The Empty Report and Football's Void-Filling Machine

Two forty in the morning in Shanghai. The file my colleague sent from Europe opens with nine analytical sections. Section one, tactics: insufficient information to assess. Section two, club finance: insufficient information. Section three, results and public opinion: insufficient information. By section nine the sentence is identical. The original article title reads N/A. The source reads N/A. The list of facts is empty.

The only thing in that file with any substance is a warning line: feed an empty input into the production pipeline, and what comes out the other end will still be a polished article, complete with a player's name, a fee, a percentage, and a source "close to the club."

I read that line three times. Then I thought about the transfer market.

Those nine empty sections look like nine chairs in a meeting room that still has to be held. Nobody wants to sit in them, but the room is booked, the water is poured, the microphones are on. In sport, an empty chair always gets filled.

The machine needs gaps to run

A mid-sized football site in Asia publishes forty to sixty articles a day. During the summer transfer window that number rises by about half. Each article needs a headline, an opening line, a character, a metric. Meanwhile the match itself supplies genuine raw material for roughly two hours a week per club. For the rest of the week, the machine runs on something else.

It runs on the gap between what happened and what might happen.

Inside a match, the gap is narrow. There is a ball, a referee, twelve cameras, footage you can rewind. In the transfer market the gap is almost limitless: a club with a need, a player who is unhappy, an agent who has to negotiate. Those three facts are enough to generate two hundred stories in ten days, every one of them logically coherent and none of them verifiable until the window shuts.

This is where the story usually gets misread. People say the transfer press makes things up. Most transfer stories do not fabricate events; they are true in the future tense and read in the present. "The club is in talks" is a sentence that can barely be wrong, because talks leave no public trace. It only becomes wrong when the reader converts it into "the player signs this week."

I have spoken with three editors at two different platforms. All three described the same pressure: the daily article quota is set first, the facts arrive later. If the facts do not arrive, the slot still has to be filled, because an empty slot on screen is an operations failure, whereas an article with wrong content is tomorrow's problem.

Before 2026, the cost of filling a slot was a reporter making phone calls. Since roughly 2026, that cost is close to zero. When text generation became cheap, the number of slots a newsroom dares to open increased rather than decreased. The pipeline does not shrink to match verification capacity. It expands to match production capacity.

What a mispronounced name taught me

In July 2026 I took over as lead commentator for a digital sports channel in Shanghai. During the city derby I got Hulk's name wrong three times in the first half: once as Rolf, once as Hulk Hogan. The forums tore it apart all night.

I did not go on air to explain myself. I opened the match footage, counted every touch, every pass, every shot by the Brazilian forward, built a spreadsheet, and cross-checked it against the movement of the opposing back line. Three days later I had a dataset nobody else in the studio had.

That episode taught me something very specific: when I have no data, my mouth will manufacture data. Not because I want to lie. Because three seconds of silence on live television costs far more than a sentence that might be wrong.

Since then I have written in the reverse order of most colleagues: conclusion last, facts first. Open with a count. Build the body on a comparison. Close with a verifiable number, or with the sentence "my sample is not big enough."

That last sentence is the hardest one in my trade to write. It does not go on the front page. It has no compelling headline. It does not get shared.

When a metric manufactures its own conclusion

In May 2026 global football stopped. The channel I worked with cut roughly forty percent of its staff, and my live commentary contract ended. I sat at home, downloaded StatsBomb's open event data, and wrote Python code to find Liverpool's pressing pattern in the 2026-2026 season. When the Bundesliga returned in mid-May, I tried predicting results with expected goals (xG) and sprint counts. I was right in eleven of fourteen matches.

Eleven out of fourteen sounds excellent in print. The three misses were the valuable part. Two of those three had near-balanced xG, which means the model was not wrong, it was saying the match could go either way. Readers want one way.

Based on my experience tracking matches through that period, I settled on a professional rule: when there is no data to answer the question, the only honest answer is to describe the gap, not to fill it with a confidently voiced prediction.

A licensed betting platform paid me 1,200 US dollars a month to write a weekly tactics bulletin, capped at 600 words, always ending with a verifiable number. I used no ornate language there. I learned to write like an internal report: blunt, short, open to challenge. It was the best environment I have had for dropping the habit of filling gaps.

There is a common misunderstanding about metrics. People assume xG is a prophecy. It is not. It is a division: total chance quality divided by number of chances. A team that wins 3-0 with an xG of 1.2 won through something else, whether that was the goalkeeper, counterattacks, or luck. If the writer cannot name that something else, they are presenting numbers as decoration, and the data becomes a new shell around the same old gap.

Transfers: the story of a buyer choosing the wrong reason to be right

Transfers, in the end, are the story of a buyer choosing the wrong reason to be right.

In a transfer window there are usually only four real facts: how long the contract runs, the current wage, the age, and the injury status. Those four facts cannot sustain daily coverage. So the rest is generated by motive: the player wants minutes, the agent wants a commission, the club wants leverage in a different deal. Motive cannot be verified, and because it cannot be verified it is never fully refuted.

For the reader, this mechanism produces a very pleasant experience. You follow a deal for six weeks. The deal collapses. You do not feel cheated, because for six weeks you had a story to follow. The emotional cost is near zero and the entertainment value was fully consumed.

Financially, the transfer market is one of the few markets where wrong information carries no direct penalty. Nobody refunds a bad report. Nobody claims damages for a transfer that never happened. The outlet that gets it wrong still gets the clicks, the agent still gets leverage, and the club can still tell its supporters it was trying.

Verification, by contrast, has a very clear price: time. In a twenty-four-hour news cycle, time is scarcer than money. Spending two hours calling three sources means two hours not publishing. Publishing an unverified item takes four minutes. That is the whole equation, and it solves very clearly in the wrong direction.

The Vietnamese cycle

In Vietnam this machine runs at a different rhythm. The domestic transfer market is short, thinly funded, and lightly intermediated, so rumours have nowhere to hide for long: a player moving between clubs is often common knowledge before the paperwork is done.

The bigger gap sits at national-team level. Since the 2026 AFC U23 Championship and the AFF Cup title the same year, every national-team camp loads the machine with material for weeks, while the actual matches number only a few. That distance gets filled with projected line-ups, unconfirmed injury news, and arguments over who was called up and who was left out. I follow those camps as a data person, and the striking thing is that the same set of numbers can support two opposite conclusions, simply by changing the order of presentation.

One thing I notice about Vietnam deserves more credit than it gets elsewhere in the region: the supporters remember for a very long time. A squad error can be brought up for years. In a market with that kind of collective memory, the price of filling a gap with a false report is not cheap. It is simply paid late.

World Cup 2026 and the fear of being forgotten

World Cup 2026 did not begin with the ball. It began with the fear of being forgotten.

That truth appeared to me in Moscow. The whole city was selling the same product: memories that had not happened yet. Tickets, shirts, flags, bar seats — all of it a bet on which team would make people remember them.

At the final on 15 July 2026 at Luzhniki Stadium, I sat in the commentary position with a dataset I had built in my head. In the eighteenth minute, Griezmann stood over a free kick angled in from the left. I had reviewed seven similar situations of his earlier in the tournament, and all seven times the ball curled into the zone between the penalty spot and the post. I said it on air before it happened: the ball would land in that zone, Mandžukić would clear it and put it into his own net.

It happened exactly that way. France beat Croatia 4-2. Mandžukić turned the ball into his own net in the 18th minute from Griezmann's free kick, then pulled one back in the 69th, one player scoring for both sides in a World Cup final.

A colleague asked which monitor I had been watching. I was not watching a monitor. I was watching frequency.

Here is the part that rarely gets told. Before that prediction I had four other free-kick situations where I said nothing at all, because the sample was too small. Nobody replays those four silences on television. What gets replayed is the one time I was right.

My way of handling it is simple. Every time a prediction lands, I also log the ones I did not dare make. My spreadsheet at home has three columns: said right, said wrong, did not say. The third column is always the longest. If a data person has no third column, they are selling predictions rather than doing analysis.

Media rights: a long contract for a product that does not exist yet

Media rights are a marriage nobody enjoys, but everybody waits to see the paperwork.

The structure is almost always the same. A league signs a five- or ten-year deal with a broadcaster or a platform, based on the assumption that viewership will hold or grow. At the moment of signing, the product does not exist: nobody knows which teams will be strong, which stars will emerge, whether the broadcast slot will shift. The contract is written over a gap, and that gap is then resold to advertisers in confident language.

This explains why most promotional material around a rights deal is not about football at all. It is about projected reach, market penetration, brand value. Football facts only appear once the ball is rolling; before that, what is being sold is anticipation.

I have sat in meetings where a single metric, average minutes watched, determined the entire value of a product package. When that metric falls, the first response is not better content but more content. More columns, more bulletins, more statistics. And the more slots get opened, the more slots have to be filled.

The reader is not a victim

The easiest way to tell this story is to cast the reader as the victim of a fabrication industry. I do not believe that framing. The reader is a co-author.

The first time I got it wrong on a big screen, the audience forgot. I did not.

Most audiences do not forget because they are credulous. They forget because the next story is already available and more entertaining. During a transfer window, supporters do not lack information; they have too much of it and too little way to sort it. When a rumour gets shared, the person sharing it usually is not asserting it is true. They are playing a collective game: imagining together a line-up that does not exist yet.

There is nothing wrong with that game. The problem is that newsrooms read that signal as a purchase order. Supporters imagine for entertainment; content pipelines imagine to fill pages. Same behaviour, two different purposes, and only one of those purposes has ever been audited.

So the fix is not to call for less news. It is to teach readers to read the structure of a story. A report has four parts: the event that happened, the source speaking, the level of verification, and the expiry date. Remove the last three and you have a story. Add the last three and you have a fact, less entertaining but usable.

Closing

I stand between revenue and emotion, and I have learned that whoever holds both is the one who wins.

That empty report I read at two forty in the morning is a fairly accurate picture of most sports content we consume: a complete frame, an empty input, and a person in the middle forced to choose between silence and invention.

Over the next ten years, as the volume of automatically generated content far exceeds the volume of real events, the scarcest thing on the market will be one plain question: who said it, when did they say it, and where can it be checked. If you read only one kind of information this season, choose the kind that carries a date and a named source. The rest is noise, beautifully presented.