When a Football Analysis Reaches the Desk With Nine Empty Fields

A second-stage football analysis has just landed on the editorial desk. It is long, properly templated, complete with nine professional sections: tactics and technique, club finance and the transfer market, results and public-opinion cycles, league landscape and team positioning, rules and compliance, management and dressing room, risk profile, media narrative and expectations, and industry transmission chains. Nine sections. Nine fields. And in all nine, the same line repeats: insufficient information.

The only thing that survived the entire pipeline was a single label — football.

No title. No source. No one-sentence summary. No author stance. No article purpose. No information points. No resolvable entities, alongside an instruction to "identify them from the information points above" when there was nothing above to identify.

In sports journalism we are used to thin articles. A match with no data, a player nobody could interview, a club that publishes no figures. But thin and empty are different things. A thin article can still be written, because an event still exists to tell. An empty article has no event at all — only a frame standing there, waiting.

What matters here is how carefully that frame was built. It is not a scrap of note paper. It is the output of a process that ran to completion, returned, and packaged itself. Technically, the pipeline reported success. It did not raise an error. It simply brought back no content.

Operators of data systems call this a silent empty. The system returns a success code, but the body is blank. In a content production line, that is the most dangerous class of failure, because it does not incriminate itself.

Three explanations are plausible, in order of likelihood.

First, the text extraction step failed. The source text never reached the parser — pipeline error, paywall, blocked page. But the topic-labelling step, which runs on metadata rather than content, completed and stamped the word football on it.

Second, the source document was never an article. An image. An embedded video. A live data widget. An ad-serving stub. These can carry a football label without ever having a body text to extract.

Third, the source was an automated feed — a scoreboard or a market-data ticker — where the concept of an author stance does not exist by design.

The distinguishing signal is this: had this been a real but shallow article, at least one of the three sections on results, transfers or governance would contain data. All three are empty. Add that author stance and article purpose — two fields any competent extractor fills for almost any editorial text — are also blank, and the conclusion is nearly certain: there was no editorial text to extract in the first place.

The key point: an analysis with nine complete sections but not a single information point is not shallow analysis — it is analysis with no raw material, and any prose filled into it would be fiction.

In other words, what sits on the desk is not a poor article. It is an article that never existed.

For a machine-run newsroom, this is a fork in the road. One path fills the frame with sentences that sound entirely reasonable: a familiar formation, a round transfer fee, a name familiar enough. Readers would not notice. The machine would publish.

The other path is refusal.

In more than fifty years at the end of the news production line, across four major technological shifts, I have come to one fairly simple realisation: the greatest risk in this trade has never been being unable to write. The greatest risk is being able to write — fluently — about something that never happened.

Our trade lives on detail. No detail, no article. A match nobody watched again has nothing to say about the stands. A transfer with no figure has nothing to weigh. A player not named has no age, no contract, no injury, nothing to love or to doubt.

Seen from another angle, that empty analysis is doing its job. It refuses to guess. It marks every place it does not know. It invents no person, no club, no match. In an era when a machine can produce a 1,500-word match commentary on a game that never took place, a system that stops and says "I have nothing" is behaving honourably.

But it is only honourable if it really stops. If that empty document is pushed straight into the publishing queue, it becomes raw material for another writer — human or machine — and the error replicates. Three months later, another analysis cites it as a source. Six months later, a fabricated number sits in somebody's spreadsheet.

The editorial fix is concrete and unglamorous. Mark this document as blocked for insufficient input. Prevent automatic publication. Retrieve the original text and re-extract. Check the pipeline logs to find which stage died. And if several other documents in the same batch are equally empty, the problem is no longer one article — it is a system fault.

One small detail deserves recording: classification and extraction are decoupled. One succeeds in labelling; the other retrieves nothing. As long as the two do not cross-check each other, this failure will return, quietly, and always with a success code.

The fix is far cheaper than the damage. A minimum condition: if the extraction output contains fewer than one entity and one information point, the system blocks itself — no human detection required.

I am not writing this to describe a technical fault. I am writing because for several years the content industry has poured most of its anxiety into whether machines write as well as people. That question is misplaced. Machines write well enough. The problem lies one stage earlier: whether the raw material is real.

A good article depends on a real player, a real scoreline, a real contract, a real sigh in the tunnel. Everything else is just the frame. And the frame stays empty until somebody brings the truth and places it inside.