When the File Is Empty: VAR, Data and the Discipline of Judgement in Football Analysis

When the File Is Empty: VAR, Data and the Discipline of Judgement in Football Analysis

In the 34th minute at Tottenham Hotspur Stadium on 30 September 2026, Luis Díaz put the ball in the net. The assistant's flag went up. Referee Simon Hooper let play continue. At Stockley Park, the VAR room began its review, drew a line, and after more than a minute delivered its verdict: no goal. Liverpool left London beaten 1-2, with the goal wiped out at 0-0. Three days later, the Professional Game Match Officials Limited issued a statement admitting a "significant human error" — the phrase the English game still treats as a rare confession from a famously closed organisation. The audio released later showed the exchange between the VAR team and the on-field referee was out of step from the very first line.

A whistle can change a fate, but it cannot change what happened on the pitch. What stands out is that almost nobody that night wanted to know what actually happened. They wanted a verdict, and they wanted it immediately.

I remember the night I sat in front of an empty file. No match name, no teams, no players, no timestamps, no sources. Just a blank analytical frame and a deadline drifting towards me. When my editor called for the piece, the most honest answer I could give was that there was nothing yet to write.

This profession does not reward that answer.

Context: a game run on verdicts

The 2026-20 season was the first in which the Premier League ran VAR across all 380 matches. Since then, almost every matchweek in England has produced at least one argument capable of lasting until Thursday. VAR did not create refereeing errors — errors existed as long as referees have. VAR changed what is visible. A wrong call used to stay on the pitch and disappear. Now it is replayed twelve times in slow motion and then lives forever online.

A referee's mistake does not disappear with the whistle; it survives through every season.

In the same period, the laws changed faster than at any time in memory. IFAB, the body that owns the laws, added the access rule: from 1 July 2026, only the captain of each team may approach the referee over contentious decisions, with any other player who encroaches liable to a caution. That change targeted the image of eleven men surrounding one — an image television sells very well and which quietly destroys the referee's authority.

Football rewrites its laws every three years, but the trust of the crowd is far harder to rewrite.

The Premier League introduced semi-automated offside technology from the 2026-25 season, after UEFA used it in the Champions League and FIFA deployed it at the 2026 World Cup in Qatar. In Qatar, the official Al Rihla ball carried an inertial sensor operating at 500 Hz — five hundred measurements per second — combined with twelve stadium-roof cameras tracking twenty-nine data points on each player's body. Technically, this was a leap. Emotionally, nothing changed: the crowd still stands and waits, still stares at a red line, still accepts a machine's conclusion.

People hate VAR because it is slow; I value it because it refuses to hurry.

That is the environment any analyst now works in. More data than ever, more camera angles than ever, more laws than ever — and less time than ever. A tackle happens at 21:07 and the first analytical piece must be live by 21:15. Those eight minutes are the entire budget in which a professional decides whether to be honest or fast.

Anatomy of a VAR check

A review at Stockley Park is not a free debate. It is an ordered sequence of questions, and the order shapes the outcome more than most people realise.

The first question is factual: what happened? Did the ball go out? Was the player offside? Where did contact occur on the body, and at which second? This is the driest layer of data and the only one answerable by image and sensor. Yet even here a problem appears: frame rate. Broadcast runs at 50 or 60 frames per second, meaning each frame sits twenty milliseconds apart. A player running at thirty kilometres an hour covers about 8.3 metres per second, which is roughly seventeen centimetres between two frames. In a marginal offside, seventeen centimetres is the distance between a goal and a goal disallowed. Semi-automated technology was built precisely because of that limit, shifting the question from "which frame" to "which millisecond" — a far higher-resolution question.

The second question is interpretive: does that event breach the law? This is where IFAB Law 12 becomes the real referee. The law does not say a shirt pull is a penalty. It grades conduct — careless, reckless, or using excessive force. Those three categories cannot be measured by a machine. They are measured by experience: how many matches an official has controlled, how many similar moments they have seen at real speed.

The gap between data and law is where every football controversy is born. Not where the machine errs, but where the machine is right and the law still needs a human to read the result into a decision.

The third question is about match management: if you punish, how? Yellow or red, penalty or free kick outside the box, a quiet word with the captain or none. A referee holds three powers: to blow, to card, and to stand firm under pressure. The first two are written in the laws. The third is written nowhere, and without it the first two mean nothing.

In May 2026, while working as an assistant editor for a football site in Manchester, I gave an entire night to one incident in matchweek 36. Referee Mike Dean did not award a penalty when Arsenal forward Theo Walcott was pulled down by Stoke City defender Bruno Martins Indi inside the box in the 67th minute. The match finished 2-2. I watched twelve camera angles, timed the contact at four-tenths of a second, and wrote a piece measuring the incident directly against IFAB Law 12. It was shared more than five thousand times and opened a career path writing about the laws.

What I learned that night was not a style of writing but a habit: every refereeing analysis must carry a still frame with a timestamp, and every conclusion must be tied to a specific article of law. No "it might be a penalty". No "it looked like". Only 67th minute, twelfth second, contact of 0.4 seconds, per Law 12, section 1.

The three layers of every decision

Across five seasons in English stands, watching matches with both my eyes and a screen, I settled on a simple classification that fits almost any controversy: every refereeing decision rests on three layers of data, and their reliability differs sharply.

The first layer is physical fact. Did the ball cross the line, was the player offside, did the ball strike the hand, was contact inside the box. This layer is measurable, and when measured it is close to absolute. Aston Villa against Sheffield United on 17 June 2026 is a cruelly elegant example of this layer collapsing. That day English football returned after the pandemic shutdown, and the Hawk-Eye goalline system failed: it did not trigger the watch on referee Michael Oliver's wrist, even though the ball had crossed the line. The match ended 0-0. Hawk-Eye issued an apology, explaining that its cameras had been obstructed at the decisive angle. This was bad data at the hardest layer, and no VAR could rescue it, because VAR has no jurisdiction over goalline decisions — it can intervene only on goals, penalties, direct red cards and mistaken identity.

The second layer is legal interpretation. Here the same image can legitimately produce opposite conclusions, and here the public loses patience fastest. A challenge that wins the ball first and the man second can, under the law, be a fair one. The same challenge, executed with a raised stud and high speed, can be serious foul play. The law does not measure force in newtons. It measures danger — whether the opponent was placed at risk of injury.

The third layer is management and psychology. It appears in no analytical model, yet it shapes most of what the crowd feels. In the fifth minute of a goalless match, a referee is slower to whistle than in the 85th minute of a settled game. In a derby, the threshold for a card drops of its own accord. None of this exists in an official document, and everyone in the trade knows it is real.

The three layers never answer the same question, and most football arguments are simply two sides talking about different layers.

Arsenal supporters see layer one: the ball hit the hand, so it is a foul. Opposing supporters see layer two: the arm was in a natural position, the distance was too short, it was unavoidable. The referee sees layer three: this team is a goal down, and if I give a penalty in the 90th minute I own the argument for a decade. All three are right from where they stand. None can win, because the argument was never on the same plane to begin with.

The trap of the empty file

Back to that night with the empty file. In this trade the situation occurs more often than the public imagines, and it carries a specific trap: an information gap always fills itself with assumptions, and assumptions always sound a great deal like facts.

An analyst's brain cannot tolerate a gap. Given no team names, it infers them from context. Given no dates, it selects a plausible season. Given no sources, it trusts memory. Each small step compounds into a fluent article with numbers, reasoning, a conclusion — and a total error.

I have been on the other side of that trap. In June 2026, at the World Cup in Russia, I published a series defending VAR after the technology missed a handball by Uruguay defender Diego Godín against Saudi Arabia. I argued that VAR was not wrong, that its operators were, and cited an average of 6.8 minutes of stoppage per match for checks. Social media labelled me an emotionless robot. A BBC editor responded bluntly that the writing was too rigid. I kept the position but conceded the phrasing needed to be softer to stay in the conversation.

Since then I have worked to a five-step routine I do not allow myself to break: gather data, filter by variable, check against the law, draft, then verify every figure again. The last step is non-negotiable. I ask myself where each number came from. If I cannot answer, the number is deleted — no exceptions, however much better it made the piece read.

That same routine carried me through the pandemic shutdown in 2026, when football stopped and the newsroom had no matches left to cover. I proposed mining VAR and offside data from the 92 matches already played in 2026-20 and built a spreadsheet by hand. The result: around 34 per cent of goals reviewed by VAR involved an offside element, and clubs such as Sheffield United lost points on marginal calls. The series lifted the outlet's readership by 40 per cent that June, and I was handed all tactical content when football returned to empty stadiums.

What I carried out of that period is one principle: when the data has not arrived, the most honest product is controlled silence. That silence is not surrender. It is a statement that we know the limits of what we know.

Pricing a mistake

No mistake in professional football is free. Every wrong call carries an invoice, and the invoice is paid by people who never sit in the VAR room.

In the Premier League, prize money between final positions differs by several million pounds per place — the gap commonly cited by football finance specialists is roughly two to three million pounds. A club finishing 15th instead of 17th can bank a sum large enough to cover a key player's wages for a year. At the bottom, the cost of relegation is estimated above one hundred million pounds once broadcast money, sponsorship and matchday revenue are combined. Three points can be one refereeing error, or those three points can be the entire future of a club with a hundred years of history.

Those errors also become a commodity. When Nottingham Forest posted criticism of VAR after facing Everton in April 2026, the club was fined 750,000 pounds in October of that year — likely the heaviest sanction ever imposed on a Premier League club for comments about officiating. It marked a boundary: clubs may complain through official channels, but they may not turn social media into a forum for personal attacks on officials.

The darker side is the human cost. Referees in England have spoken publicly about pressure and personal abuse, and some within the professional group have stepped back from VAR duties in certain matchweeks for mental health reasons. These numbers never appear on a scoreboard, but they are real expenditure. A league that exhausts its refereeing pipeline does not collapse in one season. It collapses in ten, when nobody wants the whistle in a big match.

Meanwhile money keeps flowing along routes supporters rarely see. More than four hundred million pounds was paid out by Premier League clubs in agent fees in a recent financial year. Agents are the largest hidden cost of the transfer market, and the noise they generate — leaked negotiations, half-statements to reporters, claims of interest from clubs that never called — distorts how fans price a player. Once the valuation is wrong, everything downstream is wrong: expectation, pressure, minutes played.

At the macro level, pre-season tours keep turning clubs into travelling circuses. Transcontinental flight schedules, sold-out stadiums, commercial signings at every stop, and conditioning work strip-mined for commercial return. Injuries sustained in that window never enter the season's statistics, but their consequences stay all year.

Homogenisation and the death of diverse argument

There is a deeper layer analysis rarely touches, and it bears directly on how we judge refereeing controversies.

Elite football is homogenising. Fifteen years ago, a classic winger hugged the touchline, beat his man and crossed from the byline. Now most wingers invert, becoming attacking midfielders in a narrow channel, shooting with their stronger foot from the edge of the box. The shift has sound tactical reasons: more high-value shots, more control in the opposition half.

It has an overlooked consequence. When every team attacks through the same shape, the incidents that generate controversy become identical too. A decade ago, a penalty might come from a slide tackle after a wide dribble. Today most penalties come from one of three sources: contact as an inverted winger cuts inside, a handball after a blocked cross, or a marginal offside on a through ball. Those three sources dominate the argument, season after season.

The analytical models homogenise to match. Analysts use the same metrics, the same thresholds, the same comparisons. When everyone measures with the same ruler, that ruler stops registering whatever falls outside it. The moments that should define a match — a lone run down the flank, a cross from a position the model rates as low value — become invisible in analysis before they become invisible on the pitch.

That is a genuine loss. A game with fewer player types will produce fewer kinds of controversy, and a game with fewer kinds of controversy is intellectually poorer. The analyst's job is to preserve what the metrics discard — not out of nostalgia, but to keep the complexity of the sport intact.

Governance: between power and transparency

No reform in modern football moves as fast as the reform of refereeing authority, and none is more misunderstood.

Publishing the audio of referee-VAR exchanges was the turning point. In England, a recurring programme replaying released recordings, with PGMOL chief Howard Webb explaining each decision in person, changed how the public engages with error. What used to be a scandal without a solution became a file that can be dissected: who said what, at which second, and at which step the process broke.

When the cathedral falls silent, only the laws speak. But the laws can only speak when someone agrees to switch on the microphone. This is where many leagues still lag. Releasing audio is an act of surrendering power, and not every refereeing body is willing.

The remaining problem sits in measurement. In England, an independent panel of former players and managers reviews contentious incidents after each matchweek and publishes a verdict on whether officials were right. The mechanism has value and one fatal weakness: it is staffed by people with near-identical expertise, trained in the same system, and therefore likely to share the same blind spots. A homogeneous panel is not a check. A real check requires someone who disagrees.

The counter-intuitive angle: slowness is the last line of defence

Here I will say what many colleagues would rather not hear.

The thing the public hates most about VAR — its slowness — is the only part of the modern system protecting spectators from a verdict written in advance. A match stopped for ninety seconds to draw a line is a match robbed of ninety seconds of joy. But those ninety seconds are also ninety seconds in which somebody has a chance to realise they are wrong.

Haste does not save time. It merely moves time from checking to repairing. A mistake fixed in ninety seconds costs ninety seconds. A mistake fixed three days later costs three days, plus an apology, plus a crack in public trust that never fully closes. That is the arithmetic league administrators still refuse to do properly.

And deeper still, what actually damages football is not wrong decisions. What damages football is wrong decisions delivered in a tone of absolute certainty by people who do not have enough data to be certain. Every analysis published three minutes after the final whistle, declaring definitively that this was or was not a penalty, contributes more to the noise than the original error did.

That night with the empty file taught me this: an empty file tells me more than a file full of assumptions. The empty one is real data. The full one is fabricated data. In a trade where everyone wants a verdict, the person willing to say "I don't know yet" is doing the hardest and most valuable work.

I have been criticised for writing too fast, too coldly, too technically. In July 2026, after the Euro 2026 final at Wembley, I published an analysis of why goalkeeper Gianluigi Donnarumma read Bukayo Saka's penalty in the shootout. I described Saka's five-step run-up, the slight lean of his centre of gravity to the left, and how that gave Donnarumma grounds to dive right. It went live ten minutes after the final whistle. Saka's family and many Arsenal supporters called me heartless, bookish, blind to the pain of a nineteen-year-old. I published a correction and admitted that a purely data-driven lens had ignored psychology.

Since then, every piece I write carries a dedicated passage on psychological context, even when the piece is only about the laws. I give around twenty per cent of the word count to pressure, expectation and the media's effect on players. The hardness of the data is wrapped in a layer of empathy, and the result is accurate without being bloodless.

What should happen next

If I had to propose three changes to how football handles error and how media reports it, I would start with very small things.

First, every league using VAR should publish the full audio of the exchange within twenty-four hours, unedited and unselected. Selection turns transparency into public relations.

Second, a minimum data standard for refereeing analysis should be set and published: a still frame, a timestamp, the article of law cited, and the name of the data source. Anyone without all four should not be passing judgement yet.

Third, clubs should appoint a single official responsible for refereeing matters, whose job is to file an evidence-backed complaint with video within forty-eight hours, rather than leaving a head coach to answer questions two minutes after the whistle in a state of high emotion.

None of those three changes makes football more beautiful. They only make it more honest. And in a sport where billions of people place their faith in a line drawn on a screen, honesty is worth more than any moment of brilliance.

People hate VAR because it is slow. I value it because it refuses to hurry. But are we using that slowness to check more carefully — or merely to endure for longer a mistake that should have been caught in the very first frame?