A Fuel-Price Report Tagged Tennis: How Sports Data Fools Itself

OPENING: 1215 GMT

1215 GMT, 26 September 2026. Rain fell evenly on the tin roof in Hai Phong, the kind of rain that is neither heavy nor light, just enough to keep a person seated. I opened the dashboard of the automated news-classification system our desk outsources. Items scrolled past: a quarter-final, a pre-tournament press conference, a transfer item, and one file tagged "tennis".

I opened the file.

The first line was about petrol: up 2.02 rupees, to 391.30 rupees a litre. The second line was about diesel: down 3.59 rupees, to 408.53 rupees a litre. Effective 26 to 28 September 2026. Pakistan's Oil and Gas Regulatory Authority, the Petroleum Division, ex-depot prices, Brent at 105.26 dollars a barrel, WTI at 92.78.

Not one player. Not one court. Not one set.

I laughed. Then I stopped laughing immediately.

Wrong label, right content. In twenty-eight years beside the court and beside the editing desk, I have met this exact shape of error many times, only under different names. A statistics sheet tagged with the wrong tournament. A serve-speed log assigned to the wrong player. A transfer item labelled "official" while its only source was an anonymous account with four hundred followers.

Some data never needs to shout; it only needs someone patient enough to read it. But to read it, you first have to know what you are reading.

CONTEXT: A NEWSROOM RUN BY LABELS

Sports news has changed faster in the past decade than a full Olympic cycle. Most desks no longer scan the wires by human eye at the first stage. They buy aggregated feeds, rent classification systems, plug them straight into dashboards, and hand editors a list already sorted into compartments: tennis, football, athletics, swimming, sports finance.

Every file carries a layer of labels. The label says where the file belongs, whom it concerns, whom it deserves to reach. The label decides whether the file is ever opened.

The file I opened that night carried the label "tennis" and its content was a fuel price adjustment decision. The document's structure was entirely ordinary: an old price, a new price, the difference, the validity window, the issuing authority. The original reporter did the job correctly. The tagging engine also ran exactly as it was programmed to run.

What matters lies elsewhere. Inside that document was an internal tension a hurried reader would skip. It recorded Brent up 1.5 per cent week-to-date while WTI fell 7.4 per cent over the same period. In the latest session, Brent slipped 1.3 per cent and WTI 1.9 per cent. Two benchmarks of the same market, over the same span, pulling in opposite directions with amplitudes differing almost sixfold.

An ordinary energy report can survive such a contradiction because its specialist readers are used to filtering it themselves. But once that report slips into a sports data system, it becomes a speck of dust. Dust does not break the machine at once. It simply sits there, waiting to accumulate.

The transfer window is the season when such dust multiplies fastest.

A WRONG LABEL IS SPORT'S OWN PROBLEM

There is a reflex I often meet in newsroom meetings: when data goes wrong, people blame the system. The system is soulless and cannot argue back. But a wrong label in sport has never been a foreign matter. It sits inside the way we count.

Take a recent tennis example. An unforced error is a human judgement recorded courtside. On the same rally, one recorder calls it an unforced error, another calls it a forced error. Across one match, two systems produce two different numbers. Yet on comparison pages these two numbers are still placed side by side as if they measured the same thing.

Football is the same. A player's distance covered depends on the tracking provider. Chances created depends on how a chance-creating pass is defined. Touches in the box depend on whether the data editor counted blocked phases.

When a label arrives without a definition, it is not yet data. It is only a name pasted over a blank space.

Comparing two players through their unforced error counts at two different tournaments is like comparing petrol prices in two countries with different subsidy structures. Both comparisons produce a number, and both are meaningless unless the reader knows the mechanism behind it.

So I keep a habit that has followed me for more than twenty years: I build my own tables, draw my own charts, and note the source of every column. When a data vendor changes a definition without announcing it, I am the first to notice, because my chart line breaks at exactly that point.

THE TRANSFER MARKET: NOISE WEARING A NEWS LABEL

Nowhere are labels pasted so carelessly as in the transfer window. An agent's status update, a photograph of a player dining in a city, a caption deleted after three minutes: any of these can be tagged "transfer news" and pushed to the top of the page.

My own grading of transfer information runs through three tiers. Tier one is documentary evidence: an official club statement, a registration document, published contract terms. Tier two is structural signal: how much wage-bill room remains, from which date a release clause takes effect, where the agent is, how long the registration window stays open. Tier three is pure rumour, meaning everything else.

Most of what readers meet daily sits in tier three yet is presented in the format of tier one. That is precisely the anatomy of a labelling error: tier-three content, tier-one label.

Based on my experience following matches and transfer windows, one question is always useful before believing anything: who benefits if this information spreads? An agent benefits when two clubs show interest, because negotiating value rises. A club benefits when supporters believe a star is being pursued, because season tickets sell faster. An anonymous account benefits from followers, because followers are money.

What is notable is that nobody has to lie. It is enough for a label to be pasted loosely.

THE RIGHTS BUBBLE AND THE EXPENSIVE LABEL

Higher up the industry, the story of labels becomes even clearer. For years streaming platforms bought sports rights above what they could monetise, spending investors' money to purchase exclusivity. They sold viewers a label: only here, only from us.

But viewers do not buy an exclusive label. They buy a match. When the match ends, the label is worthless too.

Cable television made exactly this mistake decades ago: paying for channel bundles, billing subscribers, then discovering that younger viewers no longer bought bundles. Streaming platforms repeat that structure with a technological skin. The money spent is labelled "content investment" when in substance it is the cost of acquiring distribution rights.

When that cash flow slows, the first budget cut is usually data verification. Nobody sees it. It generates no views. It merely, quietly, stops wrongly labelled files from entering the next news cycle.

TWO LESSONS A WRONG LABEL TAUGHT ME

In 2026 I was the only woman in the athletics press area at SEA Games 29 in Kuala Lumpur. I found that Nguyen Thi Oanh won the women's 1500 metres with a negative split: her first 800 metres were 2.3 seconds slower than her final 700. When I brought the analysis to a male editor, he laughed and said women do not understand pacing.

I did not argue. I spent three weeks rewatching the footage, splitting it into 100-metre segments, and published it on my own blog. The piece reached 50,000 views in 48 hours and was shared by the national team's head coach.

The lesson was not the view count. The lesson was this: the label "women do not understand pacing" was a wrong label, and it nearly buried a correct data finding.

In 2026 the Oanh piece took me to the World Cup in Russia as a commentator for a new sports media platform. In the semi-final I mispronounced Luka Modric's name three times in the first half and was heavily criticised online. I withdrew into my hotel room, cried for 48 hours, and cut off all contact.

Then I rewatched all five of Croatia's matches. I counted more than 90 kilometres covered by Modric across the tournament, and 14 chances created from his passes. My later portrait of that invisible labour was shared by Croatia's Sportske Novosti.

This time the mistake was mine. And the correction was mine too: from then on, every broadcast had to pass a three-source rule for pronunciation.

Data files have no self-correction mechanism like people do. They can only be corrected if someone takes responsibility for the label.

THE CONTRARIAN ANGLE: THE DEEPER THE SPECIALISM, THE MORE INVISIBLE THE ERROR

The irony is that sports media is pushing hardest at the very thing that makes labelling errors harder to detect: specialisation.

One desk per sport. The tennis desk reads only tennis files. The energy desk reads only fuel price reports. The wrongly labelled file sits exactly in the crack between two desks, where nobody is responsible for reading. Errors do not live inside a desk. Errors live on the boundary.

People look at the ranking table; I look at what the ranking table hides. What it usually hides is the process that produced it.

The second reflex is no less suspect. When a bad file is found, the newsroom's natural reflex is to delete it and clean the data. But that bad file is the only evidence that the classification system is faulty. Deleting it deletes the audit trail, and next time the error returns under a different label, better formatted, harder to find.

Rebellion does not have to be loud; sometimes it is quietly rearranging the numbers. Keeping the bad file, recording the error rate, publishing that rate for readers: a small act with far more weight than a public apology.

WHAT REMAINS

Elite sport is the art of repetition, and of breaking repetition. But a data system never breaks its own loop. It repeats until someone opens the file and reads.

The file labelled tennis that night was no minor incident to laugh off. It was a reminder that in an industry selling labels to audiences, the thing least examined is the labels themselves.

Readers have the right to ask one simple question of any sports report: where did this data come from, and who labelled it. Writers have a duty to print that answer beside the content, not in a faint footnote at the bottom of the page.

If the next generation learns one thing from my twenty-eight years of reading numbers, I hope it is this: do not trust the label, read the content.