A "Tennis" Label on a Dairy Story: The Silent Flaw in Sports Data

In Los Angeles that morning, when I opened the newsroom's internal feed, a single line sat across the screen like a spreading ink stain: "Domain Label: tennis." Beneath it was a story about a dairy company in Pakistan, a chief executive leaving his chair, a notice filed with the Pakistan Stock Exchange. Not one racket. Not one court. Not one serve mentioned anywhere.

Seventeen information points. Every one of them circled FrieslandCampina Engro Pakistan Limited — a listed company. Not a single line touched the ATP, the WTA, the ITF, or any Grand Slam. I sat still for a long while, staring at that label, and thought about something I learned early in this trade: sometimes the biggest liar in a news item is not the writer, but the labelling system standing behind him.

Behind the label

Let me be clear from the start: the source material belongs to the corporate world, not to sport. FrieslandCampina Engro Pakistan, known as FCEPL, is a dairy company listed on the Pakistan Stock Exchange, the PSX. The filing records the departure of chief executive Kashan Hasan, with a contractual notice period. The vacancy on the board of directors, the text says, will be handled under applicable legal and regulatory requirements. That is the language of company law, not of competition rules.

Hasan has more than twenty years of career behind him, having passed through Shan Foods and Reckitt, with roles stretching from Pakistan, South Africa and the United Kingdom to the Middle East and North Africa. Behind the company name sits an entire dairy value chain: more than 1,300 milk collection centres, two processing plants at Sukkur and Sahiwal, a dairy farm at Nara, and a 450 million US dollar foreign direct investment by Royal FrieslandCampina into Pakistan's dairy sector since 2026. Those figures mean something to an agricultural and consumer goods market. They mean nothing to a tennis court.

And yet the first-stage classification system tagged that item "tennis." Not because of a typo. Not because an editor misread it. But because a machine decided so, and nobody stopped it before it entered the data.

For someone who began her career at the fact-checking desk of Sports Illustrated in 2026, that detail is not small. It signals the disappearance of a quiet layer of this profession: the people who stay behind last to ask, "does this belong where it's been placed?"

A machine that cannot tell milk from a court

I have worked in newsrooms where every item passed through at least three layers of checking before reaching readers. The first layer was human. The second was a rule system. The third was a learning machine, increasingly confident in its own judgement. The paradox lives in that third layer: the faster it runs, the less it is audited. And in an age when financial wire copy moves hours ahead of sports copy, a dairy item tagged "tennis" is no joke. It is a symptom.

Classification machines rely on word frequency and surrounding context. If an item contains phrases that historically travelled with sport — "transfer," "contract," "notice period," "board," "termination" — it can be pulled toward whichever field shares that vocabulary density. In sport, "contract" usually attaches to a player. In business, "contract" attaches to an executive. The same word. Entirely different universes.

I asked myself: if the machine read the word "tennis" somewhere in its processing, where did it come from? Perhaps from faulty metadata. Perhaps from a contaminated training sample. Perhaps from the very way financial and sports wires share a single queue inside the system. What frightens me is that none of us can answer, because nobody kept a trace of that decision. The machine does not keep a diary. It labels and moves on.

And once the item carries a "tennis" label, it drifts into sports datasets like a speck of dust into a lens. It sits inside the entity graph — where every name is linked to another by relation — and begins to sow meaningless connections. FrieslandCampina may appear beside some tournament simply because an algorithm cannot tell milk from a court. Kashan Hasan may be filed among athletes with nobody checking. Royal FrieslandCampina, a Dutch group, may quietly slide into a chart of sports sponsors. Those false links do not vanish on their own. They layer, they drift, and one day they return in a table of statistics that a young reporter believes to be true.

I picture an editor ten years from now opening a database and finding the name of a dairy company sitting in a list of players. He will believe it. He has no reason to doubt it. Because data, to the modern writer, has become something close to sacred. We rarely ask where data comes from; we only ask what it says.

An empty stadium, it turns out, has a sound of its own — the sound of longing. But an empty space in the data makes no sound at all, and that is precisely the danger. An error that exists unheard will outlive the person who caused it.

Sport does not stand outside this spiral. It sits at the very centre, because it is the hungriest consumer of data there is: rankings, match breakdowns, transfers, performance indices, prediction models. Every wrong stream that leaks in here can be amplified a hundredfold within hours. And as clubs, tournaments and broadcasters lean ever harder on automated data, the number of people actually checking each item grows ever smaller. Staff cost is the first thing cut. Reputation is the last thing lost.

Across my career of following matches and following wires, I have noticed a rule: the gravest errors in sports media rarely come from writing down the wrong score. They come from believing something that was never checked. A wrong draw. A stat with no source. A name attached to the wrong club. And now, a dairy company filed as tennis.

They told me I do not understand football, but I understand what it does not say. And what the Pakistani dairy item does not say, once tagged "tennis," is a truth about the whole system: that we are building a house of data on sand, then confidently calling it evidence.

What an outsider sees

The counterintuitive angle sits here: this error is not the fault of one machine alone. It is the fault of how we now treat expertise. When a newsroom strips out its first-stage checking — the work I did at Sports Illustrated from 2026 — it does not cut cost, it cuts memory. The first-stage checker did not merely fix spelling. They were the person who knew FrieslandCampina is milk, not a tennis court. They were the person who noticed that a Pakistani name does not belong on a list of players.

What deserves saying is that when a female sports journalist like me questions an anomaly of this kind, the response is usually not "you're right" or "you're wrong." The response is usually a polite line like "the system has logged it, rest assured." That administrative silence is no less dangerous than the original error, because it turns the problem into the private business of the person who found it, rather than the shared business of the whole process. An error that is logged but not fixed is an error still alive.

Women don't understand football, they say. But I have noticed something: the people who catch a wrong label are usually outsiders — people not invited into the data room, people without editing rights, people forced to read with their eyes rather than their faith. In Moscow in 2026, when I asked Luka Modrić a question a male colleague called "poeticising everything," I learned that a naive question is sometimes the right one. Insiders are so used to the label that they can no longer see it.

Before they are a contract, they are children carrying a dream in search of a home. And before it is a line of data, every news item is a person, a decision, a morning when someone knocks on an office door to say goodbye. The machine sees none of that. It sees only the word "contract," and applies a label.

I wonder what would happen if we held sports data to the standard we hold a match to. In tennis, a point counts only when the umpire confirms it. In a ranking, a position has value only when the points system agrees. So why can a domain label — the thing that decides where an entire item belongs — be applied without anyone confirming it?

What needs watching, to my mind, is not the fate of one chief executive. It is whether this labelling error recurs. If, in the coming weeks, other financial items also carry a sports label, then the problem is no longer a speck of dust but a whole system gone blind. At that point the task is not to delete one line, but to recalibrate the classifier at its root.

What is left after the label

I do not conclude that the machine is the enemy. The machine only does what it was taught. The question lies elsewhere: who among us still has the patience to doubt a "tennis" label when it is stuck onto a story about milk? Who still has the curiosity to open that item, read its seventeen points, and notice there is not a single racket in it?

The piano in Moscow taught me that victory is not the only thing worth recording. Perhaps one day the best sports writer will not be the one who reads the numbers fastest, but the one who knows how to stop and ask: where did this figure come from, and who put that label on it?

The pandemic froze sport, but it could not freeze what we tell each other about ourselves. I still believe that. I only hope that, amid streams of data running faster than a serve, someone still stays behind — slow enough — to notice when a label is lying.