The Discipline of an Empty Dossier

In November 2026, in Turin, I sat at my desk until two in the morning with 240 minutes of tape from the second leg of the Italy–Sweden play-off. The score was 0-0; no goals, no bright moment worth quoting. My editor dismissed the first draft with one short line: "Girls write about tactics just for decoration."

I rewound the tape, drew 14 pressing maps, marked every minute in which Italy's midfield was cut off from both flanks, and noted how the pair of Ciro Immobile and Andrea Belotti were starved of service in the middle. The second submission had data, timestamps and diagrams. It ran, not because the editor changed his mind, but because he had run out of reasons to refuse.

Eight years later I met the same lesson in another form: an analytical dossier with a full table of contents, a full set of templates, and not one line of data inside.

My trade runs on a four-stage pipeline. The source stage establishes the event. The fact stage breaks the event into verifiable units. The inference stage turns facts into causal relations. The conclusion stage turns those relations into a judgment that can be contested. When the second stage is empty, the other three do not grow weaker — they turn wrong.

In 2026, when I first joined Autosport, I learned a principle from the engineering room: telemetry data without a baseline lap to compare against is just noise. A failed sensor does not return zero. It returns a null. The distance between the two is the entire foundation of trustworthy analysis.

Null data and zero are two different kinds of information, and confusing them is the most common error in sports analysis.

In football that error shows up in a familiar sentence: "We have no pressing data on this team, so they do not press." Wrong. No data means the analyst has not measured. The conclusion "there is no factor worth worrying about" is only valid once enough measurement exists to rule factors out; if no measurement was ever taken, there is no right to conclude.

In a standard scouting template, the cell for "time to recover the ball after losing it" sits in the fourth column. When that cell reads "not available", the report still looks clean, still fills its pages, and can still be presented to a coaching staff. But every judgment behind it about transition defending loses its value.

The grey zone is not a place where light is missing. It is the place where football is most real. But the grey zone is also no excuse for a writer to stand still.

Based on my experience watching matches, a tactical conclusion only holds when the sample is large enough and a control group exists. In 2026-20, Atalanta scored 98 goals in Serie A — the highest tally in the club's history. I logged those goals across two seasons and classified them by origin: transitions after winning the ball, set pieces, wide combinations. Only after separating them did the pattern appear: most of the opponent's damage came from three or four passes immediately after losing the ball, not from settled possession. Duván Zapata and Josip Iličić were the clearest outputs of that pattern.

In 2026, football returned in empty stadiums. I built a dataset of 120 matches, comparing the pressing intensity of home teams with and without crowds. The result: home teams lost roughly 15 percent of opposition pressing. With only three matches, I would have no right to speak of 15 percent — only the right to speak about three matches.

At this point the transmission mechanism of a null error becomes clear. A blank cell read wrongly infects the cell behind it. A downstream reader sees the words "insufficient information" and understands "there is no problem". A technically valid analysis then produces a factually wrong conclusion.

There are 22 players on the pitch, but the match is really played between two brains. A brain working with null data needs a different discipline: it must state plainly what is missing, instead of filling the gap with plausible-sounding guesswork.

There is a flip side. Discipline around null results can easily become a shield that lets a writer never commit to a judgment. Readers do not need a list of possibilities. They need to know which possibility you chose, and what you lose if you are wrong.

In 2026, at the World Cup in Russia, I wrote a long piece on how Isco moved into the gaps between the lines during Spain's 3-3 draw with Portugal. The edit cut it in half, on the grounds that nobody reads that level of detail. The lesson lay elsewhere: put the main claim in the opening paragraphs, use diagrams instead of long-winded passages, and keep the core section on Spain's central rotation intact. Since then I write the way one proves a theorem: premise, evidence, conclusion, limits.

My World Cup theorem does not predict the champion. It predicts who collapses first. Precisely because of that, I have to write the alternative scenarios before presenting the collapse scenario. If the team does not crack, the analysis must explain why it held, rather than quietly look away.

Every new contract is a hypothesis. The match is the experiment. An empty dossier, in the end, is also an experimental result: it tells you where the measurement method failed.

What is worth watching in the next round is not which team wins. It is this: when the data arrives, who among us dares to change an earlier conclusion, and who will keep it by calling the new data noise.