The Empty Payload: When Football Analysis Fills Tables Without Evidence

At 2:47 in the morning I opened a nine-dimension analysis report. Table after table — tactics, club finance and the transfer market, results and the public-opinion cycle, league landscape and team positioning, rules and governance, management and dressing room, risk profile, media narrative, industry transmission. Every cell returned the same sentence: insufficient information. Not one name, not one figure, not one date across all nine dimensions. The only true thing in the document was a single word — domain label: football.

What caught my eye harder was a field near the end. Under 'additional notes' the template's own instruction was still standing, untouched: "identify from the information points above." The step whose job was to read the article and pull out facts had never run. Only the imprint of its instruction came back. Technologists call it template echo. Journalism has an older name for it: walking onto the pitch with an empty notebook and coming back with a full pen.

I know this scene. Not from a screen — from the press box.

June 21, 2026, Goodison Park. Everton against Liverpool, the first Merseyside derby behind closed doors. 0-0. When the whistle went I had no goals, no curled free kick, no late drama. I had silence — the sound of the ball in an empty ground, a row of vacant seats, twelve supporters waiting outside. That piece ran three thousand words and its entire foundation was an absence.

That night I learned that emptiness can be a finding. Writing 'nothing happened' is not a failure; often it is the only true sentence available. The document on my screen had done exactly that — nine times it said: I do not know.

It is worth being precise about what happened. Modern football analysis often runs in two stages. The first breaks an article down into information points, entities, the author's stance and purpose, time sensitivity, and source quality. The second applies the analytical framework to those points. Here the first stage came back empty-handed — zero information points, zero summary, zero source. Only the domain classifier worked: the piece was about football.

Inside that asymmetry sits a telling clue. The system knew it was football, but never read the article. The classifier succeeded; the body extractor failed. Anyone who has worked online knows how this happens — a paywall, a JavaScript-rendered page, a retrieval path that broke before the text was ever lifted. The fault was in fetching, not in analysis.

Why should a football reader care? Because much of the machinery now producing transfer rumours, tactical explanations and takeover coverage runs on the same two-stage architecture. The risk of confident sentences being generated from an empty evidence base is not science fiction; it is a plausible output of an everyday production process.

I know these failure patterns, because they are old diseases of football journalism — only faster now.

The most dangerous sentence in the transfer market is 'they overpaid'. That judgment needs at least three things: the actual fee, recent prices for comparable players, and the contract structure — how much is fixed, how much contingent, over how many years it is amortised. Without them, 'panic premium' is not analysis but a feeling wearing a suit. When not a single figure exists for net debt, wage bill or a financial rule's threshold, saying 'the club is under financial stress' means writing a guess in the register of news.

Statistics fall into the same trap. Passes allowed per defensive action (PPDA) measures pressing intensity — lower values mean more aggressive pressure. But a PPDA figure with no source, no time window and no opponent stops being analysis and becomes decoration. The same holds for xG; an unsourced xG can prove 'bad luck', never 'good structure'.

Dressing-room stories are the clearest case. 'The dressing room is lost' requires an interview, a public remark, or a verifiable training-ground incident. Without any of those, the sentence is a rumour with correct grammar. In my experience such sentences are born most often in the forty-eight hours after a defeat, when nobody has said anything and a explanation is still due.

Then there is the wonderkid label. Ten bright matches make 'the new Messi' easy to write; the harder work is checking the base rate — how many teenagers have hit that streak, and how many stopped exactly there. Praise written without a sample-size check becomes a promise to the reader, and the debt for breaking it lands on the player.

Staleness is the quietest trap of all. A report with no publication date can be undone by a table two weeks old. I have my own memory of this. July 11, 2026, a World Cup semi-final, England against Croatia. Kieran Trippier scored a free kick in the fifth minute; Ivan Perisic equalised on 68; Mario Mandzukic won it on 109. I was in a packed Liverpool pub, and what I had written by the hour mark was a festival report. By full time every sentence of it had turned from true to false — even though it was true when written. Analysis published without a time-sensitivity check travels from truth to falsehood exactly like that, without a single error in it.

And source tier? A report with no named source cannot place a rumour in any tier — elite journalist, major outlet, aggregator, low-quality feed. If the tier cannot be graded, the rumour fails at the first gate of verification. In that state it deserves the lowest possible weight, because the distance between an unattributed claim set and a fact is the widest there is.

Industry transmission stalls the same way. A record fee is not split only between two clubs; through the solidarity mechanism and training compensation, the clubs that raised the player also take a share. Tracing that flow needs the player's youth-club history and the actual fee — both of them. Without them you can still write 'the market has heated up', but that describes weather, not flow.

Club positioning is equally unreadable without a name. Star exporter, star destination, or stepping stone — which rung a club occupies shows up only against squad market value, wage bill and recent finishes. Without the name that comparison is impossible, and club analysis without comparison is just affection.

This is where the central observation sits. A table is not evidence; a table is a shape that evidence is supposed to fill. When evidence is absent the shape still stands, and a standing shape demands to be filled. A nine-dimension template wants nine answers. The analyst sitting before an empty cell faces two paths — leave it empty, or put something in it and call it an estimate. The second path pays far better socially, because a full table looks like work.

That is why the bravest sentence in that report was 'insufficient information'. It is not a failure. It is a finding.

The fix is just as plain. Before the second stage runs, three conditions are enough to check: is there a title, is there a source, and was at least one information point extracted. If all three fail, the document should return to ingestion instead of proceeding to analysis. The cost is near zero, and it prevents the most expensive error of all — a confidently stated wrong name, wrong fee, wrong formation.

Attached to this is the question of genre. We learned the piece was football — but which kind of football? Match report, transfer, financial crisis, governance decision, player profile? That distinction is decisive, because it determines which dimensions apply and which do not. Knowing only 'football' while not knowing which teams played is roughly like knowing a match finished goalless without knowing who was on the pitch.

There is a fine but vital difference here: 'not applicable' is not the same as 'insufficient'. A match preview owing nothing to governance is not at fault; that is a legitimate finding. A question that applies but goes unanswered is a fault. Confusing the two inflates reports with empty scaffolding and buries the real gaps.

One further signal is technical but significant. Template echo may not be a single accident. If other articles processed in the same batch return the same instruction in the notes field, the fault is systemic rather than article-specific. In that case repairing one document is pointless; the whole flow needs auditing. The old newsroom rule holds: when one error is published, ask what else went out the same night.

There is a human side to all this, and I learned it from the stands. I found my first sentence in the Kop — August 2026, the night Liverpool beat Arsenal 4-0 at Anfield, in front of 54,074 voices. The scoreline, I realised, was the least interesting part of the piece. Supporters know when someone is bluffing. However neatly the press-box table is arranged, the people in the Kop can tell from the body language which sentence was written standing on the touchline and which was assembled on a laptop.

Now the uncomfortable part. We usually blame the machine — AI hallucinates, invents, errs with confidence. But this document did the opposite. The restraint required to write 'I do not know' nine times is rare in human writing. The most honest document I read that week was the one that admitted, nine times, that it knew nothing.

So where does the blame actually sit? In the mould. A nine-dimension structure demands nine answers, and most falsehood is born from the pressure to supply them. Give a human writer nine boxes and a blank page and most will fill the boxes, because an empty box looks like professional failure. The machine was faster, but in the end it refused. We do not.

And that is precisely why a null result is unpublishable. Nobody shares a zero in football media. What travels is the confident sentence, not the doubt. Two supporters come back to me — one said, "I'd rather read a wrong analysis than nothing"; the other said, "I stopped reading transfer news, it all feels made up." Both are right, and holding the tension between them is the job — not smoothing it away.

So what can be done going forward? Three small things a football reader can demand outright. Every number should carry its source and its time window — fee, xG, PPDA, wage ratio. Every table should carry a publication date at the top, because an old table cannot be the basis of a new decision. And the right to write 'insufficient information' should survive in the journalist's hands — with editors, and with readers.

The last question is easy and the answer is hard: if the blank cell were allowed to stay blank, what would the football page look like today?

At 2:47 in the morning one word was lit on the screen. Football. It was the only true thing in that document, and it was not enough. But the truth is, it was not false either.