The Empty Cells Report: Why "I Don't Know" Is Football Analysis's Bravest Answer

At two in the morning in my London flat, I opened a file I had just downloaded. Nine tabs. In every tab, rows and rows of empty cells — tactics, club finance, the results cycle, league geography, governance, management, the risk matrix, media narrative, industry transmission. Every cell repeated the same sentence: insufficient information, assessment not possible. No team names. No player names. No scoreline. An analytical framework with zero raw material inside it.

I laughed first. Then I felt uneasy. Because I know that in the machinery of football media today, if someone filed this document, it would be filled in within three hours. By whom? By the analyst who wakes up already knowing that a confident paragraph can be built on zero data. And three decades of watching this industry tell me this file may be the most honest football document that has crossed my desk this month.

Because the rarest skill in football analysis is no longer building a model — it is recognising when to say "I don't know."

Football writing has settled into a fixed mould over the past decade. Within ninety minutes of full time we are expected to deliver a verdict. Who wins, why they won, whether the manager survives, whether the club broke financial rules, whether there is a crack in the dressing room. Nine dimensions, one clean answer for each, and the faster the answer the better. The economics of social media teach exactly this: hesitation does not sell, certainty does.

I am part of that machine myself. In 2026, after Conte's Chelsea won thirteen league games in a row, I wrote a thread showing that Chelsea held only 52 percent possession during that run but generated 1.9 xG per game. The headline was clickbait; the numbers underneath were real. Two thousand replies followed, the BBC put me on radio to argue about it, and I gained fifty thousand followers in three months. That success taught me a lesson — and set a trap.

The lesson was that a hot take without numbers does not survive. The trap was that once numbers are present, people assume the analysis is finished. Yet football has regions where numbers simply do not exist, and never will, at least not publicly. The power structure of a dressing room. The cold war between a manager and a sporting director. The confidence inside a footballer's head. Which agent's phone was answered on the final night of a January window. There is no optical tracking for any of it, no PPDA, no passing network.

And that is precisely where our profession commits its worst offence. Where there is no data, we write sentences that sound like data. I have a name for it — the null record problem: confident analysis built on zero information.

Walk through the nine dimensions and you see how dangerous zero data really is, and where we routinely manufacture false certainty.

Start with tactics, because tactics are the easiest thing to fill in. A formation was visible, a pattern was visible, so an explanation gets built. But formation and tactics are not the same object. People still sell that 2026-17 Chelsea run as "Conte's revolution." The maths in my file is black and white: it was not a philosophy. It was a math problem with wing-backs. Three centre-backs covered the space, two wing-backs pushed the opposing full-backs back, two inside forwards drifted into the half-spaces and triggered counters. Possession fell; shot quality rose. A coach who lowers possession and raises xG should be described as an accountant, not a philosopher.

Now the opposite example. At the 2026 World Cup, Germany lost 0-2 to South Korea having taken 26 shots, scoring none, with just 0.8 xG from open play. Germany took 26 shots, scored zero, and the xG shrugged. That night plenty of writers concluded that German mentality had shattered. It was not a story about mentality. It was a story about slow, predictable, centrally-loaded build-up that cannot break two deep blocks and a low defensive line, no matter how many shots it accumulates.

This is where the trap of treating xG as a deity appears. xG is a smoke detector, not a fire. In 2026-23 Erling Haaland scored 36 Premier League goals, far beyond his model expectation. That is not proof that xG is broken; it is proof that shot quality, game state, goalkeeper positioning and finisher skill must be read separately. An analyst who reads only the xG number has dropped the scoreline story and replaced it with another single-number story.

Finance is crueller still. Nobody sets prices in a January window; time does. In January 2026 Chelsea spent more than £300m in a single month, including £106.8m for Enzo Fernández — a British record. Was that fee the price of Fernández's ability, or the price of panic before the window shut? If a file contains only transfer fees and wages while broadcasting revenue, commercial revenue and net debt sit empty, that is not analysis — it is a price list.

The results cycle produces an even bigger error. Three defeats become a crisis, three wins become a fairytale. Yet divergence between xG and points is the most ordinary event in football. If a team generates 2.0 xG a match for five games and scores twice, that may be a finishing problem, a goalkeeper's form, or simply small-sample noise. Writing about sample size does not earn clicks; writing about crisis does.

League geography and governance repeat the same mistake. Everton were docked ten points in November 2026 for breaching Profit and Sustainability Rules, reduced to six on appeal; Nottingham Forest lost four points in March 2026. Here the legal reality is separate, the financial reality is separate, the sporting reality is separate. Where three records exist, a one-sentence verdict denies two of them.

Management and the dressing room are the room with the door shut. No tracking camera enters. Yet this is where the most confident prose is born — "he has lost the dressing room." On what evidence? Usually on the evidence of losing. That is circular logic: you lose, therefore the dressing room is gone; the dressing room is gone, therefore you must lose. Building a risk matrix is often ritual too — write "medium" in every cell and nobody asks a question while the file looks full.

Media narrative is the noisiest dimension of all. A transfer rumour has no source, yet it becomes news, and then returns as a "report." An agent's interest, a club's negotiating strategy, a journalist's deadline — the sum of those three forces is a headline, not information. Industry transmission — academy to broadcast, agents to capital — operates on a timescale of years, not three days. Yet we explain that on the night of the match too.

Our embarrassment over environmental variables is greater still. I grew up in Bangladesh and now work in London, and I have seen with my own eyes the difference in football infrastructure between the two. A waterlogged pitch, thirty-degree heat, a fifteen-hour journey, two matches in three days — these are not excuses, they are inputs to a prediction. But there is one condition: the weights must be set beforehand, not narrated afterwards.

These nine empty cells mirror a much larger problem — the problem of verifying whether information is true. That is the core idea behind blockchain: every piece of data carries a timestamp and a record that cannot be altered. Imagine if a transfer fee, an injury report, a points deduction existed not only in a club press release but on a ledger everyone could see. How much would "sources say" and "I have heard" be worth then? Football's real use for blockchain is not selling fan tokens. It is issuing birth certificates for information.

This is where I have to stand against myself, because "insufficient information" can itself become a comfortable shelter.

The analyst who always says more data is needed is never proved wrong — and never teaches anyone anything. It is the safest position in our profession. The uncomfortable truth is that the best analysts decide on sixty percent of the information and admit it openly. They write: my confidence is seventy percent, here is the basis, and here is the condition that would prove me wrong. That is the honest route — neither false certainty nor permanent hesitation.

If I am wrong, it will happen like this. Over the next two years the volume of football data will grow so fast — player tracking chips, semi-automated offside, published club accounts — that the null record argument becomes irrelevant. Perhaps by 2027 there will be measurable indices even for dressing-room power structures, and saying "I don't know" will be plain laziness.

Still, I have a warning. The empty stadiums of 2026 taught us that football cannot be understood with environmental variables stripped out. The empty stadiums of 2026 did not merely kill the home-advantage myth — they showed us which data we had never measured at all. Home advantage was never only a crowd; it was refereeing tendencies, sleep cycles, travel fatigue, the pressure of a familiar stand. When the machinery removed the crowd, the remaining variables stood alone and introduced themselves. Anyone who treated the crowd as the sole cause took a hit that year.

And one more thing worth remembering — pre-season tours abroad. Turning clubs into circuses and parading them through trophy cities moves money from player fitness into commercial revenue. Congestion and commercial travel: without writing those two accounts separately, no explanation of August injuries holds up.

So here is my prediction, and you can verify it. By 2027 at least one major football publication will add a null-disclosure rule to its style guide — where a question has no evidentiary basis, the analyst must state plainly that information is insufficient, and any estimate must be labelled as an estimate. If that does not happen, you will know we are not selling analysis. We are selling stories.

And the nine empty cells in my file? I will not delete them. They stay on the desktop as a reminder — an empty cell is always worth more than a wrong answer.