Two Funeral Vehicles Labelled “Football”: A Misclassification and a Sourcing Lesson in Transfer Season

Near midnight, a line of data lit up the screen in my small flat in Saigon. The classification label read, plainly: football. I opened it, one hand still around a cup of coffee that had gone cold long before. Inside there was no team, no scoreline, no player. There was a short report from Islamabad, saying that a religious centre in the Pakistani capital had just completed two mobile units dedicated to carrying the deceased to washing and shrouding, free of charge to the community.

I sat still for a while. More than thirty years of watching football taught me a reflex: hearing the wrong note in the orchestra before knowing why it is wrong. Two funeral vehicles. A football label. Between them lies a gap I know well, because I have stood inside it more than once: the place where our trade lets through the things that are not ours.

The pitch does not lie. Only the people in the stands lie to themselves every weekend. This time, the one lying was not a spectator. It was a classification machine.

That report sat inside a dataset most sports newsrooms now use daily: a system that reads, labels and redistributes content from thousands of sources. It names the topic before a human finishes the first line. When the label is right, we save a few minutes. When the label is wrong, the mistake does not stop at one article. It travels into aggregated feeds, into the morning radio bulletin, into the data package sent to partners, into a page in another country where nobody knows where Islamabad is.

The file on the Islamabad report carries eight information points. Two purpose-built vehicles. A free service. One cost figure: 250 million rupees, roughly USD 0.9 million, or about 125 million rupees per vehicle — around USD 450,000 for a converted unit. I circled that number, because in my trade an unsourced number is a number that does not yet exist. The three most consequential facts in the file — that the vehicles exist, the cost, and the interior design — carry no source at all in the original record. The only attributed statement belongs to Abdul Wajid, General Secretary of the Quba Islamic Centre, who said the project was completed under the directives of the centre's president.

Above the report, the system inserted a Reuters photo caption of a mobilisation and peace-talks event, with the Faisal Mosque visible in the frame. Two things sitting side by side, unrelated to each other, yet enough for an algorithm to bundle them under one topic: Islamabad, region, flashpoint, sport.

That is why I do not treat this as a joke about machines. We are in the middle of the transfer window, the phase when sports feeds are densest, loudest and most easily mislabelled of the entire year. A name, a number, a photo of a player stepping off a plane, and a file is tagged as a transfer and flows into hundreds of pages within hours.

After more than thirty years of watching football, I have built a five-tier filter for ranking rumours, and I still use it every day during the window. Tier one is paperwork: contracts, release clauses, registration documents. No paperwork, no story. Tier two is a named insider speaking directly and staking their name on it. Tier three is two independent sources that are not copying each other. Tier four is an original photograph, original data, original documentation. Tier five is the anonymous source — the tier I read to learn what people are thinking, not to learn what happened.

Where does the Islamabad report land in that filter? Tier two, thanks to one named statement. Tier five, because three important facts have no source. And it does not belong to my field. That is the only honest conclusion available, and I wrote it down before writing anything else.

In sports reporting, the most expensive error is not publishing something false; it is placing something true in the wrong slot. A correct fact about a community project, set beside a correct fact about a transfer, produces a third thing: the illusion of a link that does not exist. The transfer window lives on that illusion. Fans do not consume individual facts; they consume the links between them. And links are the easiest thing in the world to invent.

A transfer is not a transaction. It is a separation signed in legal ink. But in mid-July, when every newsroom races to report the same name, that legal ink is usually replaced by the watery ink of a social media post. An account with a few hundred thousand followers posts a line. An aggregator copies it. A television bulletin reads the copied line. A podcast discusses the bulletin. By the fourth pass, the story has a subject, a verb and a past tense, and nobody remembers it began with a nameless account.

In a transfer window, three things tell the truth: money, contracts and the behaviour of agents. Money tells the truth because it has to be wired. Contracts tell the truth because they have to be signed. Agents tell the truth because every call they make has a purpose, even when the words inside the call are a lie. Rumours tell you nothing. A rumour is only noise with a shape, and the shape is usually drawn by the reader.

In June 2026, when short-video platforms were just emerging in Vietnam, my desk handed me an odd experiment: retell a 2-2 draw between Ho Chi Minh City FC and Hanoi FC in the V-League in exactly 90 seconds, in the language of poetry. I shot it, cut it and voiced it myself. The video drew 45,000 views in a week, the highest for the sports section at the time. Telling a story in 90 seconds taught me that a moment does not need to be long, only heavy enough. But it taught me something less poetic too: to earn 90 heavy seconds, I had to watch the match three times and check every phase against the original footage.

On 30 June 2026, from the commentary cabin in Kazan, I watched Kylian Mbappe, 19 years old at the time, score twice inside four minutes, in the 57th and 61st, sending France past Argentina 4-3 in the World Cup round of 16. I cried on shift, threw away the script I had prepared that morning, and rewrote the whole piece from scratch that night, 2,500 words long. It was shared 12,000 times in 24 hours. The tears in the France–Argentina cabin taught me that this job was never a job; it is a life. They also taught me that emotion cannot replace verification: I still had to reopen the footage to be sure I had logged the 57th and 61st minutes correctly.

In March 2026, global competitions stopped. I sat in a city without football, listening to motorbikes outside my window instead of a crowd. By June 2026, when the Bundesliga returned to empty stands, I wrote a series on the sound of absence: a coach shouting, a ball meeting a boot, a player gasping in a Bayern–Dortmund match that finished 1-0. The series, “Nine Minutes of Silence,” drew 30,000 reads, the highest in the sports department that year. In that silent summer, I heard a city breathing to the rhythm of a ball nobody touched.

Since then, I have believed that absence has its own weight. A file containing no football content is still an event, if we read it as a pause rather than a void. When something disappears from the pitch, it does not disappear from the story. It merely changes place: from the singing of the stands to the breathing of the players.

Saigon watches football on a small screen, but its heart still hangs in the stands. That night, I too was watching football on a small screen, except mine held a data table. And my heart hung elsewhere: on the belief that every line must have a source, every number must have documentation, every name must have someone accountable behind it.

Back to the Islamabad file. If I handled it as a transfer file, that is, with the highest discipline my trade possesses, I would do four things.

First: re-establish the domain. A community service project in Pakistan's capital does not belong to football. The file is mislabelled. Full stop.

Second: measure the sourcing gap. The three most consequential facts have no source. In a transfer file, that is the threshold of unverified rumour, and no outlet should cite it as a fact.

Third: test the number. 250 million rupees for two vehicles is a very high outlay if it covers only vehicle conversion. It may also include cold-storage equipment, dedicated shrouding systems and years of operating costs. It may. But the word “may” is not a source.

Fourth: read the photo frame again. A Reuters frame from a mobilisation and peace-talks event sits above a report about funeral services. There is no causal relationship between the two. It is a presentation error. But presentation errors enter collective memory faster than events do.

The familiar reflex here is to blame the machine. I do not think so. The machine learned from us.

Sports has taught its own data with very specific habits: publish first, check later; use the headline as the body; call a draw a shock; call a 20-year-old who has never played in Europe a blockbuster signing; let a photograph of a stand stand in for a whole article. A system trained on that archive will conclude that anything with a crowd, a big number and a foreign name might be football.

We grow old between generations of players, but we never grow old between matches. The collective memory of fans is built from highlights: goals, stepovers, moments. Nobody saves the 12th minute of a match in which nothing happened. So when a machine learns from that memory, it does not know that football is mostly minutes in which nothing happens. And it does not know that a funeral vehicle does not belong to the pitch.

The blind spot is not in the algorithm. The blind spot is that we have quietly agreed speed matters more than sourcing, and that a line with a wrong source is still better than a line with none at all.

That night, I closed the file and wrote one line in my notebook: mislabelled, civil domain, removed from the football dataset. Then I sat for another fifteen minutes, thinking about two vehicles carrying the deceased in a capital more than four thousand kilometres from Saigon, and a machine that called them football.

If a machine can label a funeral vehicle as football, who taught it that? And if we do not teach it again — by checking sources before publishing, reading documents before reading numbers, keeping a place for the minutes in which nothing happens — how many names will the next transfer window write in ink nobody ever verified?

I still believe one simple thing: if every line had to pay for itself with one verification, the noise would quieten on its own. And when the noise quietens, people can finally hear the ball meeting the boot.