Nine Pillars, Nine N/As: When the Football Data Chain Lost Its Own Record

The document that landed on my desk on Tuesday morning had no headline on its first page. No author's name. No publication, no date, no category. In all four metadata cells the same word was typed: not applicable. Below it, nine analytical pillars — tactics, club finance, results, league landscape, governance, dressing room, risk, media climate, industry transmission. Under every one of them the same sentence: insufficient information. Every 'evidence' line pointed at an empty cell.

The first document was boring. That was the point.

Because the document did not lie. It did not guess either. It stood still and said: I do not know. In the current market of football journalism that is the rarest sentence. Where every transfer rumour is welded to a 'reliable source', where a one-word verdict is printed after every match, a system that stops and admits a null is not a failure — it is a form of testimony.

In the economy football now lives inside, an empty cell is not an innocent thing. Betting feeds pull xG and possession figures before dawn. Transfer tickers update fees and contract lengths every hour. Club press offices distribute fact sheets the day before a match. Compliance desks reconcile the wage-to-revenue ratio once a year. All four layers eat empty cells, and none can digest them — each has to fill the gap with something. The only question is who fills it, and with what.

I arrived at this question holding a filed account, not a match report. In March 2026 I walked out of a nine-year compliance job at a Manchester accountancy firm and self-published a 14-page financial autopsy of the Championship's 24 clubs, built entirely from accounts filed at Companies House. Three clubs had wage bills above turnover; Birmingham City's stood at 129 per cent against £29.4m of declared revenue. I had no editor, no accreditation, and 340 email subscribers. Two national outlets cited the piece within a week. By December the subscriber list stood at 6,000, and I had not written a match report all year.

The following year I spent 32 days in Russia and filed zero match reports. Instead I scraped FIFA's official hospitality resale listings every morning, logged 41,700 seats offered above face value, and traced the three largest resellers to a single registered address in Nicosia. FIFA's own ticketing terms prohibited resale above face value. I published 9,000 words and attached the raw spreadsheet as an appendix, not a footnote.

From those two episodes a rule has held, and it is the standard by which today's empty document must be judged. A number with a date and a filing address can be checked by a stranger; a quote cannot. The quote is the technology that has been used to fill empty cells for two centuries.

In twenty-nine years at the edge of the pitch, in the press box and in front of a screen, I have seen one thing repeatedly: the crowd forgets who won, the scoreboard never does. The same rule governs the information chain. A system that keeps no record of itself survives on someone else's record.

This week's document was divided into nine pillars, but all nine rested on a single emptiness. The tactics pillar had no formation, no style, no coaching decision — because the input contained no match. The finance pillar had no wage bill, no debt, no broadcast revenue. The results pillar had a sample of zero matches. The governance pillar named no governing body — even though FIFA, UEFA, national associations and leagues impose materially different obligations.

Here the first diagnosis is not football but plumbing. Four metadata cells — headline, source, type, date — all returned 'not applicable' at once. Even a content-free article yields a headline. Four different kinds of cell failing in the same way means the article was not empty; the article never arrived. Somewhere between fetch and parse the document died, and no log recorded the death.

Why that death matters is explained by a rule I keep: every document must be paired with one concrete consequence. A document does not speak alone; it speaks through a tribunal date, a wage figure, the name of someone absent from the pitch.

[Table: tactics — zero match sample; finance — no wage or revenue figure; results — zero points, zero fixtures; transfer — no fee, length or age; governance — no named regulator, each with a real-world consequence.]

Let us stand at five rows. An empty tactics row does not mean the team has no style; it means nobody supplied match data. The model does not stay silent — it inserts last season's average, and within five matchdays that average circulates as fact in three reports. The finance row is more dangerous: a compliance checklist is a template that demands a risk level in its final column. A template that demands a verdict will get one, and an FFP verdict without data means a baseless allegation against a real club. The results row, with zero points and zero fixtures, produces not a table but a mood. The transfer row lacks fee, length and the age curve, which makes resale recovery uncomputable. The governance row is the most sensitive, because there the price of error is not money but reputation.

An empty cell never stays empty. Whoever touches it first writes its content.

An absence of information is not a neutral condition; it is itself a claim — and whoever fills that claim owns it.

Spreadsheets do not lie. They wait for the right question. Today's blank page had been waiting a year for one question: did the document ever arrive? Nobody asked it, because nobody kept a log.

The remedy I want is not a bigger model. I want an ingestion log: when the request went, when the answer came, how many bytes came, in which encoding. I want a cell called 'out of scope', where an empty document can be closed instead of forwarded. And I want every data vendor to publish its null rate — what share of records it could not populate itself. None of it is exciting. None of it is a headline.

The easy critique is that machine learning hallucinates. It does. But the original technology for filling empty cells is not a machine; it is a human sentence — 'sources say'. Before 2026 that was the technology in my own copy, because where there was no document there was a sentence. I abandoned the quote as a unit of evidence for exactly this reason.

The more uncomfortable point is that this empty document behaved correctly. It refused to guess. The football industry does not reward that behaviour, because a guess has a deadline and honesty does not. Betting models do not publish their null rates; transfer exclusives pass through no grading; fan channels must print something daily. A system that says plainly 'I do not know' looks like a failure and is in fact the only one that cannot be held to account tomorrow.

This is where the critics miss the point. They think the problem is model intelligence, so the solution is a bigger model. The problem sits beneath the model, in the drain, where no log exists. And it is structural: the handful of data vendors that feed every outlet will fail everywhere at once. Nobody notices, because every outlet prints the same blank in a different font. Just as every academy now produces the same inverted winger and nobody hunts for the old touchline-hugger, the information economy runs on one template, one gap, one silence. When variety ends, failure becomes homogeneous too.

For every football data report you read in the coming months, ask one question: which cell was empty, and who filled it? Before signing with a vendor, demand its null rate. When an extraction fails, mark it out of scope and close it — do not pass it downstream. Because a forwarded empty document does not stay empty. It fills with someone's imagination, and that imagination becomes a file.

I do not chase villains. I chase inconsistencies. Today's inconsistency is this: a system knew that it did not know, and nobody kept the record. When the next transfer exclusive lands, ask which cell was empty on Tuesday morning.