Empty Analysis: When a Blank Spreadsheet Still Produces Football Conclusions

On the night of July 14, after the final whistle of a semifinal at a major tournament, I opened my spreadsheet as usual. Three tabs, forty columns, not a single cell with a number. The twelve-page analysis an associate had sent me that morning — heat maps, pressing-direction arrows, a scoreline forecast — had been built from an empty input file. The writer had not invented numbers. The writer had simply not checked the source. And so a smooth conclusion was born out of nothing, and it spread faster than any report carrying footnotes.

That was the moment I realised the greatest danger in sports journalism is not the fake news deliberately manufactured. It lies in conclusions generated unconsciously from data that does not exist. I call that phenomenon the empty analysis.

A major-tournament season is a season of compressed emotion. Fans are swept up in flags and stories, broadcasters race hour by hour, and every newsroom feels the pressure to publish before its rivals. In that rush, an analysis that looks complete becomes a more valuable commodity than a slow truth. The fast writer is rewarded; the slow writer is called lazy.

I have worked in this trade for ten years, most of it in Lyon, reporting on football for the French market. I learned to open a spreadsheet during the interview itself, to ask about contract structure before asking about form. Experience taught me that most errors do not come from malice. They come from a process lacking a control gate, where an empty dataset still passes through and dresses itself in the appearance of certainty.

In recent years, the flow of information in football has accelerated beyond the speed of verification. A transfer rumour can travel from an anonymous account to a front page within hours. A tactical metric can be quoted again and again while nobody can trace its origin. When speed beats accuracy, the empty analysis becomes the default, and the person who checks sources becomes an obstacle.

Picture the journey of an empty analysis. Step one, an unverified source supplies a single information point. Step two, an editor turns it into a dataset. Step three, an empty input file is still pushed through to the interpretation stage. Step four, the interpretation stage has nothing to say but must still produce a product, so it speculates. Step five, the speculation is presented in a confident voice. The result is a report that looks utterly credible while its information value is zero.

The fatal point is step three. A good system must have a gate: if the input data is empty, the process stops and flags an error. A good newsroom should work the same way. When there is no data, the most honest answer is to admit there is no data — not to fill the gap with plausible-sounding guesswork.

I once fell into this very trap, in the opposite direction. In June 2026, while a first-year journalism student, I spent an entire month rewatching video of forty-eight World Cup group-stage matches in Russia. After the France–Australia match on June 16, I noticed that nine of twelve penalties in matches with odds differentials above twenty-five percent favoured the underdog. I hand-recorded one thousand two hundred and forty-seven refereeing decisions and cross-checked them against open data from Opta and five Asian bookmakers. One referee made seventy-eight percent of his foul calls in favour of the weaker team, two point five standard deviations from the mean. The piece published on my faculty blog was taken down after forty-eight hours, but I kept the entire spreadsheet. The lesson I took from the 2026 World Cup: referees can read a spreadsheet too. Since then, I never reach a conclusion first and hunt for evidence afterwards.

Three years later, that lesson became a method. In April 2026, when Ligue 1 was cancelled mid-season because of the pandemic, Olympique Lyonnais published a one-hundred-and-twelve-page emergency financial report on the Euronext exchange. I spent three weeks checking every line item against DNCG records. COVID-19 shut every stadium in the world, but the holes in a financial report never social-distance. I found a seven point eight million euro brokerage fee transferred to a company in Luxembourg incorporated two months earlier, whose director shared a name with the agent of the substitute player number twenty-four. A textbook shell-company structure. The data-analysis skill from 2026 let me spot it, and a financial-investigation lecturer guided me in sending the proposal to the Mediapart newsroom.

In September 2026, I received a scanned contract worth forty-five million euro between a FIFA subsidiary and Qatar Energy. An access-fee clause of three point two million euro flowed into an account in the Bahamas, where the receiving company had a registered address but no physical office. I confirmed the company had been set up two months before the signing. Three harmless data points stitched together form a map of money flowing into a village with no football pitch. The four-thousand-eight-hundred-word investigation ran on November 15, 2026. FIFA opened an internal audit and admitted that sixty percent of the costs had no verifiable documentation.

All three cases share one logic: when the data is empty, the conclusion must be empty too. People call me a sceptic; I call myself someone who reads the books behind the pitch. The difference between the risk of wrongdoing and the evidence of wrongdoing is my entire trade. A suspicious payment is only a risk until I hold the contract, the date, and the final beneficiary.

Here I must argue against myself. There is a legitimate part to the habit of fast analysis I have just criticised. Football is a sport of small samples, and the intuition of someone who has watched thousands of matches has its own value. Sometimes empty data is itself the story: a team that cannot muster a single shot in the first half, and that emptiness is the clearest proof of a collapsed system. If I absolutised the principle, I would miss the things only a human eye can see.

The problem is not with inference. The problem is with presenting inference as though it were data. A commentator saying this team presses poorly because that is how it feels is normal. A chart saying the same thing with no source is something else entirely. The same sentence, two different responsibilities.

I have also erred by trusting small samples too much, in the opposite direction. Once I nearly published a piece on an amateur team that reached a final, arguing that their system was succeeding. I paused for a week. On closer inspection, that run came from a favourable draw and exactly one explosive match. A lucky draw is not a system. The transfer market never lies if you bother to read the brokerage-fee column instead of the player-price column. But to read that column, you must accept that you may have to wait.

I do not demand that every article come with a methodology appendix. I demand transparency about the level of certainty. A guess should be called a guess. A trend based on three matches should be labelled three matches. When writers are honest about where they stand on the evidence scale, readers can decide for themselves how much to believe. That is the difference between an opinion and an accidental lie.

I once saw an empty analysis do real damage. A young player was branded a failure after a match in which the dataset used to judge him had never been loaded. He lost his starting place, lost a small sponsorship deal, and lost faith in himself. No one in that production chain meant him harm. They simply did not stop to check.

Every major-tournament season will produce thousands of empty analyses. They look flawless, they spread fast, and they leave a faint residue in how fans understand the game. The only way to resist is to build control gates at every stage: which source, which date, who verified it, and what would make this conclusion collapse.

I still open my spreadsheet every night. Most of the cells are empty. But those empty cells remind me that the cheapest conclusion is the one that needs no evidence. If one day you read an analysis of a match without a single sourced figure, perhaps the writer is telling you about a dataset that never existed, rather than about the match. This industry only improves when readers begin to demand what writers should have supplied from the start.