When the Audit Returns Empty: The Real Cost of 'Not Assessed' in Football's Information Economy

Hook

On 17 June 2026 the ball crossed the goal line at Villa Park and nothing appeared on the screen. In that Premier League match against Sheffield United, Hawk-Eye sent no message — not goal, not no-goal. The technology stayed silent, so referee Michael Oliver's on-field decision was allowed to stand: no goal. The match finished 0-0, and without saying a word football established a principle it has never since abandoned — when the technology is silent, the event did not happen unless a human saw it happen.

That same week, on my desk in Manchester, there was another silence. A form was open on screen with every field populated and every field empty. Title: N/A. Source: N/A. Article type: Unclassified. One-sentence summary: blank. Information points: zero. Author stance: N/A. Purpose: N/A. Time sensitivity: "not assessed".

This was not a trivial glitch. It is the most common and least-documented offence in football's information economy: dressing an empty result as a filled one. A document that is perfectly valid in schema and perfectly hollow in substance. When I reopened my own audit file 40 minutes later, there were no facts in it — only a completely filled-out, completely blank sheet.

And that is where the real football story begins. Because the Hawk-Eye silence and the form's silence ask the same question: when the evidence is absent, what does the institution write — "not assessed", or zero? The first is honest. The second sells.

Context: what the law says, and what the practice does

Start with the statute, because in football even a null has a legal status. Law 5 of the IFAB Laws of the Game gives the referee full authority over the match, and the VAR protocol's founding principle is that the on-field decision will not be changed unless a "clear and obvious error" is established. In other words: in a state of doubt, the decision does not change; it stands. Law 5 and the VAR protocol together produce a system in which "we are not certain" translates into "the original call applies".

It is no accident that the same logic governed the Hawk-Eye failure. The system provider later acknowledged publicly that it had not tracked the ball crossing the line, and the remarkable detail was that the first restart match of the 2026-20 season had seven cameras deployed, none of which covered the angle where the incident occurred. The failure was blind, and the remedy for a blind failure was already in the lawbook: the human decision.

On time, the law is even clearer. Under Law 7 the referee adds time at the end of each half for all playing time lost, and from 2026-23 IFAB pushed for a more accurate calculation — goal celebrations, injuries, substitutions, VAR checks and deliberate time-wasting all counted. The result was visible in Qatar: on 21 November 2026, England versus Iran produced 27 minutes of added time across the two halves, 14 in the first and 13 in the second. That was a record-scale figure, and it was an administrative decision, not weather. Whoever controls the count controls the outcome, and almost nobody is checking the count.

In financial regulation the same question gets more expensive. In February 2026 the Premier League charged Manchester City with 115 breaches of financial rules. On 17 November 2026 Everton received a 10-point deduction for breaching Profit and Sustainability Rules, reduced to six points on appeal in February 2026, with a further two-point deduction in April 2026 for a second breach. In March 2026 Nottingham Forest were docked four points. In Italy, Juventus were handed a 15-point deduction in January 2026, revised to 10 points in May 2026. The whole record says one thing: the rule is not what is written in the book; the rule is what survives enforcement.

Back home, dates matter even more. In August 2026 FIFA suspended the All India Football Federation over third-party interference; the suspension was lifted on 26 August 2026. The Indian Super League ran closed from 2026 until promotion arrived in 2026-24, and Mohammedan Sporting played the ISL in 2026-25. An article that still opens on the glow of the 2026 ISL launch and mentions promotion without naming a season is not selling facts. It is selling an archive.

What changed — a four-line ledger

| Area | Before | Now | In force from | |---|---|---|---| | VAR | Unclear threshold for overturning | Strict "clear and obvious error" standard | IFAB VAR protocol, effective 2026 | | Added time | Estimate and tradition | Precise accounting for goals, injuries, VAR, substitutions | 2026-23 guidance | | Handball | Blurred shoulder-to-elbow boundary | Arm begins at the bottom of the armpit | 2026-22 Laws of the Game | | Indian transfers | Loose force-majeure interpretation | Written approval mandatory | AIFF circular, 12 September 2026 |

Core: three kinds of null, and what each costs

Football's "we don't know" arrives in three distinct forms, and each has a different price. The first is the referee's null — he sees the incident, doubts it, and lets play continue. The second is the compliance null — a club filing in which every field reads "not applicable". The third is the information null — a report, a transfer story or a claim that circulates with no source behind it. All three share one property: a null is never neutral. It, too, has a beneficiary.

Take the referee's null first. When a referee declines to give a penalty for a fall in the box, commentary says "the referee did nothing". That is a false description. Under Law 12, doing nothing is a positive finding — that the contact did not reach the threshold of a foul. A non-decision is still a decision, and it still has a beneficiary. When I was coding at the 2026 FIFA U-17 World Cup, this was the most uncomfortable distinction: in 18 of 52 matches, restart management was logged as "undetermined" — no foul, no play-on, just a blank field. And a blank field did not mean nothing happened. It meant nobody wrote it down.

Fifty-two matches, and the minutes nobody counted

Across the tournament staged in India from 6 to 28 October 2026, I coded all 52 matches for fouls, cards and added time. The finding: referees in India averaged 4.1 minutes of stoppage time; the equivalent average at the 2026 UEFA U-17 Championship was 2.9 minutes. A gap of 1.2 minutes. Small? Across 52 matches, that is roughly 62 minutes of football that existed in one competition and not in the other.

I counted the minutes nobody counted, and the stoppage-time gap appeared. But I refused to stop there, because the next question was compulsory: was this a refereeing habit, or a consequence of the matches themselves? So I split a second time — by referee, by score state, by team strength. When I isolated score state, a large share of the gap disappeared: tight matches naturally carry more added time, and the U-17 World Cup knockout rounds contained more of them than the UEFA equivalent.

What survived was small, but it was not zero. Inconsistency in restart management survived, and it was a tournament-level problem, not a referee-level one. That distinction matters, and it is precisely where most analysis goes wrong. When you split a second time and the effect shrinks, you have to say so in the same paragraph. I said so. The report ran to 2,000 words with a methodology appendix and was shared 4,000 times. A fan-site editor asked me to drop the tables and use players' initials, citing "reader bias". I refused both.

The bio-bubble looked neutral

The 2026-21 ISL was played entirely inside a Goa bio-bubble. I was then a junior rules analyst at a Manchester sports data firm, assigned to code 40 of the 115 matches. Home teams received 0.09 penalties per game, down from 0.21 in 2026-20; away teams received 0.18.

The bio-bubble looked neutral until I separated home penalties from away penalties. And that was the trap: the moment the gap appears, the easy conclusion is refereeing bias. I did not take it. I split again, by penalty-box entries. The picture changed. Home box entries had also fallen in the bubble, because the normal sources of home advantage — the absence of travel fatigue, crowd pressure, a familiar pitch — did not exist there. What survived was a very small difference in penalty rate that could not be reliably separated from ordinary variation on a 40-match sample.

That season I logged 112 second-half stoppages for "cooling breaks". One hundred and twelve. That is a number that never appeared on a television graphic, never entered a match report, and was never recorded in any regulatory document. But those minutes existed inside matches, and they had a distribution: who was benefiting from them, and who was not.

That work produced something I still value more: a 12-point incident-report template that eight ISL clubs adopted for referee feedback. It had one condition — on every disputed call, the official had to cite the IFAB law number. Instead of writing "the referee got it wrong", a club had to write: "Law 12, contact inside the penalty area, insufficient intensity, no card." That was the moment rule-writing began changing behaviour rather than merely explaining it. I stopped accepting anonymous referee quotes from that point. If you cannot name the law, your complaint is an opinion, not a document.

Fifty-four transfers, three holes, one circular

In August 2026 FIFA suspended the AIFF. Writing from Manchester, the first job was a 1,200-word FAQ clarifying the two-week timeline — what stops, what does not, which registration stalls, which does not. Then I audited the 54 transfers of the 2026-23 window against the salary cap.

Three clubs had registered players under force-majeure clauses without AIFF approval. Two had misclassified agent fees. On 12 September 2026 the AIFF issued a circular requiring written approval for such clauses.

That taught me something I will state as plainly as I can: the distance between the rule on paper and the rule in force is where football's real politics happens. The book said the national association's obligation was binding; in practice three clubs stepped around it and no sanction followed — only a circular that closed the door going forward. Whoever registered a player inside that window took advantage of a rule that had not yet been written. That is not corruption. That is pure timing, and timing belongs to nobody unless you date it.

Rulebook versus enforced rule: the handball lesson

Handball is the best example available. In 2026-20 IFAB tried to clarify the boundary of the arm, and penalties flooded the season. In 2026-22 the line was pulled back further — the arm begins at the bottom of the armpit. And in 2026-21 it was clarified that a goal would not be ruled out if the ball accidentally struck the arm of a teammate in the build-up; only the scorer's own accidental handball immediately before a goal counts.

All of that looks like rulebook progress. But my question is always the same: how many times was this clause invoked, how many times waived, and who was sanctioned? On handball, the answer is that no central enforcement log exists. How many decisions were overturned on the armpit line after 2026-22 is not recorded in any official, public, searchable register. Where no log exists, the correct professional answer is to write that the log does not exist.

The same applies to financial rules. Manchester City's 115 charges were filed in February 2026, and years later their final disposition has not been publicly completed. That does not mean the charges are false, and it does not mean they are true. It means we have a process whose duration matters as much as its outcome — and during that duration, teams are competing in a competition whose financial basis is undetermined. Undetermined is also a state, and that state also benefits somebody.

The arithmetic of a small sample

Here I have to play my least favourite role: the person who applies arithmetic to enthusiasm.

Take the bubble numbers. On a 40-match sample, a rate of 0.21 penalties per game predicts roughly 8.4 home penalties. The observed figure was about 3.6. The gap looks large, and it is — but on 40 matches it can still be small enough that ordinary variation cannot be ruled out. On 52 matches, by contrast, the 4.1 versus 2.9 minutes gap was more solid, because it was a variable present in every match rather than in every tournament.

| Measure | Sample (n) | Period | What survived the second split | |---|---|---|---| | Stoppage-time gap | 52 matches | 2026 World Cup vs UEFA U-17 | Partial — shrank when controlled for score state | | Restart inconsistency | 18 of 52 | 2026 World Cup | Survived — tournament-level, not referee-level | | Home penalty rate | 40 matches | ISL 2026-21 bubble | Near zero — after box entries | | Cooling-break stoppages | 112 incidents | ISL 2026-21 | No split possible — a count, not a conclusion | | Force-majeure breaches | 3 of 54 | 2026-23 transfer window | Survived — documentary evidence existed |

The final column is the most important and the least published, because readers want thrill and "the effect shrank" is not thrilling. But if your conclusion does not survive the second split, it was never a conclusion — it was just the first split, looking clean.

The contrarian angle: the market buys a confident number, not an honest blank

Now the part that makes me uncomfortable.

The way I filed that empty form was professionally correct — every field marked "insufficient information", no guesses, no invented narrative. But if I sent that same file to an agent, a club's recruitment department or a broadcaster's data desk, what happens? The answer is obvious: they send it back. In scouting, the person who says "I don't know" loses the contract, and the model that returns 0.62 keeps it — regardless of what produced the 0.62.

The null is the most expensive sentence in football's information economy. Which is exactly why it is written so rarely.

But here I have to argue against myself, because there are two kinds of null and conflating them is dangerous.

The first is an editorial null — evidence was gathered, checked, and the result was that no pattern exists. The near-zero home-penalty effect in the bubble belongs here. That is a finding, and it should be published. The second is a technical null — the data never arrived. That empty form belongs here. Writing "no evidence was found in this case" about it would be a lie, because the evidence was never sought successfully; the pipeline simply broke.

That form had three possible causes, and I will give them with their confidence levels, because labelling an estimate as an estimate is the whole of the audit discipline. First — medium confidence — the extraction pipeline failed: the page was JavaScript-rendered, paywalled or bot-walled, and the scraper wrote a schema-valid empty object instead of raising an error. Second — low to medium — the source was not an article at all (video, image, podcast). Third — low — the source was genuinely hollow: a headline and no body.

The first is the most probable, and it makes a testable prediction: if the scraper parsed metadata but not body text, the source's metadata should still hold the outlet name and headline. That takes two minutes to check. And it is my professional motto in one line — every rule leaves a trail; my job is to timestamp it.

I also accept that my professional habit carries a risk. If you stop everything with "insufficient information", you build a culture in which nobody ever says anything. That is also a failure. The remedy is not to hide the null but to label it: "this is a null, this is its basis, and this is the cost of re-running it." That is the difference between an honest organisation and merely a cautious one.

The second contrarian angle is the one I opened with: the referee's non-decision. Fans read it as "no opinion". In law it is a positive assertion. And that asymmetry is the blind spot in football's entire accountability architecture — we review decisions, not non-decisions. I have never heard a broadcast say "the referee assessed this contact as below the threshold of a foul, and therefore did not give a penalty", even though that is precisely the information a review requires.

One more thing belongs here: the exile's archive problem. I write about Indian football from a Manchester desk, and I have to admit that the older material looks cleaner and more trustworthy from a distance, because it was better written. But if I write that the ISL has promotion, I must name the season — 2026-24. And if I mention Mohammedan Sporting, I must say they were in the ISL in 2026-25. A structural fact without a season is a claim, not a fact. That rule is strictest on myself.

Takeaway: the count nobody keeps is the count that matters

Next season, when a club says its financial model is sustainable, my first question will not be its table position. It will be how many seasons sit behind the claim and which seasons were excluded. When someone says a referee changed the course of a match, I will ask how many matches, which referees, and what survived the second split. And when someone says the data shows nothing, I will ask how the data arrived — because often the data is not silent; its mouth has been shut.

I have one reform proposal, and it has three points.

First: publish n beside every percentage, and publish the excluded seasons beside every n. If a claim rests on 40 matches, write 40 — not "an analysis of multiple matches".

Second: pair every clause with its enforcement record — how often invoked, how often waived, who was sanctioned. Where no log exists, state plainly that no log exists. That is not weak journalism. That is the only journalism that can be checked.

Third: label the null and gate it from flowing downstream. The day a newsroom adopts the rule that a report with zero information points does not advance, the most expensive error in the trade — the confident guess — stops by itself.

The question is not whether we have the data. The question is what we write when we do not. And the answer will decide whether football analysis over the next five years is an audit or an advertisement. The referee's eye is just an audit that never blinks — and that is the only eye I am willing to write with.