Empty Data, Full Report: The Quiet Failure of AI Pipelines in Football Coverage

Over the past few years, a large share of football coverage has begun passing through a three-stage machine process. The first stage reads an article and extracts its facts — which team, which player, which number, which quote, which date. The second stage applies an analytical framework to those facts: tactics, finances, the results cycle, league structure, governance, the dressing room, risk, media narrative and industry transmission. The third stage turns all of it into language a reader can consume. The entire structure remains honest only as long as the first stage keeps working. And that is exactly where the biggest failure happens — quietly, without any noise.

The basis of this report is an analytical document that openly declares it received no material at all. Its title field reads "not available", its source field reads "not available", and its list of information points is entirely empty. And yet look at its shape: nine dimensions, tables in every dimension, rows in every table, and at the end a summary, a risk matrix, a signal-tracking table and a glossary. This is the contradiction football media needs to take seriously: a document completely full of paper and completely empty of substance.

A polished report with nothing behind it

In any newsroom, the first reaction to such a document would be that it looks professional — because it contains everything that looks professional. But read it cell by cell and the same sentence repeats everywhere: "not applicable due to insufficient information." That sentence is itself a mark of honesty, but when it appears forty times, it becomes obvious that the problem is not the analyst but the system upstream.

This kind of failure has a clear fingerprint in football. When an article is not read properly, the fields do not all go blank at once — some usually survive. The outlet's name survives, the date survives, the sport survives. Here, the only surviving field is the sport label itself: football. The title is gone, the source is gone, the date is gone, the teams are gone, the players are gone, the claims are gone. When everything disappears at once in the same way, the likeliest cause is not that the article contained nothing; the likeliest cause is that the article was never read at all.

That distinction matters, because the two situations have entirely different remedies. If the article genuinely contained nothing — say, an official channel issued a one-line statement — then the correct analytical response is to write very little. But if the material existed and simply was not read, the problem is technical: a paywall, a JavaScript-rendered page, broken encoding, a changed page structure, or a parser that no longer recognises the website.

Diagnosing the failure: when everything is blank in the same way

At a professional level there is a straightforward test. After any content-extraction system runs, someone should count how many mandatory fields were actually populated. If the fill rate drops below sixty per cent, that payload should not move forward. This is a simple rule, but most newsrooms fail to enforce it, because the system does not look like it has failed — it just looks empty, and empty is less frightening than failed.

A second test is entity counting. If the sport label is football and the number of extracted entities is zero — no club, no player, no coach, no competition — that is not content poverty, it is a parser fault. No football story, however small, is without at least one club name. Zero entities is a signal that cannot be ignored.

A third test concerns time. Football news decays fast; most of it is worth little after twenty-four to seventy-two hours. If a payload carries no publication date at all, no time-based analysis is possible. This sounds trivial, but it has a direct consequence: you cannot determine which part of the season the story belongs to — the opening sprint, the festive pile-up, the January window, or the run-in. Without that, any analysis of pressure or fatigue hangs in the air.

"Not applicable" never means "no risk"

A methodological point here extends well beyond this single document. When every cell of a risk matrix reads "not applicable", a reader's brain unconsciously reads "no problem". That is a human tendency, and it is the most dangerous part of the whole affair. Absence of evidence is not absence of risk.

Football offers daily examples. Until an official financial figure appears, no one can say a club is compliant with financial rules — only that we do not know. Without revenue, wage bill, debt or amortisation schedules, the conclusion is "indeterminate", not "low risk". And that distinction must be stated explicitly for the reader, because once "not applicable" is printed, nobody revisits it.

The same applies to tactical analysis. Modern football analysis rests on data — expected goals, which estimates the probability that a given shot becomes a goal; expected goals against, the same measure applied to chances conceded; and passes allowed per defensive action, which captures pressing intensity. Without them you cannot tell whether a good result came from a strong process or from luck and an outstanding goalkeeper. That gap between process and outcome is the most valuable part of contemporary analysis. A report where the gap does not even exist is not analysis; it is a spreadsheet.

The temptation of the template

The question is why this happens. The answer is partly economic. Demand for football coverage is insatiable — every match, every transfer, every press conference, every day. A template is an attractive tool for meeting that demand, because it buys speed. When you already have a nine-dimension framework, you do not have to think from scratch each time; you simply fill the cells.

But templates come with a condition that is easy to forget: a template organises thought, it does not generate thought. If the material exists, nine dimensions make it quickly professional. If the material does not exist, those same nine dimensions create an illusion — the illusion that because something has been written under every heading, some analysis has taken place. That illusion is the most expensive thing in an editorial process, because every later stage treats the document as finished and moves on.

There is another aspect that is often missed: accountability. In a conventional analysis, a mistake can be traced to a person's name. In a multi-stage machine process, the error occurs somewhere in the middle and attaches to no one. The first stage says it passed on what it received; the second says it received nothing so wrote nothing; the third says it forwarded what was written. Every link is technically correct, and the reader ends up with a report that answers no question at all.

Source tiering: the real currency of the rumour market

Much of football journalism runs on transfer rumours, and there the most valuable thing is not money but sourcing. A veteran correspondent's claim and a rumour launched from a social-media post are not the same, however similar they look. Professional analysis therefore grades the source: who made the claim, which outlet they belong to, what their track record is, and what they stand to gain. Now consider what happens when the source field is empty. The entire apparatus for weighing credibility becomes useless, because there is nothing to weigh. The risk is not that you will believe a bad rumour; the risk is that you will publish it in neutral language without knowing who started it — and neutral language manufactures credibility on its own.

The same logic applies to sensitive areas. Financial breaches, points deductions, transfer bans, allegations of improper approaches — commenting on any of these requires knowing which authority has jurisdiction and what conduct is alleged. Without that, writing "low risk" is the gravest error. Likewise, pressure on a manager, factions in a dressing room, the standing of a captain: these emerge from statements, reports and incidents, not from a table.

When failure propagates silently

The greatest danger is not a single flawed report but where it travels next. Suppose such an empty document reaches the third stage. An article is produced. The article is full of cautious language — "assessment is not currently possible." But caution does not travel on social media; only the headline does. And if the headline says "questions over club's situation", then within hours that phrase becomes a fact requiring no source at all.

This is not new in football — it is an old process in a new form. Rumours once moved from a café to a newspaper; now they move through a pipeline with no human at any step. Speed has increased, and with speed has come the illusion that speed equals accuracy. In reality the two are unrelated. An empty pipeline can run as fast as a full one.

Minimum conditions for newsrooms

If the solution is not to change the framework, where is it? The most direct answer: impose strict conditions on the input. Before any second stage begins, it should be established that the first stage delivered at least the following — the article's title, the publisher's name, the publication date; at least five discrete information points, each with its own source attribution; at least one named club and one named competition; at least one named individual — player, coach or executive — with role and club; and at least one quantitative anchor: a transfer fee, a wage, a contract length, a league position, a points total, or a data point.

One subtlety must be added. Where the source contains no number at all, the system should state explicitly: "no data in source." That sentence and a blank cell are not the same thing. The first says the reporting contained no figure; the second may mean the figure was there and was not extracted. Distinguishing the two is the mark of a trustworthy pipeline.

A further habit is needed: sample testing. Any new or changed system should periodically be run on an article whose correct conclusions are already known. If the system cannot extract anything from a known-good article, the fault lies with the system, not the article. This simple practice would prevent a great many empty reports.

How readers can spot a hollow report

Readers have simple defences. First, count the names. If the entire report contains no club, no player and no coach by name, it is not analysis. Second, look for numbers. Where a figure should be and only adjectives appear, you know the material is thin. Third, ask for the source: who said it, when, and on what basis. Fourth and most important, do not mistake cautious language for a conclusion. "Assessment is not currently possible" is an honest sentence, but it is not a story.

It is also worth remembering that an incomplete report is not in itself a failing. Sometimes the correct journalism is precisely to establish that nothing is yet known. The problem begins when that is concealed — when inferences are inserted to fill the emptiness and given the appearance of data.

Conclusion: verification before speed

The role of data and analysis in the football industry keeps growing, and that is a good thing. Data-driven decisions are now common inside clubs, and media is moving in the same direction. But that shift brings a responsibility that technical enthusiasm tends to forget: the value of an analysis is set not by its framework but by its evidence.

The document that prompted this discussion is, in its own way, a good example — it did not lie. It stated plainly, everywhere, that it had nothing. The danger is not in the document but in the moment a reader sees nine sections and assumes nine findings. That moment will determine whether the next decade of football coverage runs on information or on the appearance of information.