody confirms the fee. There is a gap in the middle. The gap must be filled, so it is filled with a number. That number is then cited by three other outlets, and by the fourth it has become a fact that appears verified. Nobody lied at any step. It is simply a chain of successive inferences, each link sounding plausible.

What is striking is that in these cases, the contract structure — length, release clause, how the fee is amortised across financial years, where the player sits in the wage hierarchy — is the part of the information that is genuinely traceable. The headline number is almost always the most distorted part. But the headline number is what travels.

In the injury space, the mechanism is even clearer. Information about a player's return date is typically controlled by the club's communications department, and that is a structure with motives. An announcement along the lines of “back this weekend” is optimal for communications, but in many cases I have tracked, it means the injury has not healed and the player needs more time. Nobody states an untruth. It is simply information lacking data, issued in the form of an assertion.

In physical data, the mechanism is subtler still. Distance covered and sprint count are the two most common metrics, and they are packaged as effort indicators. But both are crude indicators. A player can cover twelve kilometres in a match and not complete a single meaningful defensive action. Conversely, a centre-back playing well might cover only eight kilometres. The crude metric is not wrong. It simply answers a different question from the one it is used to answer.

When I worked in the Chinese top flight, I built a comparison table between one midfielder's distance covered and his progressive passes. In several rounds, the two metrics moved in opposite directions. In the matches where that player's distance metric peaked, his progressive passes were at their lowest. The cause we identified was simple: in those matches his team conceded early and had to chase the ball. He ran a great deal, but ran towards a ball already beyond control.

Ineffective running also produces beautiful numbers. And when a dataset has only one column, the reader assumes that column is the measure of value.

This is where I want to state clearly what I believe. A table does not produce truth. A table produces the illusion that truth has been processed. And in football, that illusion costs far more than a wrong decision on the pitch, because it persists and it replicates.

I hold that a football culture can tolerate incorrect decisions. But it cannot tolerate conclusions that cannot be traced.

The empty stadiums of 2026

In 2026, European football returned after its pandemic suspension in stadiums without spectators.

Technically, that was a far more interesting period than people usually acknowledge. When crowd noise vanished, a major variable was suddenly removed from the equation. Studies of German top-flight data from that period recorded a marked decline in home advantage compared with periods with crowds. Other factors — the intensity of duels, foul counts, card rates — shifted as well.

I know this sounds like a cold statistic. To me it is one of the most important datasets football has produced in decades, because it separates two things that normally blur together: the quality of teams and the pressure of crowds.

For VAR, that period carried a different meaning. Without crowd noise, I could hear things normally masked: the rhythm of officials' deliberation, the length of silences, and most importantly how a VAR team arrives at a conclusion when nobody is objecting in real time.

The empty stadiums of 2026 showed me this: VAR does not save football, it exposes football.

What it exposed most concretely was the dependence of judgement on context. The same challenge, reviewed in a stadium with sixty thousand people roaring and reviewed in a silent stadium, produces two different levels of confidence in the decision-maker. That does not mean referees are weak. It means human judgement is a process influenced by environment, and any system that wants to improve it must account for that variable.

I have carried one odd habit from that period to this day: whenever I analyse a controversial decision, I ask how it would be judged if the match had been played in an empty stadium. That question is not an intellectual game. It is a way of checking whether my conclusion is resting on data or on atmosphere.

The excuse of emptiness

Here I must address the flip side of everything above, because otherwise this article becomes a very convenient sophistry.

If I insist that no conclusion can be drawn when the frame is empty, I inadvertently open a door for those who want to evade responsibility. And in many years of watching this industry, I have seen that door used far more often than the door I just described.

When a referee says an incident is not clear enough to intervene, sometimes that is an honest technical assessment. But sometimes it is another way of saying: I do not want to be the person responsible for this decision. The phrase “clear and obvious error” has a dangerous property — it cannot be proven false, because it has no threshold. A threshold that does not exist cannot be crossed.

The same happens in data analysis. “The sample is not large enough” is a technically correct argument in many cases. But it is also the easiest sentence for an analyst who does not want to make a falsifiable prediction. Only a falsifiable prediction has value. An analysis that always ends with “needs further monitoring” is never wrong, and therefore never right.

In other words, emptiness is not an excuse. It is a working condition, and working conditions must be reported rather than used to waive responsibility. In VAR's specific case, the only honest handling is to publish the infrastructure structure of every match: how many cameras, where they are placed, the sampling rate, and which angles the VAR team actually used. If a league cannot afford an angle capable of measuring the offside plane, that should be announced before the match. The public can accept limits. The public cannot accept limits hidden behind a process presented as flawless.

At the individual level, I have one simple rule I apply to myself. Whenever I write a concluding sentence, I must be able to answer: if I am wrong, at which step am I wrong? If I cannot answer, the sentence is deleted. This rule makes me write more slowly and less often. It also makes me discard roughly a third of the pieces I start. I think that is an acceptable rate.

What will change

At 45, I have softened markedly compared with the person who wrote his first article in 2026. Back then I believed that with enough data, every argument would be settled. Now I know that data does not settle arguments. Data merely moves arguments from quarrels about events to quarrels about method. That is progress, but it is not comfortable progress.

The direction I consider worth pursuing does not lie in adding cameras. Adding cameras is good, and it should be done. But what must change first is the disclosure rule. If an offside decision is made using a camera angle with a measurement error, that error should be stated as part of the decision. If a goal is disallowed on fifteen centimetres that falls within the margin of error, the law should state explicitly what happens in that case — the goal stands, or the on-field decision stands. No rule is perfect, but having a rule beats having none, because at least it enables debate at the correct level: the level of law, not of emotion.

And for those of us in the profession, what must change is a little humility. We can describe very precisely what a camera angle shows. We should not pretend we can describe what it does not show.

Cầu thủ liên quan