The 'Football' Label on a Report With No Football

On my screen, on a late weekend night, there was a file weighing less than two kilobytes. Its first line read: Domain Label — Football. The seventeen information points stacked below it contained not a single player's name, not a club, not a competition, not a score, not a transfer fee, not a line of broadcasting revenue. The only thing carrying the word 'football' in that file was the label stuck onto it. The real content sat in the lower half of the page. It was a demonstration by members of the trans community at the Cuauhtémoc borough government headquarters in Mexico City, which ended in material damage to the facilities; the borough authority then set up a dialogue table to address demands relating to workspaces, housing, community dining halls, street-vending stalls, and an administrative review of a hotel. Walls were broken. Windows were smashed. Slogans were sprayed across the stone. No team. No coach. No match. I sat for a long time in front of that word 'Football'. Not because the content confused me. The content was clear, coherent, even serious. What made me stop was the label. A file with not a single grain of football in it had been filed in the football drawer, and that did not happen in some dark corner of the internet. It happened inside a professional pipeline, where there are people who label, people who check, people who publish. I once mispronounced the name Perišić, but I am never wrong about what I have witnessed. And what I am witnessing here is a system error, not a typo. Football does not end at the ninetieth minute; it runs to the last line of the bank statement — and this time, the last line of the bank statement was a wrong label. The machine that cannot say 'no' In ten years of watching the sports-content industry, I have never seen the gap between how much information is produced and how much is actually read carefully grow as wide as it is now. Every Premier League round has ten matches. Every match has about twenty-two players, every player has hundreds of metrics, every metric is sliced by the minute. Add the lower divisions, the women's game, the youth game, the national cups, the transfer windows, the board meetings, the financial reports filed with regulators. No newsroom, not even the largest ones I have worked with, has enough people to read every line. That gap is filled by systems. News is collected automatically, classified automatically, labelled automatically, then pushed into automated analysis pipelines. This is a rational economic development. Without it, fans in Vietnam would never read about a League Two match on the night it is played; they would wait days, or never read it at all. Automation has widened coverage in ways nobody could have imagined ten years ago. But there is one thing automation still cannot do, and I suspect it never will unless humans teach it: say 'no'. Say 'I don't know'. Say 'this document is not in my field'. A machine built to always answer will always answer. That is its nature. And when a machine always answers, its final answer is no longer the truth — it is the shape of the question. The file in my hands is proof of that. It was fed into a football-analysis pipeline. That pipeline has a nine-dimension framework built specifically for the football industry: tactical and technical analysis, club finance and the transfer market, sporting results and the public-opinion cycle, league landscape and team positioning, rules and governance compliance, management and the dressing room, risk profile, media narrative and expectations, and finally football-industry transmission. A framework like that exists to read a match, a contract, a crisis at a club. It does not exist to read a protest in a borough of Mexico City. Yet the file went through it anyway. And the first question I asked was not 'what did the pipeline do', but 'who let it through'. Seventeen information points I started by counting. This is a habit I formed at seventeen, when I spent three weeks in a small London office cross-checking line by line the financial report of the Leyton Orient youth academy. Back then I found thirty-seven sponsorship contracts with unusual refund clauses, with money flowing through a shell company in the British Virgin Islands. I took it to a senior male editor. He laughed and said girls usually watch football with emotion, not with the books. I did not argue. I started counting again from the top. This time, too. I counted. Seventeen information points. Not one of them mentioned a team, a competition, a player, a coach, a match, a goal, a card, a transfer fee. Seventeen points, all revolving around a civic protest and a borough government's administrative response. Points one and two describe the protest and the demands raised. Point three records the borough authority as a party to the events. Points four and seven describe material damage: walls, windows, doors, graffiti, and an allegation of stolen furniture. Point five records that a dialogue table was set up. Points eight, nine and fourteen concern an administrative review of a hotel. Points ten, eleven and twelve concern street-vending stalls, housing and community dining halls. Point thirteen lists demands relating to sex-work commercial premises. Point seventeen notes that administrative procedures are under way. I read it a second time. A third. Then I built a cross-check table between the nine-dimension framework and these seventeen points. The result was the same on every line. Nine dimensions, nine nulls For the first dimension, tactical and technical analysis, I need a formation, a playing style, a coaching duel. I need expected goals, passes allowed per defensive action, possession share. I have none of it. All four cells of this dimension — sophistication, execution, personnel fit, key data — are empty. Not because I am lazy, but because there is nothing to put in them. For the second dimension, club finance and the transfer market, I need broadcasting revenue, commercial revenue, wage expenditure, net debt. I need a deal, a contract structure, a panic premium. This document talks about an administrative review of a hotel and about street-vending stalls. That is municipal governance, not football finance. I cannot blend the two without lying. For the third dimension, sporting results and the public-opinion cycle, I need a table, recent form, a fixture list. I need pressure on the manager, on the key players, on the board. The only thing that could be called 'public-opinion pressure' in this document is a civic protest. These are two different phenomena in nature. Merging them is an analytical error, not a creative act. For the fourth dimension, league landscape and team positioning, I need a league, a set of competing clubs, a squad-value table. The only thing resembling a 'landscape' here is the urban and commercial environment of the Cuauhtémoc borough. That is administrative geography, not a football league. For the fifth dimension, rules and governance compliance, I need financial fair play, transfer-registration rules, disciplinary sanctions, competition eligibility. The 'governance' in this document is a local authority handling citizens' demands, a category outside the rule systems of FIFA, UEFA or national associations. The destructive acts mentioned, if they carry legal meaning at all, fall under civil or criminal law, not football disciplinary regulations. For the sixth dimension, management and the dressing room, I need an owner, a recruitment board, a leadership hierarchy inside the dressing room, a manager-player dynamic. The 'management' in this document is a borough government. There is no dressing room here to analyse. For the seventh dimension, the risk profile, I need sporting risk, financial risk, personnel risk, rules risk, public-opinion risk, systemic risk. The risks actually described in the document — property damage, an allegation of theft, unresolved civic demands — belong to an urban risk register, not to this framework. For the eighth dimension, media narrative and expectations, the source document is recorded as neutral, objective, informational in purpose. It is a standard local news report, not a football story. There is no hype cycle, no expectation gap, no transfer rumour whose credibility needs checking. For the ninth dimension, football-industry transmission, I need a chain running from academy to club, to competition, to broadcasting, to commerce, to derivative markets. That chain does not exist in this document, because its subject is urban commerce, housing and borough administration. Nine dimensions. Nine nulls. I wrote the same conclusion for each: insufficient information, cannot assess. That is the sentence a machine that always answers hates most. But it is the most honest sentence I can write. A muckraker who loses accuracy is nothing but a gossip. And a gossip does not need a nine-dimension framework; he only needs a mouth. What is striking is that the pipeline itself recognised the problem. It stated plainly that this document is not football content, that the label 'Football' is a misclassification, and that forcing the document through a football framework would invite fabrication. It even recommended re-routing the document out of the football stream and correcting the label. In other words, the software knows how to say 'no'. The problem is that the humans have not yet listened. The reasonable part of an error I do not want this story read as an indictment of automation. That would be a lazy conclusion, and I have learned that lazy conclusions are usually wrong at exactly the point that matters most. There is a reasonable part to this error. First, scale. With thousands of documents passing through a pipeline each day, a small error rate is mathematically unavoidable. A labelling system that is ninety-nine per cent correct will still mislabel dozens of documents a day. That absolute number sounds alarming, but it says nothing about the quality of the system. It only says that the system is running. Second, the label itself is not a claim about truth. A label is a routing instruction. When I stick the label 'football' on a file, I am saying 'send this to the football desk', not 'this file is about football'. The two are close enough to be confused, but they are different. A routing error is not a cognitive error. It is more like a letter dropped into the wrong drawer than a letter that lies. Third, the fact that the pipeline caught its own error is a good sign, not a bad one. A bad system would have fabricated a football analysis out of this document. It would have called the protest a 'dressing-room crisis', the dialogue table a 'board meeting', the hotel review a 'financial investigation'. It would have produced a fluent analysis, plausible in form and utterly wrong in substance. This system did not. It chose emptiness over fabrication. In an age when false content is generated faster than true content, that is a choice worth noting. Fourth, the reader is not necessarily harmed by a routing error. I have read a great many sports reports in my life, and I do not recall a single time I encountered a protest in Mexico City presented as a transfer story. This error lives at the operations layer, not the publishing layer. It has not yet reached the eyes of a fan. But — and this is where I have to stop — there is a difference between 'has not yet reached' and 'cannot reach'. A routing error at the operations layer becomes a publishing error at the exact moment nobody checks again. And in a pipeline running on speed, that moment is always closer than we think. I think of an evening in Moscow in 2026, when I was eighteen and the youngest reporter in the press room. I mispronounced the name Ivan Perišić three times on camera. A local broadcaster mocked me directly, saying I should go back to the keyboard rather than stand in front of a microphone. I was humiliated. But that night, instead of crying, I opened the English Football Association's financial records. I found a 'match compensation fee' worth two point seven million pounds, tied to a murky sponsor and absent from the official report. My independent investigation was later bought by a major newspaper. The lesson I drew was not 'be careful with players' names'. The lesson was: every small error can be the thread of a large one, if you are willing to count. A wrong label today can be a wrong line tomorrow, and a wrong line tomorrow can be a wrong investment decision next week. In football, where billions of pounds flow through contracts priced with numbers that very few people actually read, accuracy is not an abstract ethical standard. It is a fence protecting the money flow. A contract exists only on paper; the money evaporated long ago. And news pipelines are the same: credibility exists only on the homepage, while the truth evaporated long ago — right at the labelling layer. The question that remains I did not write this to convict a file. A file is not guilty. I wrote it to raise a question I believe is the central question for the sports-content industry in the coming years: when speed becomes the measure of success, who is responsible for stopping? Fans are the ones who pay, but they are usually the last to see the books. They pay for broadcasting rights, for subscriptions, for advertising, for shirts printed with a player's name. In return, they deserve a news pipeline that dares to say 'I don't know' when it genuinely doesn't know. A pipeline is only trustworthy when someone dares to ask: what is this label based on? I put that question to three different people this week. All three gave the same answer: 'It's a system error.' None of them said 'it's my error'. And perhaps that is the most troubling thing — not that the system makes errors, but that no one still feels responsible for them. I keep that file in a separate folder on my machine. I named the folder 'wrong label'. It is small, only two kilobytes, but it is a reminder that in this industry the most dangerous thing is not a wrong number, but a right number given a wrong label. Because a wrong number still gets checked. A right number given a wrong label gets believed at once. An academy can train talent, but it cannot train honesty. And a news pipeline is the same. It can be built with the best algorithms, but its honesty must still come from people — from an editor willing to stop, to ask, to say 'no'. The last question I leave with the reader is not whether this pipeline will make errors. It will keep making them. The question is: when the next error happens, who will be the first to open the file and count?