Mislabeled and Sourceless: When Sports Journalism Forgets to Ask 'Who Says'

A February morning in Boston, the temperature below freezing and the surface of the Charles the colour of dull tin, I sat in front of a monitor with a file an editor had just forwarded to me. The data label stated one word: tennis. I opened it. There was not a single player inside. No tournament, no scoreline, no serve statistic, not one name belonging to a court. Eighteen information points sat neatly in the file, and all of them concerned gold, silver, platinum, palladium, US Treasury yields, and a policy meeting of the Federal Reserve. I read it from start to finish twice, then a third time, more slowly. A report about the price of gold wearing the skin of a tennis report. Across forty-seven years in this trade, I have grown used to checking every number before it reaches the page. This time the first thing I had to check was the label itself.

People watch the match, I watch the match breathe. I still say that to the young reporters in the newsroom whenever they ask me where to begin. The breath of a match lives in very small places: the pause between two serves, the way a player tightens his grip on the handle when he is behind, the sound of shoes dragging on a hard court at the changeover, the opponent's breath sinking after a long rally. For forty-seven years I have learned to read the things that are never written into the official record. And for all those years I have also learned a simpler rule: if you do not know who said it, do not write it.

That rule sounds old-fashioned in an age when anything can be produced in seconds. But precisely because it is old-fashioned, it has become the thing most worth keeping. The file the editor sent me was not a joke. It was a symptom. And symptoms never appear alone.

I began checking the file the way I check a transfer story during the transfer window. First, the source. Of the eighteen information points, fifteen cited no provenance whatsoever. The remaining three pointed to a single market analyst from a brokerage firm. Every qualitative statement in the file converged on one name. That is the structure of a single point of failure: if that person is wrong once, the whole file collapses with him. In my trade, an analysis resting on one living source has never been considered enough. We want at least two independent sources for a fact, and three for an accusation. Fifteen out of eighteen points with no source means fifteen out of eighteen points that cannot be verified. They are not data. They are sentences repeated until they look like data.

Next, time. In the file, the federal funds rate was recorded at 3.75 per cent to 4.00 per cent, a figure belonging to a distant period. At the same moment, the ten-year Treasury yield was described as having just touched 5 per cent, the first time since October 2026. And the head of the Federal Reserve was given a name that matched no term I could trace. Three fragments from three different moments, forced into one sentence. In a match report, the equivalent would be a writer stating that a player led 6-4, 5-7, 7-5 when the real score was 6-4, 6-4. A reader has no way of catching it, unless they watched the match themselves.

Then the prices. Spot gold was recorded at 4,300.96 dollars an ounce and silver at 63.28 dollars an ounce. To anyone who has followed that market over the past decade, such figures belong to a scenario that has not happened, or to a hypothetical future, not to the period the file itself describes. October 2026 saw gold near the 2,000 dollar threshold. The gap between those two numbers is not a rounding error. It is the signature of an assembled text, in which fragments of data from different sources were mixed without anyone cross-checking them.

Finally, the language. Sentences such as gold is seen as a hedge against inflation, and it often loses its appeal when rates rise, are encyclopaedia-style explanations, correct but meaningless in a news report. A genuine report does not explain basic concepts to its reader; it tells them what changed since yesterday. Encyclopaedia-style explanation appearing densely is the fingerprint of a text generated from a template rather than from a telephone call.

I noted all of this in my notebook, the way I do after every closed training session. The conclusion was clear: this was a file misrouted within a data pipeline, or a synthetic product assigned the wrong subject domain. Either way the consequence is the same: if nobody reads it with the eye of someone who knows the subject, it goes straight onto the page under a tennis headline.

The mislabelling itself did not worry me as much as the question behind it. What allows a file like this to exist? And what allows it to travel this far before somebody stops it?

In recent years I have read hundreds of sports articles a week, partly for work and partly out of a habit forty-seven years have not broken. I have come to recognise a pattern. More and more pieces have every form of a sports report while completely lacking its spine. They have headlines, openings, statistics, quotations, conclusions. But there is nothing inside. No named source. No absolute date. No detail that only someone present could know. They resemble stadiums assembled from photographs: the stands are full of people but nobody is breathing.

During the transfer window, that pattern swells many times over. Thousands of lines a day about contracts never signed, fees never confirmed, release clauses nobody has read in the original. Clubs issue statements of denial simply to stop the story writing itself. Players post photographs to prove they are still somewhere. And the fans are submerged in a current without a bank, where credibility is measured by the number of shares rather than by who is speaking.

I do not hate the noise. Noise is part of this sport, and I have sat in the loudest stands in the world. What I will not accept is noise without provenance. A shout in the stands is still real, because a person opened their mouth. A line with no source has nobody behind it at all.

In 2026, when I was a training-ground observer for a Boston paper, I followed a young midfielder through forty-seven consecutive sessions. I recorded his movement paths, his reactions to every substitution decision, where he stood when the team lost the ball. The dataset ran to 212 pages. When the long-form series ran, it drew three thousand divided comments, and a senior male editor said publicly that a woman could not feel tactics.

I did not argue. I went home, opened the comments, read every line, and noted down the criticisms that had logic in them. From then on, every piece I wrote had two layers: the training-ground evidence and the voice of the community. The second layer is not decoration. It is my verification system. If a number I publish does not match what the audience saw with their own eyes, they will say so. And I will learn where I went wrong.

That is why the file irritated me so much. It had no self-correcting mechanism. Fifteen unsourced points cannot be caught out, because there is nothing to check them against. A statement with no source is a statement immune to being proven wrong. And in my trade, the thing immune to being proven wrong is the most dangerous thing there is.

The Moscow door opened, and I walked into the world of the fans. On the night of the 2026 quarter-final in Russia, I had registered for the mixed zone, but the system had failed and my name was not on the list. The security guard refused me outright. A group of Croatia supporters led by a man named Ivan recognised me from a small newsroom podcast. Ivan raised his scarf and shouted that they should let her in, that she writes for us. The guard looked at the crowd's eyes and opened the door.

I walked in and interviewed the Croatia head coach that night. He spoke about the pain of winning a penalty shootout. What I remember most is not the answer but the way a crowd of strangers used their unity to open a door for someone they had never met. Trust in my trade is built like that: one person at a time, one session at a time, across decades.

A name built over forty-seven years can be blurred in forty-seven seconds, when an unsourced line is attached to it. I have watched it happen to colleagues. A sentence cut from its context, a quote attributed to the wrong man, an analysis retitled for more clicks. The real author takes years to recover; the reader has already believed the wrong version.

When the stadium is empty, I hear the match more clearly. In 2026 the pandemic closed every training ground. Boston fell silent and the stadium stood bare. I was fifty-seven and, for the first time, saw my career at risk of breaking. I could not bear the feeling of uselessness, so I launched a live stream series, three sessions a week, inviting reserve players, kit staff and veteran supporters to tell their stories.

A twenty-one-year-old forward wept in the second session because he feared for his career. I received two thousand thank-you messages in the first month alone, with an average audience of eight thousand per episode, four times the readership of the articles I had written. That number taught me something I still carry: readers are not looking for perfection. They are looking for truth, and they will sit with anyone brave enough to tell it for a very long time.

If a livestream with eight thousand people can build trust, then those same eight thousand can be deceived by an assembled data file. That is why I treat verification not as a step in a process but as the soul of the craft.

In 2026 I was granted special accreditation for a group of female journalists over fifty-five at a World Cup. I got into the United States dressing room after a group-stage match that finished 1-0. The only scorer suffered a pelvic injury. Inside, players of different origins wiped their tears; some held each other without a word. The coach said quietly that today they were healing, not winning.

I did not switch on a recorder. I stayed silent and took in the atmosphere. They treated me as a witness, not a hunter. After that night I abandoned the habit of jotting down every passing sentence. I learned to write about the fragility of a dressing room in a language that respects boundaries, holding enough distance, like embracing the air without squeezing it.

But respecting boundaries does not mean softening the truth. In that dressing room I still had to record the exact score, the exact scorer, the exact injury, the exact date. Kindness and precision do not exclude one another. They are two hands of the same writer.

From all those years I draw three checks that any reader of sports news can apply within a minute.

The first is the named source. Not an anonymous source, not a vague group of sources, but an organisation, a person, a title. A sentence spoken by someone who can be challenged. A sentence spoken by nobody is right forever, and that is precisely the problem. When a report cannot be wrong, it cannot be right either.

The second is the absolute date. No yesterday, no today, no this week. A specific date always leaves a trace. It lets the reader cross-check against history and find contradictions the eye would otherwise miss, exactly like the contradiction between a 3.75 per cent rate and 4,300 dollar gold. A text afraid of dates is a text afraid of inspection.

The third is entity consistency. A person, a team, a tournament must be named the same way from beginning to end. In that file, the head of the US monetary authority was given a name matching no term. In a sports piece, the equivalent error is calling one team by two different names, or crediting a goal to someone who never came on.

Those three checks sound simple. But when I gave them to the young reporters in the newsroom, many said they had never been taught them. They were taught to write fast, to write attractively, to write for clicks. They were not taught that an unsourced number is a number that does not exist.

The core of the problem sits here. The problem is not that machines produce false content. The problem is that people read it without asking for the source.

For many years my industry built its credibility by naming its speakers. A reporter at the scene is trusted because he was there, because he made the call, because he asked three times. That credibility does not live in the pen; it lives in the process. When the process is skipped, the pen goes with it. And when an assembled data file can wear the skin of a tennis analysis and pass through many hands without anyone stopping it, the problem is no longer at the end of the pipe. It is at the beginning, where nobody asks the gatekeeper what exactly he is guarding.

I will say plainly what many colleagues avoid saying during the transfer window. Most of what we read in these weeks is not news. It is hypothesis written in the present tense. A fee never confirmed, a release clause never quoted in full, an agent's move never verified. What genuinely has value is not the name mentioned most. It is the structure: who pays, how the clause is shaped, how the wage bill is affected, and who benefits if this story spreads.

That logic holds far beyond the transfer market. It holds for any report that passes through my hands.

A tactic never dies; it only waits for someone who understands it. A source is the same. A good source can lie dormant for years, waiting for someone to return with the right question. A bad source dies the first time it is challenged, and it dies so fast the writer does not notice what he has lost.

What makes me optimistic about this incident is one small detail. The file was stopped. It was never published. It was stopped by a human being, on a cold Boston morning, when she opened it and saw there was no tennis player inside.

A machine trained to detect synthetic content would see a text with a valid structure, fluent sentences, complete figures. It would score it highly. What caught the anomaly was not an algorithm. It was the memory of a person who has lived long enough to know that gold has never touched 4,300 dollars, and that the Federal Reserve has never had a chair with that name.

A machine cannot detect a lie. A person can, because a person has memory. This is the counter-intuitive point few want to hear in an age when everything is said to be automatable. The more content is generated, the more the careful reader is worth. The more data flows through the pipe, the more the person who knows how to stop is worth.

During the transfer window I see newsrooms chasing speed. Whoever publishes first wins. Whoever is faster takes it. But fans do not stay with the fastest. They stay with the most accurate. I have seen it across forty-seven years. Writers who rise and vanish within a season. Writers who move slowly and are still here, their audiences still waiting.

Observation is not standing outside; it is standing in the right place. I stand on the training ground, not in the press room. I stand where there is the sound of the ball, where there is sweat, where there is somebody breathing. And precisely because I stand in the right place, I know at once when something does not belong there.

A file about gold labelled as tennis is a small incident. But it opens a larger question for the whole industry: are we checking the input, or only the output? Are we raising writers who ask for the source, or only systems that push the story? Are we building a newsroom, or building a pipe?

I do not have a complete answer. At sixty-three I have learned that some questions do not need to be answered at once. They need to be asked again every day.

What I will be watching in the coming weeks is how newsrooms handle files like this one. If they publish the process, if they name the verifier, if they use absolute dates, that is a good signal. If they stay silent, delete the piece, offer a vague apology and carry on, that is a bad signal. Silence is a label, and it is always stuck in the worst place.

I am old, but the pulse of the ball never grows old. In the same way, the principle of verification does not grow old. It is only forgotten, and then picked up again, by a new generation of writers brave enough to ask a question nobody wants to hear: who said this, and why should I believe it?

That question is not new. It is as old as the trade. And perhaps precisely because it is old, it is the only thing still standing when every label can be changed in an instant.