An EV Advertorial Inside a Football Database: The Classification Error and the Price of Trust

Late at night in Marseille, I opened the data file the newsroom had sent over. The label at the top read simply: “Football”. A few lines down, what appeared instead was a length of 4,740 mm, a wheelbase of 2,840 mm, a 150 kW electric motor with 280 Nm of torque, and a claimed range of 450 km. Not one team. Not one player. Not one minute of stoppage time.

Thirty years in the trade, from local radio shifts in 2026 to flights to Moscow in 2026 and Doha in 2026, I am used to opening a file and finding a match. This time, I found a car.

It was the VinFast VF MPV 7, a three-row electric vehicle sold in Vietnam. The source piece — in truth an advertorial dressed as a news article — walked through dimensions, drivetrain, range, air-conditioning layout, a testimonial from the family of Ms. Nguyen Thuy Duong in Nghe An, and a stack of buyer incentives: a list price of 750 million VND, a promotional price of 682,500,000 VND, a “Zero-dong car purchase” plan financing up to 100% of the vehicle value, and a free-charging programme on the V-Green network.

Someone had stuck the label “football” on all of it. And here I sat, a sports editor, reading a spec sheet as if it were a match report.

Context: when a database learns the wrong name

In the sports industry we have grown used to content pipelines running through many stages: an article is collected, tagged, classified, then pushed into a database feeding commentary, statistics or prediction models. Every stage looks harmless. But a single wrong label poisons the entire stream behind it.

In Vietnam, where football is the most widely followed sport, such databases are multiplying. Newsrooms, platforms and analytics firms all want a clean football store to feed their algorithms. And it is precisely that ambition that lets a car advertorial slip through before anyone notices. The speed of collection has outrun the speed of verification.

I once witnessed a far smaller episode with no small consequences. In 2026, while editing at a sports magazine, I replaced a short piece on pressing with a 3,000-word article counting 127 ball recoveries by André Zambo Anguissa in the opponent's third. It drew 1,400 shares, more than any piece the magazine published that year. I learned that data can become poetic material. But I learned the opposite too: bad data can become a lying poem, read aloud very fluently.

Core: what is actually inside the file

Strip away the shell and the source article is fairly clear. The price arithmetic checks out: 750 million VND minus 9% is exactly 67.5 million VND, leaving 682.5 million VND. The 450 km range carries no test cycle — the kind of disclosure common in promotional copy, where CLTC or NEDC figures are more generous than WLTP, and far more generous than real roads. A multi-link rear suspension is a genuine plus for second- and third-row comfort. A 150 kW motor is adequate, not exceptional, for a fully loaded seven-seater.

At 4,740 mm long with a 2,840 mm wheelbase, the car sits squarely in the compact MPV class. The consequence is that the third row and the boot with all three rows up are structurally constrained — something the source never measures, even while asserting the car is “enough for the whole extended family”.

On the incentives, two things stand out. The “Zero-dong car purchase” plan runs to 31 December 2026, while the discount ends earlier, on 19 December 2026. That twelve-day gap is a familiar technique to push buyers to decide before the price lapses. The conditions for the promotional price are called “conversion conditions”, yet are never defined anywhere in the piece. A genuine information gap.

The free-charging package is the same story. Up to ten sessions a month on the V-Green network, running to 10 February 2029. Multiplied by the claimed 450 km, one arrives at more than 4,000 km a month. That multiplication only holds if every session is near-full and hits the claimed efficiency — an optimistic assumption. The source also never says whether the battery is included or leased separately, a variable that decides total cost of ownership.

And the testimonial? One family, one case, one voice. In research, that is called a sample size of one. That lone testimonial rings out like an empty stadium: it does not reflect a crowd, it reflects the seller's voice.

If this were a match report, I would write: the home side controlled possession, but every pass was aimed at its own goal.

Contrarian angle: where the fault really lies

People will call this a technical glitch, a speck of dust in the machinery. I don't think so. A car advertorial tagged as football exposes three layers of trouble at once.

The first is classification. Someone, or some machine, read a headline, spotted a loose keyword, and applied a tag. The machine does not understand content; it only matches patterns. This is the blind spot of every automated system: it is good at imitating understanding, but it does not possess it.

The second is editing. A file labelled football should have passed a human eye before entering the store. That it slipped through shows the checking stage is thinning, perhaps for speed, perhaps for staffing, perhaps out of blind faith in machines.

The third, and the layer I care about most, is the boundary between advertising and journalism. The source has no outlet name, no byline, no disclosure of a commercial relationship. It wears the coat of a news article to sell a car. When an advertorial mimics the voice of journalism well enough to enter a database, the problem is no longer a car. The problem is a whole system's capacity to tell real from fake.

Here I want to use an image I still carry from my years writing about football: a team is not merely eleven people; it is a system of equations that can run. A database is the same. It is not merely thousands of articles; it is a system of beliefs that can run. A single wrong variable and the equation yields the wrong answer — and that wrong answer gets read aloud as truth.

One small detail I don't want to skip. The source calls the incentive programme “the second Green Future programme”. That word “second” says there was a first. It hints at a repeating promotional cycle, not a one-off event. The transfer market is a match with no referee, where every number is a free kick — and so is the EV market.

Takeaway

People ask me why I bother with a car. I bother with what stands behind it. Modern football lives on data: transfer data, fitness data, distance-covered data. A contaminated database will gradually teach algorithms things that are not true, and then the algorithms teach us. A car advertorial today can be fake news about a player's injury tomorrow.

Data does not score goals, but it knows where the ball is going. And if we let it take the wrong road from the very first label, then every analysis behind it — however sophisticated — is just a shot into empty space.

The football dream never lives in the result, but in the moment the ball has not yet touched the ground. The same holds for data: its value lies in the moment before it is labelled, when we still have to ask ourselves what we are looking at. Perhaps the task is not to delete the file and keep things clean, but to pause long enough to ask: who applied that label, and why did we believe it so quickly.