The "football" Label Stuck on the Acapulco Flood Report: An Error at the Edge of the Frame

Late on 9 October, when my data dashboard surfaced a row tagged “football,” I had already poured coffee for a night of transfer-window analysis. That row contained no player. No club. No scoreline, no line-up, not a single transfer fee. It described Tropical Storm Simón, the flooded streets of Acapulco in Guerrero, Mexico, downed utility poles and landslides. A civil-protection bulletin wearing the shirt of a football match. I found the error not at the centre of the pitch, but at the edge of the frame. In 29 years in this trade I have learned one simple thing: before analysing, be sure that what you are looking at is a match. In 2026 I started out at Báo Bóng đá, then worked as a correspondent for Báo Thể thao Thế giới in Madrid. I have covered eight Olympic Games, eight World Cups, and many editions of the Giro d'Italia and the Tour de France. At 36, I sat in an editing room in Chengdu, reviewing 147 contentious refereeing incidents across 28 rounds of the Chinese league, and found 12 incorrect offside decisions directly tied to camera placement. Since then I have understood: bad data is not loud. It slips quietly into the system, and then quietly keeps being wrong. The Acapulco bulletin is one such case. Local radio and the civil-protection agency Protección Civil, citing Quadratín Guerrero, warned of fast-moving water across main roads, stranded vehicles, and a state order suspending classes. Those are clear, verifiable facts. Yet the label attached to them was “football.” Inside the VAR room, when a referee reviews an incident, people tend to think the problem lies in the final decision. My experience says the problem lies one step earlier: which camera angle you choose to watch. In July 2026, in the France–Croatia final, referee Nestor Pitana used VAR and awarded France a penalty after the ball struck Ivan Perisic's hand. I measured six main broadcast cameras. Only one angle showed Perisic's arm extended in a way deemed “unnatural,” and that was the only angle shown in the VAR room, for one minute and 47 seconds. My 2,400-word analysis was shared more than 50,000 times in 48 hours. Something similar is now happening with sports data. An automated classification pipeline chose the wrong “camera angle” for the Acapulco bulletin. It saw words, saw figures, saw the structure of a news item, and labelled “football” something with no football entity at all: no club, no player, no coach, no competition, no transfer, no club finance. Multi-angle verification is my first principle. Facing a contentious incident, I never accept a single camera angle. I reconstruct the decision chain: timing, camera placement, player movement data, and the referee's cognitive limits. With the Acapulco bulletin, that chain breaks at the very first link, because the thing called “football” has not one football link in it. It took 37 replays before I understood that the human eye is not a measuring instrument. Machines are the same, except they are wrong faster and at greater scale. A wrong label does not vanish. It flows into the aggregate table, into the prediction model, into the report sent to a client. The next day, someone may be “analysing the form” of a club that does not exist, based on a flood that does. Based on my experience following matches, I always re-check the origin of every figure before using it. In 2026, in the round-25 match between Guangzhou Evergrande and Shanghai SIPG, Wu Lei's goal was disallowed for a 15 cm offside, yet no camera captured the correct horizontal plane. I set about building my own database of “viewing-angle error.” Every time a system assigns a label, I want to know where it is looking from. With Acapulco, it is looking from an angle that does not exist. What we call “seeing with our own eyes” is only one camera angle, and that angle can be set wrongly. During the transfer window, noise always drowns out signal. Every day brings hundreds of rumours, thousands of data rows, and readers need a credibility filter. But that filter is only as good as the labels on its input. A “football” label wrongly stuck on a flood bulletin is a speck of dust. When thousands of such specks pile up, the filter starts returning meaningless results, and readers gradually lose faith in the whole system. There is a counter-reading worth weighing. When a system labels a flood bulletin “football,” our first reaction is to blame the machine. But people make exactly the same mistake, only more slowly. I have seen a sports editor push a traffic-accident story into the football section simply because the victim was a supporter. Machines learn from precisely those habits. The empty stadiums of 2026 showed me this: VAR does not save football, it exposes football. Likewise, a labelling system does not create information chaos; it exposes the chaos already present in how we define a sports story. The problem is that we lack a mechanism to catch the error before it spreads. In football we have a VAR assistant to review an incident. In the data pipeline, we usually have no one standing at the edge of the frame. The misclassification of the Acapulco bulletin harms no one. But it is a signal: if a flood bulletin can wear football's shirt, then a false transfer story can wear the shirt of truth. A system that can call a flood in Acapulco football can also call a football match something else. Machines will keep erring, and erring at scale. The job of the professional is to stand at the edge of the frame, long enough to recognise when what they are looking at is no longer a match.