One Wrong Label, One Collapsed Verdict: Why VAR Fails at Naming, Not Seeing

Minute 34. The referee leaves the penalty area, walks to the touchline monitor, and stops. The stands erupt. He is not rewinding the footage to see what happened. He is rewinding it to answer a different question entirely: what do we call this.

On the screen there is only one incident. Under the Laws, that incident can carry three different names — accidental handball, handball from an unnatural body position, or handball denying an obvious goal-scoring opportunity. Three names, three sanctions: nothing, a penalty, a penalty plus a card. Same frame, same frame rate, same camera angle.

Five years observing VAR rooms and thirty-seven years reading match footage taught me something uncomfortable: most major errors on the pitch do not begin with a referee seeing the wrong thing. They begin with a referee labelling the right thing wrongly. One incorrect label inside the VAR room collapses the entire chain of decisions behind it, and no camera angle can save it.

Football runs as a classification system

Law 12 of the Laws of the Game does not describe incidents. It describes categories of conduct: careless, reckless, using excessive force. A referee does not declare that a player fouled. A referee declares at which level the player fouled. The level is set by the label. The same challenge from behind carries three different names, three different sanctions, three different suspension lengths.

The 2026 World Cup in Russia was the first finals with VAR. A livestream platform brought me in as its rules analyst, and I built my own tracking sheet across all 64 matches. It logged 335 VAR interventions, 20 overturned decisions, and 10 penalties that originated directly in the VAR room. The figure that held my attention longest: correction accuracy reached only 68.4 percent in the group stage, then climbed to 91.2 percent in the knockout rounds.

That 22.8 percentage-point gap did not come from knockout referees seeing better. It came from knockout incidents falling more often into familiar classification boxes — penalties, offsides, direct red cards. The group stage produced more incidents sitting on the edge of the catalogue, where the label had never been defined conclusively and each referee had to decide which catalogue he was reading from.

Since 2026, IFAB has rewritten the handball definition at least twice. The 2026 revision introduced the phrase unnatural body position, with guidance that an arm raised above the shoulder is presumed unnatural. In the very next season, the Premier League recorded a sharp surge in handball penalties. In 2026, UEFA issued its own guidance: no penalty when the ball strikes a hand in a natural position, or after it deflects off the player's own body.

One incident, three governing bodies, three catalogues. A defender handling the ball inside his own area can be punished in the Premier League, escape in the Champions League, and receive a third outcome in an Asian league. To supporters this is injustice. To lawmakers it is three overlapping classification systems, none of which answers to the others. In Vietnam, VAR entered V.League from the 2026 season, and the same problem appeared immediately: Vietnamese referees do not lack camera angles, they lack a catalogue that matches the rest of the region.

Three ways a label goes wrong

From my own dataset, I separate classification failure into three distinct types. Each demands a different remedy, and mixing them together is the fastest route to fixing nothing.

Type one — the wrong box. The incident belongs in one category and the referee files it in another. On 15 July 2026, in the World Cup final at Luzhniki, the ball struck Ivan Perišić on the arm inside the France penalty area. Referee Néstor Pitana consulted VAR, awarded the penalty, Antoine Griezmann scored to make it 2-1, and France went on to win 4-2. The argument raged for years afterwards. The question was never whether the ball hit the arm — the footage settled that. The question was whether the contact belonged in the deliberate box or the body-position box. Where was Perišić's arm, how many metres from the shot, did he have time to react. That is a classification problem, not an observation problem.

Type two — right box, wrong threshold. On 6 March 2026, at the Parc des Princes, Manchester United visited Paris Saint-Germain. The ball struck Presnel Kimpembe on the arm inside the area. Referee Damir Skomina went to the monitor, reviewed it, and awarded United a penalty in the 94th minute. Marcus Rashford scored, United won 3-1, and advanced on away goals. The category was chosen correctly — handball inside the penalty area. But the threshold for turning it into an offence was pushed extremely low. Kimpembe's arm was slightly away from his body as he turned his back. How many degrees of separation become an offence? The Law states no number. There is no number to state.

Type three — right box, right threshold, broken process. On 30 September 2026, Tottenham hosted Liverpool. Luis Díaz put the ball in the net. The goal was disallowed for offside. Nobody disputed the category — offside is the clearest box in the entire book. But the VAR team misread the on-field decision, and the communication chain collapsed. England's refereeing body later issued a public statement acknowledging a significant human error. This is the most dangerous type, because it happens at the lowest layer: one input data point mislabelled, and every downstream step — each one perfectly executed — automatically produces a wrong output.

Based on my experience tracking matches, the three types are not evenly distributed. Types one and three are rare. Type two is the most common and the hardest to admit, because in that failure nobody is visibly wrong. All three share one troubling feature: none of them lives in the referee's eyes. They live in the catalogue the referee is reading.

In my tracking sheet, every contentious incident occupies three columns: what happened, what it was called, and what it should have been called. The third column is the one nobody wants to fill in, and the only one that reveals where the system is breaking. In my own work auditing player contracts, a record carrying the wrong label drags every conclusion behind it into error, even when every figure inside the record is accurate. Football is no different. An incident described correctly down to the last frame can still generate a completely wrong decision, simply because it was filed in a box that was never built for it.

The machine cannot repair the catalogue

Supporters believe the problem is technology. They demand more cameras, more angles, a chip in the ball, semi-automated offside. Since the 2026 World Cup, semi-automated offside has been on the pitch, accurate to the millimetre. But it answers exactly one question: where was the player at the moment the ball was played. It cannot answer the larger question of which clause this incident breaches.

Here is the counter-intuitive point. The more precise football becomes visually, the more clearly it exposes that its bottleneck sits in language. The fiercest arguments now are not about whether the ball touched the hand — mostly that is settled. They are about what that contact is named.

Through a referee's eye, you cheer for nobody. You only look for who is right. But to find who is right, you must first name the act correctly. Name it wrongly and you can punish the right person for the wrong reason, or the wrong person for the right reason. Both are system failures, and neither can be repaired with an eighteenth camera.

There is one more layer. The clear and obvious error threshold that VAR operates on also conceals classification errors. A referee is only summoned to the monitor when an error is obvious enough. But classification errors are rarely obvious in the first frame — they become obvious only when you place two names side by side. The mechanism is excellent at catching errors of observation and rather poor at catching errors of naming.

Name it first, judge it second

In recent seasons, refereeing bodies in several leagues have begun requiring referees to announce decisions over the public address system. Most read out only the outcome: penalty, no penalty, yellow card. I would take one more step: announce the label too. Handball, unnatural body position, penalty. One extra sentence, and every contentious decision becomes a line of data that can be audited, cross-checked and counted later.

335 VAR interventions, 335 times the Law was called by name in the middle of the pitch. If each naming were recorded publicly, football would gain something it has never had: a classification register. After each season, people could sit down and ask which box is being misused most often, which league is applying a stricter threshold than the next, and why one act receives two different verdicts on two sides of a border. Repairing one label costs far less than repairing an entire system. And every refereeing reform of the next decade will start in exactly that place.