The Empty Analysis Sheet and the Limits of the Right to Decide on a Football Pitch

Kazan, 16 June 2026, 58th minute. Andres Cunha, the Uruguayan referee, left his position in the middle of the pitch and walked towards the pitchside monitor set up beside the touchline. Seventeen seconds earlier he had waved play on. Joshua Risdon went into Antoine Griezmann's left leg inside the penalty area; the ball rolled on towards Mathew Ryan. Cunha reviewed the incident three times, at two different speeds, then pointed to the spot. Griezmann converted, Mile Jedinak equalised from the penalty spot, and France won 2-1. It was the first penalty in World Cup history awarded through the video assistant referee system.

I was sitting roughly two metres from a television in Nagoya, twenty years old, writing down every second of that sequence in a notebook: the moment Cunha touched his earpiece, the moment he drew a rectangle in the air with both hands, the number of replays, the distance between where he stood and the screen. Not one line in that notebook asked whether Cunha was right or wrong. What I wanted to know was where the authority to decide had moved, and who was holding it after the whistle.

Six years later, a nine-column analysis sheet landed on my desk. Nine columns: tactics, club finance, results and the public-opinion cycle, league landscape, rules and governance, the dressing room, the risk profile, media narrative, and industry transmission. Each column carried four to six rows. Every cell was empty. No team names, no player names, no dates, no sources. The sheet was the output of a data-extraction pipeline that had failed at intake, and it arrived with a single instruction attached: analyse this.

How you respond to a sheet like that says more about this profession than any match does. There are two options. The first is to fill the cells with things that sound plausible: a big club, a coach under pressure, an expensive transfer, a freshly amended law. Nobody can check any of it, because nobody knows what the original article was about. The second option is to return one sentence: there is not enough information to conclude anything. In twelve years of watching this industry, I have seen the first option chosen roughly ten times more often than the second.

What makes this worth writing about is that football already has a rulebook for exactly that situation, and the people running the system ignore it every single week.

Context: a season in which the handball law contradicted itself

In the 2026/21 season, IFAB introduced a new interpretation of handball built around the phrase about an arm making the body unnaturally bigger. Before that, in September 2026, the panel had already had to clarify that an accidental touch by a teammate would no longer wipe out a goal. Then in March 2026 it clarified the wording a third time: the unnatural-size test only applies where the arm position is not a consequence of, and cannot be justified by, the body movement in that specific situation, and an arm above shoulder height is likely to be a breach. Three clarifications in seven months for one clause. The Premier League that season hit a record number of penalties awarded, and most of the arguments centred on incidents where the referee himself was not certain what he was punishing.

The paradox sits here: the more the clause was clarified, the blurrier the line became. A committee of many people in Zurich tried to model the human arm in language, while the human arm moves roughly two orders of magnitude faster than language does.

At the same time, the VAR framework operates on a completely different logic. The system may only intervene in four categories: goals and the incidents leading up to them, penalty decisions, direct red cards, and cases of mistaken identity. The VAR does not blow the whistle for the referee; he can only recommend an on-field review when he detects a clear and obvious error, or a serious incident that was missed. That phrase, clear and obvious, has become the shield in every argument that never ends.

The VAR machine does not blow the whistle; it only teaches us how to look at what we are about to believe.

When the pandemic season arrived, I used the long, directionless free time to test whether that claim held up in the numbers.

Core: 47 incidents, and a criterion that does not exist in the law

I pulled Opta data on 47 handball incidents punished in Europe's leading leagues and coded every one of them across four variables: the angle between the arm and the body's vertical axis at the moment the ball left the opponent's foot, the angle between the arm and the shoulder line, the distance from the shooter's foot to the arm, and the reaction time between ball release and the first movement of the arm.

Several of my initial hypotheses died. Distance predicted almost nothing. Reaction time predicted nothing either: most punished incidents had a window shorter than normal neural reflex time, meaning the arm had no chance to withdraw, and the law does not account for that. The only variable that separated punished arms from unpunished ones with any reliability was the deviation of the arm from what I came to call the natural body silhouette, constructed by extrapolating the torso outline if the arm had stayed in its anatomical position at the current posture.

Referees are not punishing handball under IFAB's clause. They are punishing a geometric standard that IFAB has never defined in any document.

In other words, an unwritten rulebook exists. It is transmitted through training camps, through pre-match referee meetings, through the collective memory of incidents that once caused a storm. That unwritten rulebook operates fairly consistently, and it is not in the Laws of the Game that anyone can buy.

My empty nine-column sheet belongs to the same family of problems. It forced a choice between two kinds of expertise. The first produces a document that looks substantial: every column has a judgement, every judgement has a strong verb, every verb has an adverb manufacturing certainty. The second publishes the fact that there is nothing to analyse, adds a diagnosis of why there is nothing, and requests that the extraction step be re-run on a verified source text.

The only valid conclusion when the data is empty is the conclusion that there is no conclusion yet, and in football that is the sentence nobody wants to say.

Nobody wants to hear it, because football runs on a binary command. The ball is in the net or it is not. Foul or no foul. Yellow or red. A referee has no standby mode. He must produce a decision within seconds, even when the input sits below the threshold required for any decision at all. You cannot raise a flag that reads: I do not have enough data.

Clear and obvious - how sports law names its own helplessness.

The contrarian angle: we do not need a more accurate VAR, we need a law that admits uncertainty

The Euro 2026 semi-final at Wembley, the night of 7 July 2026, 104th minute, England against Denmark. Raheem Sterling went down inside the area after contact with Joakim Maehle. Danny Makkelie pointed to the spot without leaving the field to check the monitor. Harry Kane missed, then scored the rebound. Europe split in half.

I was a first-year staffer at a sports broadcaster then, and the only person in the room defending Makkelie. My argument was simple and dry: if the error is not clear and obvious, the referee is entitled to stand by his original decision. A shoulder-to-shoulder contact at live speed, replayed in slow motion from a camera behind the stand, does not prove a clear and obvious error. Colleagues pushed back hard, accusing me of defending the system. A veteran editor noticed the unusual meticulousness and handed me a regular column on the laws.

Looking back, I think both sides framed the problem wrongly. Those demanding VAR intervention were demanding something the system never promised: absolute accuracy. Those defending Makkelie's call, myself included, were defending a decision using a clause written to avoid making decisions. Both camps were arguing about an incident the footage cannot resolve.

I do not watch matches through a spectator's eyes, but through the eyes of the man the spectators are judging.

From that vantage point, the real problem is that the law offers no exit. A system designed to settle disputes with evidence has no button marked insufficient evidence. The consequence is that any referee, forced to choose between two wrong readings, will choose the one that generates fewer enquiries the following week. That is a psychological problem, not a legal one.

The same disease spreads across the industry. A team keeps 62 per cent of the ball and is described as controlling the game, while most of those passes travel sideways between two centre-backs and change nothing about the opponent's defensive structure. Expected goals get used to conclude that one side played better, even though the metric describes chance quality, not refereeing decisions, not individual form, not the quality of a single passage. Esports betting is eroding competitive integrity faster than traditional sports, simply because the regulatory framework lags the technology. What all three examples share is not that people misread data. It is that they conclude before the data exists, then build the argument backwards.

With the nine-column sheet, I applied a twenty-four-hour rule. I finished the diagnostic, left it overnight, and read it the next morning with an editor's eye for faults. The only change I made was trimming the technical explanation of the extraction process, because readers do not need to know where the pipeline broke. They need to know that it broke, and that nobody is allowed to fill empty cells with imagination.

Takeaway

If IFAB added an official category to the VAR protocol called insufficient grounds to overturn, with a duty to publish a confidence level for every major decision in the match report, this industry would save thousands of hours of pointless argument every season. Not because referees would get more calls right, but because for the first time they would be permitted to say honestly that they are unsure.

A system willing to write the words not certain into the record is more mature than a system that always knows it is right. On the pitch, we are standing in front of a gate nobody has dared to open, while the empty cells remain crowded with conclusions that were written in advance.