Anatomy of Football's Empty-Report Machine: A Sport Written by Template

The Empty Dossier

The document ran to four thousand three hundred and eighty-two words. Nine sections. Forty-two tables. Every section had a heading, a skeleton, an orderly sequence. And every section closed with the same sentence: insufficient information to assess.

I read all of it, first line to last. When I folded it shut, I had learned nothing about football.

What stopped me was not the emptiness. It was its shape. A machine had taken an empty input, run nine full processing cycles, and returned an output that looked exactly like a real one: same typeface, same confidence labels, same three-tier structure of conclusion, evidence, and hidden information. Had I skimmed it, I could have cited it as a source.

Thirty-two years earlier, I met the mirror image of the same problem. On 9 July 2026, at Foxboro, Massachusetts, I sat in the fourth row of the press room after the quarter-final between Italy and Spain. Italy won two-one. I was twenty-eight, newly switched from competitive swimming to journalism, and the only woman in the room. I asked the Spain coach about the positional error of the number four defender on the second conceded goal. He smirked and told me I did not understand football.

I did not argue. That night I borrowed a colleague's tape, rewound it thirty times in my hotel room, and found that the real error was not the number four defender. It was the central midfielder, who failed to drop at the right moment. I wrote two thousand words and sent them to a small sports paper in Barcelona. Three weeks later, that same coach called to concede the point.

In 2026, my problem was that nobody would give me information. In 2026, the problem has inverted: I am given far too much that resembles information.

They shut the door of the press room in 2026; three decades later I opened the files wide. But opening the files no longer means finding something inside. It means checking whether anything real remains.

Context: the content supply chain and the price of speed

For the first two decades of my career, I worked a news flow with two clear ends. One end was the club, the other the newsroom. Between them stood a reporter, a notebook, and a telephone. To learn a contract fee, I called the intermediary. To learn why a defender was pulled out of position, I borrowed a tape.

That flow has been replaced by a chain with at least seven links: the club's communications office, the player's agency, the data company, the betting company, the aggregation platform, the newsroom, and the automated text tools. Every link has an incentive to publish before the next link does. Nobody is paid to stop.

The arithmetic consequence is blunt. One European summer transfer window produces tens of thousands of content items. That volume far exceeds the number of real events that can occur in the same period. The gap has to be filled with something: speculation, recycling, template.

I call that gap the production gap. And the production gap is always filled with the cheapest material available.

What matters is that the reward mechanism of digital platforms makes no distinction between dense and empty content. It counts views, clicks, dwell time. A vague headline sometimes holds a reader longer than a precise one, because the reader has to scroll to find out what is inside. They scroll, find nothing, and scroll on. The metrics improve. The content thins.

The 2026 search algorithm carries a requirement called information gain: every article must give the reader at least one thing they did not know. The requirement is right in principle. It has also created a trade: manufacturing the appearance of information gain without the substance. Add unattributed numbers, add technical vocabulary, add confidence labels. The shell thickens. The core stays hollow.

In Spain, my trade has an old convention that widens the gap further: transfer fees are not required to be disclosed. A club may write "undisclosed fee" in its official statement. That means that in many deals, the only party publishing information in the entire system is the one party not accountable for the number.

In China, where I was born and began writing in 2026 at local radio stations, the structure differs but the mechanism is identical. Transfer news passes through three layers: the agent layer, the local media layer, the aggregation layer. Each layer strips out part of the sourcing, and the stripped part is never recorded. After three layers, the reader receives a complete story with no anchor left in reality.

The production gap crosses borders. It merely wears different shirts in different markets.

Two sentences, two levels of information

To separate information from the appearance of information, I use a simple test.

First sentence: this team is in good form.

Second sentence: over the last three matches, this team's PPDA has fallen from 11.4 to 8.2, meaning they allow opponents fewer passes per defensive action, which means markedly higher pressing; their ball-recovery rate in the opponent's third has risen from 6.1 to 9.8 per match.

PPDA is the number of passes an opponent is allowed before your side makes a defensive action. The lower the figure, the earlier and more aggressively the team engages. That is a dry definition, but it does something the adjective "good" cannot do: it allows me to be wrong.

The first sentence cannot be wrong, because it says nothing. The second can be wrong, and that is exactly why it has value.

Based on my experience of watching matches, I draw one rule: information is what remains after you subtract everything that could be said about any team.

"This team has great fighting spirit" could be said of all twenty La Liga clubs. "This team switched from a back four to a back three in the 58th minute, and afterwards opponent entries into the box fell from seven to two" can only be said of one team, in one match, at one moment. The second is information. The first is noise with polish.

Six markers of an empty text

From my own archive, I extract six markers of a piece of sports writing that is formally complete and substantively hollow. They rarely appear alone. When three or more appear together, the probability that the text contains no verifiable information is very high.

Marker one: no entity is fully named. The text speaks of "a big club", "a quality player", "an experienced coach". No proper noun means no accountability. No accountability means no verification.

Marker two: no number arrives with a unit and a date. A number without a unit is decoration. "Millions of euros" is not data. "12 million euros, announced 14 July 2026" is data.

Marker three: no primary source. Nobody said it. No document records it. The claim exists only as "understood to be", "believed to be", "according to a source close to".

Marker four: no statement can be proven false. This is the most important marker and the hardest to spot. A text can be long, detailed, and correctly terminologised while containing not one falsifiable sentence.

Marker five: complete structure, empty semantics. This is exactly the dossier I received. Nine sections. Forty-two tables. Every cell filled with a polite negation.

Marker six: confidence labels unattached to evidence. Writing "high confidence" beside a conclusion with no data underneath is more dangerous than writing nothing. It borrows the authority of method to underwrite emptiness.

These six markers are not a machine speciality. I have found all six in human-written journalism, in famous newsrooms, long before machines could write.

Case one: the bot network at Girona

In 2026, Girona were promoted to La Liga for the first time. I was fifty-one, filed among the veterans, and forced to learn how to read social-media data so as not to be left behind.

The club sold a twenty-two-year-old defender to an English side for a reported 25 million euros. That figure was roughly ten times the valuation on the analytics site I used. I did not believe a gap that large was random.

I downloaded every interaction on the player's social account over six months. About forty thousand. I rebuilt the list of interacting accounts, cross-checking activity times, comment patterns, account creation dates. Around twelve thousand accounts shared one technical fingerprint: they had been created through the same application-programming key. In other words, they ran on one system.

Girona inflated the player's value with a bot network; the real value sat in the server logs.

I traced a contract between the club president and a media company. That company was run by his younger brother. The service purchased was "image management and expansion of digital reach". Digital reach was expanded with fake accounts. The player's value was lifted by digital reach.

My three-thousand-five-hundred-word investigation was ignored by the federation. But by 2026, UEFA began requiring that player valuations rest on real indices.

The point here is not the fraud. The point is the mechanism. Twelve thousand fake accounts produced a formally complete text: numbers, charts, trends, comparables. That text contained not a single unit of true information. It contained only repetition.

A bot network and an empty article run on the same principle: generate enough volume that volume is mistaken for value.

I have since appended a method section to every investigation I write: how the data was collected, with what tools, at what level of source reliability. An article that does not state its method denies the reader the right to check it.

Case two: the index in the Betis file

In 2026, I was assigned to transfer-contract investigations after making my name on video analysis. In the summer window, Real Betis spent 12 million euros on a Brazilian winger named Márcio, then playing in the Brazilian third tier and known to almost nobody.

A deal like that produces two kinds of text. The first is long, excited, describing potential, quoting insiders who say the player will explode. The second is a blood-test table.

I chose the second.

Over three consecutive years I collected Márcio's blood-test results. His haematocrit, the proportion of blood volume taken up by red cells, rose from 43 per cent to 52 per cent within eight months. That is a rise outside the normal variation band of a healthy athlete not living at altitude.

I drew no conclusion. I recorded the numerical anomaly, with dates, sources, and reference ranges. Betis hid the doping in a contract annex; I read backwards through every page to find it. By 2026, Márcio was banned for two years for erythropoietin, a hormone that stimulates red-cell production.

My newsroom at the time threatened to fire me for touching a big club. Events later showed the direction of the investigation was right, but what I keep from it is not the victory.

A number with a date and a source carries more information than five thousand words of commentary about potential. I do not trust transfer valuations. I trust the numbers that were struck out.

Case three: the two words "undisclosed fee"

Now look at what appears daily on club homepages.

A typical La Liga transfer statement has a fixed structure: a headline with club and player; a first paragraph giving position and nationality; a second quoting the president or sporting director; a third on contract length; and a closing line stating the fee is undisclosed.

The whole statement may run four hundred words. Of those, the words that actually carry verifiable information usually number fewer than twenty: name, date of birth, nationality, contract length, signing date. The rest is template.

I do not object to undisclosed fees on commercial grounds. I object to an entire industry's information system being built on a deliberate gap, after which everyone turns to arguing over figures estimated by third parties.

When the official number does not exist, the unofficial number is treated as official. That is how a gap becomes a market.

I have tracked hundreds of such statements. In my own file, the share of La Liga statements that state a fee is far lower than in the English Premier League, where clubs listed on the stock exchange must disclose the figure to shareholders. The difference is not cultural. It is a legal obligation.

That yields an actionable conclusion: the information gap is not a natural property of football. It is the product of a specific regulatory framework, and frameworks can be amended.

Method: thirty minutes to check a text

I close my investigations with a method section so readers can repeat the work. This is the thirty-minute routine I apply to any piece of sports writing.

The first ten minutes go to counting. I count fully named entities. I count numbers with units and dates. I count sentences that can be proven false. Those three figures divided by total words give a metric I call anchor density. A text with anchor density below one per cent is almost always empty.

The next ten minutes go to sourcing. For each claim I ask: who first said this, and on what date. If the chain stops at another article that also cites no source, I mark the text as having no primary origin.

The last ten minutes go to reading backwards. I start from the annex, the footnote, the closing note. That is where real information tends to hide, because it is the least-read place. In the Betis file, the suspicious clause was not on the front page. It was in an annex.

The routine requires no special equipment. It requires time, and time is precisely what the modern content chain denies anyone.

The paradox of information gain

I mentioned the information-gain requirement above. Let us go into its paradox.

To prove an article delivers information gain, the writer must show what is new. But in football, most of what is new sits in very small places: a positional switch in the 58th minute, a pressing metric falling after a midfielder is withdrawn, an automatic renewal clause in the third year of a contract.

Those details demand time, tape, and sometimes documents. They do not fit the publishing rhythm.

So fake information gain is manufactured in three ways. The first is more numbers without more understanding: adding many figures unrelated to the argument. The second is more terminology without more analysis: using technical language to cover a hole in the reasoning. The third is more length without more density: writing three thousand words around a single idea.

Deployed together, the three make a text impressive. And information density per thousand words collapses towards zero.

That is the paradox: a tool designed to reward information becomes an incentive to manufacture the appearance of information.

When the machine fails silently

Back to the dossier from the opening.

Its most notable feature is not the emptiness. Its most notable feature is that the process which produced it reported success.

Empty input. Formally complete output. No error signal in between. If the final reader does not check for themselves, the system keeps running on the belief that it has just produced a valuable report.

I call this a silent failure. In my trade it is more dangerous than an obvious mistake, because an obvious mistake gets caught. A silent failure goes straight into the archive and becomes a reference for later work.

I have met silent failure in many places. In a payroll table where an empty cell was read as no bonus, when the contract specified a bonus. In an injury tracker where a two-month absence was logged as short-term. In a scouting report where the "weaknesses" field was left blank, and the blank was read as "no significant weaknesses".

An empty cell is not zero. An empty cell is an unanswered question. Treating the two as the same is the most expensive error in sports data analysis.

Here is what I want the reader to carry away: an empty report is not dangerous because it is empty. It is dangerous because it looks full.

I have also seen the reverse, and it is more memorable. In July 2026, in Saint Petersburg, during the semi-final between France and Belgium, security stopped me at the commentators' gate because the accreditation carried a woman's name. They said the area was for men. I did not raise my voice. I bought a ticket in the stands, hid a small camera in my coat pocket, and recorded the entire match.

In the 67th minute I captured an incident the referee missed. The clip was fifteen seconds long. I sent it to a refereeing body I knew. The result did not change, but the clip was used to train referees for the 2026 European Championship.

Fifteen seconds of footage carried more information than all the commentary written about that match in the following forty-eight hours.

The contrarian angle: the reasonable case for emptiness

There is another reading I am obliged to acknowledge, because I have pursued it for forty years.

A report that declares it lacks sufficient information is more honest than one that invents nine conclusions. Forced to choose between the two documents, I always choose the one brave enough to say it does not know.

The problem is not emptiness. The problem is that this industry will not let emptiness exist naked. It forces emptiness to wear clothes. It makes people fill every cell, and a cell filled with a meaningless sentence is still counted better than a cell left blank.

When a system rewards filling cells rather than filling them correctly, it will receive a great many filled cells.

I also want to speak plainly about something else. There is a wave of concern that automated text tools are flooding football with low-quality content. The concern is not wrong, but it aims at the wrong target.

Before deepfakes came transfer rumours; both are tricks that deserve exposure. Humans have done this work very well for decades.

The tools only accelerate a process that already existed. When I read an unsourced transfer item written by a named human being, I find exactly the six markers I find in a machine-generated file: no clear entity, no dated number, no primary source, no falsifiable claim, complete structure, confidence labels without evidence.

The wave of concern about machines quietly conceals human responsibility. It lets a newsroom say the fault lies with the tool, when the fault lies with an editing process dismantled long ago.

Emptiness is not the enemy. The enemy is the counterfeit fullness.

And to the extent I dare say it, the reader has a part too. A market only produces what it can sell. As long as a vague headline outsells a precise number, the chain keeps running.

Who benefits from the gap

Gaps in football information do not appear by themselves. They have beneficiaries, and the beneficiaries always sit on the side with more information.

When a transfer fee is undisclosed, the beneficiary is the party that knows the real number. When test indices go uncross-checked, the beneficiary is the party that knows the index is drifting. When a report returns nine empty sections without raising an error flag, the beneficiary is the party that commissioned the report in order to have something to present.

I have no right to ask the public to read more slowly. I have the right to ask information producers to be accountable for the information density of their product.

A number must arrive with a unit, a date, and a source. A conclusion must be capable of being proven false. An empty cell must be recorded as empty, not read as zero.

From the 2026 press room to the automated data files of 2026, power has only changed shirts. Back then they stayed silent so that I would have no information. Now they speak at length so that I cannot tell information apart.

And my job, at sixty, remains the old job: open the files wide, read backwards through every page, and trust only the numbers that were struck out.