When a Criminal File Gets Tagged 'Football': A Lesson in Clean Data for the Sports Newsroom

At two in the afternoon on 8 October, inside a closed courtroom in London, counsel for Andrew Mountbatten-Windsor stepped up to address Mr Justice Hilliard. The subject: an application for judicial review of the lawfulness of search warrants obtained by Thames Valley Police, in an investigation linked to an allegation of “misconduct in public office.” My phone buzzed. A colleague sent a link with a short line: “football piece, take it.”

I opened it. No team. No player. Not a single contract clause, not one PPDA figure, not one wage bill. Only a classification tag reading “football,” sitting awkwardly above a criminal file. I sat still for a few seconds. Twenty-eight years in this trade, I had learned to distrust numbers that look too good and deals that look too easy. But I had never thought I would have to distrust the very label stuck on top of a news item.

The story itself is almost too simple to believe. Andrew Mountbatten-Windsor, the former Duke of York, is seeking to have a court quash search warrants tied to a Thames Valley Police investigation. That investigation is said to be connected to his association with Jeffrey Epstein. Two locations named in the warrants are Royal Lodge in Windsor and the Sandringham Estate. The warrants were issued by the Central Criminal Court, the Old Bailey. The Judicial Office confirmed the hearing was held in private. The subject himself denies all allegations.

Reading this far, a clear-headed football reporter would close the laptop. There is no football here. But the question worth asking is not about the content of the item. It is about how a file like this slipped into exactly the “sports” drawer of a newsroom.

I have spent many years in the corridors of transfer windows, where everyone believes speed is everything. Filing a minute early is winning. Based on my experience tracking and cross-checking thousands of items every season, I learned something few will say out loud: most errors do not come from reporting the wrong thing, but from classifying it wrong at the very start. A mislabelled item will pass through every downstream checkpoint without anyone bothering to open it again, because the label has already told them this one is fine.

In the modern news system, the classification tag is what stands in for a human being. When a newsroom runs at the scale of thousands of items a day, no editor has time to read them all. They rely on tags. The tag says “football,” they send it to the sports desk. The tag says “transfer,” they file it under transfers. It is that tag, not the content, that decides the fate of a piece.

And this is where I want to linger. The market does not run on money; it runs on information. I have always believed that, whether on the transfer market or in a newsroom. A single piece of information that goes down the wrong path spawns a chain of wrong-path consequences behind it: an editor who accidentally writes a skewed headline, a recommendation algorithm that pairs it with the wrong context, a reader who finishes it and comes away believing their club is tangled up in a criminal investigation. No one intends any of that. But together, they produce it.

I once built a process of cross-checking three independent sources. It sounds rigorous, but it only counters the first kind of error: information that has not been verified. It cannot counter the second, more dangerous kind: information that has been verified correctly but placed in the wrong drawer. A true fact in the right place is knowledge; a true fact in the wrong place is fake news, even though nobody lied. That is a class of error no source-checking machine can catch, because it is not wrong at the level of the fact. It is wrong at the level of meaning.

Looking closely at this file, the source quality is far more interesting than the wrong label. Thames Valley Police and the Judicial Office are on-record, citable, named sources. The subject's denial is also citable. But most of the remaining claims — about the dates of the searches, the locations, the procedural sequence — carry no source at all. That is the typical aggregated brief: it sounds certain, but in truth a single source stands behind many sentences.

In my trade, we call that a single source presented as though it has been verified many times. It is dangerous because it does not look like a rumour. It has dates. It has named agencies. It has a judge. It wears the clothing of precision. Data points the direction; instinct points to the door. And my instinct, after twenty-eight years, says there is one thing here worth scrutinising: the timeline. The events are dated 18 and 19 February 2026, while the hearing is dated 8 October. Those figures sit ahead of normal reporting convention. Either this is a prospective schedule, or a timeframe error that needs re-verification. To me, this is exactly the kind of detail a serious newsroom must pause over before publishing.

What stands out is that the item itself is fairly careful. It states plainly that seeking judicial review of the warrants does not mean the police did anything wrong. That is an important caveat, one many items of this kind skip. It also shows both sides are managing disclosure tightly: one asks for a private hearing, the other declines further comment. In this trade, when both sides go quiet, that is when a reporter must work the hardest, not the least.

This is where I want to go against my own instinct. The first reaction of a veteran is to blame the classification algorithm, to demand it be replaced, to demand humans re-check everything from scratch. But I do not think that is the root.

The root is that we have quietly agreed speed matters more than correct placement. A sports newsroom today does not only report on football. It reports on club finance, on law, on the politics of federations, on the legal cases surrounding famous figures. The line between sports and current affairs has blurred so far that a criminal file can slide into the sports drawer without anyone finding it odd. The problem is not that the machine misclassifies. The problem is that we have stretched the definition of sports until it means nothing clear at all.

And here is the paradox: that mislabelled item was handled more carefully than many correctly labelled ones. It has on-record sources, a caveat, balance. Meanwhile, thousands of transfer items a day go up with a single source and not one caveat. The biggest shock is not on the pitch; it is in the balance sheet. And this time, the shock is not the wrong item, but that we worry about it while ignoring the right-place items that are hollow inside.

Andrew Mountbatten-Windsor will have to wait for the court. My newsroom does not have the time to wait. The real question is not how to stop the classifier from mislabelling again, but this: when a true fact is placed in the wrong drawer, who among us will be the one to open it and read it from the start? If no one does, then every label can become an accidental lie. And in this trade, the accidental lie is the hardest kind to forgive.