An Empty Spreadsheet in Marseille: The Discipline of Refusing to Conclude

Three in the morning, August, Marseille. The mistral slips through a narrow window, and on my screen there is a spreadsheet with exactly one header row. Column A says "Minute." Column B says "Player." Column C says "Action type." Below that, white space runs to the bottom of the page.

I was waiting for the feed from a qualifying match. The positional tracking system at the stadium had lost sync, and the data provider sent a short notice: the entire event package for the second half would be re-released after manual verification. In my inbox, the editor in Paris wrote: "Need 1,500 words before eight. You watched the game. Just write what you felt."

I answered in four lines. I watched the match. I have no data. I have verified nothing. And I will not write it.

The next morning, a younger colleague filed on time. The piece flowed, it had a climax, a paragraph about "fighting spirit" and another about "the character of the away side." It was shared several thousand times. Three days later, when the data package was re-released, the only goal of the match came from a corner that the assistant referee had wrongly allowed, and the team praised for its "spirit" had in fact produced two shots on target across ninety minutes.

I do not tell this story to mock a colleague. I tell it because it is my trade, it repeats a few times a year, and it describes the condition of football analysis in Europe today: an industry that produces conclusions faster than it collects evidence.

When a spreadsheet is empty, the only professionally correct conclusion is an empty one — and being willing to publish it is a skill, not a failure.

Context: from a strange column of numbers to an industry

In the summer of 2026, I learned to trust something nobody had named yet: xG. I was 57 that year, working as a transfer market administrator in Marseille. When Opta first published xG tables for Ligue 1, I did not believe them immediately. I hand-logged 1,204 shots from 20 teams across the first half of the 2026-18 season, matched each one against actual goals, and arrived at a correlation of 0.84. That number was enough to build my own striker valuation dataset. Colleagues said my reaction was slow. I need verification before I use anything, and I still think that is the only way a metric survives inside your own head.

Nine years later, xG has travelled from a mysterious column of numbers into the shared language of television, of transfer meetings, of arguments in cafés. Alongside it came PPDA, player valuation models, positional tracking at twenty-five frames per second. A modern football match generates millions of data points, and the content industry built around it has swollen in proportion.

But there is an inverse equation that few people mention. The volume of data grows exponentially; the time available to verify it stays almost still. During a major tournament, that gap becomes brutal. A match ends at eleven at night, a bulletin goes out at six in the morning, and between those two points hundreds of newsrooms each demand a different angle. The pressure creates a professional habit: fill the gap with language when the data has not arrived.

I have wondered why readers still consume analysis built on sand. The answer lies in the structure of attention. Someone watching a football match does not remember precisely whether a team held 54 or 57 percent of the ball. They remember a sprint, a roar from the stand, a single moment. Writing that speaks to emotional memory is always easier to read than writing that speaks to a table. The problem appears only when such writing becomes the basis for a real decision: a contract, an academy scholarship, a renewal offer.

That is the point at which the white space in my spreadsheet becomes a serious matter.

Nine analytical dimensions and the anchor each one needs

I work with a nine-dimension framework: tactics, finance and the transfer market, results and the opinion cycle, league landscape, rules and governance, management and dressing room, risk profile, media narrative, and industry transmission. It sounds heavy, but the operating principle is simple: each dimension needs at least one concrete anchor. Without an anchor, that dimension returns a null.

The tactical dimension needs a formation, a player, or a metric. In the 2026 World Cup semi-final between Croatia and England, I sat counting PPDA for both sides. Croatia allowed England only 8.2 passes per defensive action, while England allowed Croatia 12.5. I wrote a note predicting Croatia would win through pressing in extra time. They won 2-1. I did not shout. I reopened the spreadsheet to hunt for outliers. Croatia won a tournament of low PPDA? Then PPDA is only a letter. It describes a behaviour; it does not explain a result.

The financial dimension needs a number with a provenance: a fee, a contract length, a wage bill, an amortisation schedule. In 2026, my report on matches played in empty stadiums reached a Ligue 2 club. Le Havre used it in a negotiation, driving down their offer for a young striker whose goalscoring record at home had been excellent before the pandemic. I did not feel pleased about that. I saw clearly that a correct dataset can also be used as a weapon, and that the writer is responsible for both ends of it.

The results dimension needs a large enough sample. In 2026 I analysed 81 matches played behind closed doors in the 2026-20 season. Home teams won only 26 percent of them, against 43 percent before the pandemic. An empty stadium is the finest laboratory for someone who loves data, because it removes a variable that is normally impossible to isolate: crowd noise. But 81 matches are still 81 matches, not 810. I stated the error margin plainly and made clear the conclusion held only under no-crowd conditions; it should not be generalised into a claim about players' psychology.

The remaining three dimensions — rules, dressing room, industry transmission — are stricter still. On FFP and PSR, an analysis has value only when tied to a club with a specific loss, a specific accounting period, a specific ruling. Precedents such as the 115 charges brought against Manchester City, or the points deductions handed to Everton and Nottingham Forest, are tools for understanding the enforcement framework. They are never evidence for accusing anyone else.

On the dressing room, this is the most source-sensitive dimension of all. I will say it plainly: it is where I refuse to write most often. A story about a rift between a manager and a captain needs at least two independent sources, ideally three, and it needs a clear distinction between an ordinary tactical argument and a genuine power struggle. Without that anchor I have nothing to say, however far the rumour has travelled across forums.

On transmission, the rule is tighter still: only an originating event can create a path. A transfer, a broadcasting contract, a change of ownership, a ruling. Without an originating event, every inference about knock-on effects is decoration.

Read carefully, this nine-dimension structure is really a gap detector. It does not help me write faster. It tells me precisely where I have nothing.

The contrarian angle: a null report is a product, not a failure

In this trade I am considered unnecessarily difficult. People say: if everyone waited until they had enough data, there would be nothing to publish.

That argument is commercially sound and professionally wrong. And the mismatch between those two standards is where modern football's most expensive mistakes are born.

Take the transfer market. A player is a variable, the market is a function, but most of my life has been a constant. A valuation model can state with great precision that a 19-year-old is worth thirty million euros, based on minutes played, key passes, and age curve. It cannot measure something anyone who has sat in a dressing room knows matters more: whether he can spend three months on the bench without fracturing the group. Models overrate young potential and underrate dressing-room chemistry, not because nobody knows this, but because chemistry has no unit of measurement.

For the same reason, I do not believe a club should issue a full analytical report before it has a data anchor. A null report, published openly, has one practical function: it halts the transmission of a false conclusion before that conclusion is used to make a decision. Along that chain, the cost of retrieving an error multiplies at every node. Refusing at the first node is hundreds of times cheaper than correcting at the last.

I see the same pattern in two fields I have followed for years. The first is club IPOs. When a club lists, the emotion of its supporters is converted into an asset class, and the pressure of quarterly financial reporting begins to weigh on sporting decisions. A manager is judged by broadcasting revenue in Asia rather than by points after ten rounds. A young talent is sold to balance cash flow mid-season. A null report, in this case, is a reminder that some decisions should not be justified with data, because they are financial decisions wearing the clothes of sporting ones.

The second is esports. A single mouse click on an esports screen carries the shape of a pass. Professionalisation turns players into assembly-line products: a fitness programme, a psychology programme, an opponent-analysis programme, and an individual style smoothed flat across each digitised training session. When every dataset is clean and identical, difference no longer lives in the metrics. It lives in whoever dares to keep something the spreadsheet cannot measure.

Finally, there is another kind of white space I cannot ignore: the cancelled match. A cancelled match is not a lost point, it is a lost page of the diary. For an analyst, a postponement for crowd, weather, or security reasons is a data point that will never exist. A whole season is woven from such diary pages, and every torn page leaves a hole in every model that follows.

There are matches won on the pitch but lost on the spreadsheet — I choose the spreadsheet. Not because the spreadsheet is always right. Because it is the only thing that lets me go back and check where I was wrong.

What to watch in the next round

In the current major-tournament cycle, I will track four specific signals.

First, the share of analytical pieces that carry a publication date and a named data source. A number without a date is a number that cannot be verified, and across a tournament lasting several weeks, a group-stage metric may already be meaningless by the semi-final.

Second, how newsrooms handle data gaps. When a provider re-releases an event package, how many earlier articles are corrected? This is the most direct measurement of a sports desk's seriousness.

Third, the frequency of publicly issued null reports. A club willing to say "we do not yet have enough data to assess this player" is a club that will buy better than the rest.

Fourth, the conditions that accompany every new tactical fashion. In 2026, when the profession praised Achraf Hakimi for 142 sprints and roughly 2.3 chances created per match, I dug into the data and found the corridor behind him stood empty for 34 percent of match time. Morocco stayed safe because their centre-backs ran above 31 km/h. I wrote a note warning that the fashion held only if the back line had enough speed. Against France, the opponent funnelled the ball repeatedly into Morocco's right flank. I no longer praise a system without listing the necessary and sufficient conditions for it to function.

I am 66, old enough to know a number never tells a story unless you ask it a question. The first question I always ask is not what this number means, but where it came from, what it measures, what it leaves out, and how many observations stand behind it. When the answer is none, I close the spreadsheet, shut the machine, and go to sleep. The next morning, I send the newsroom a single line: not enough data to conclude.

That is the whole article. And to me, it is the most honest piece I could have filed that night.