An Empty Report in the Paddock: The Verification Discipline of an F1 Analyst

Milan, 6:40 a.m. on a Friday. On my desk lies a nine-page document, printed double-sided, fastened with a black plastic clip — exactly the format the coaching staff has used for years. I open the first page. "Title": N/A. "Source": N/A. "One-sentence summary": blank. "Information points": empty list. "Core viewpoints": empty list. Only one field is filled in: the domain label — F1.

Nine pages. One data point.

The assistant stands in the doorway with a coffee, waiting for me to say something. I say the sentence that thirty-eight years around racetracks have taught me I must say: "We have nothing to analyse. Go get the source again."

In this profession, that is the hardest sentence to say. Harder than admitting I predicted a goal in the 90+3rd minute wrong. Predicting wrong is only wrong once. Publishing an analysis with no root corrupts the credibility system from the inside, and that system takes several seasons to rebuild. I am 57 now, and I have sat in enough technical meetings to know that one honest blank page is worth more than ten pages filled with plausible-sounding guesswork.

The trade of analysis under transmission pressure

Every race weekend, a modern car pushes hundreds of telemetry channels back to the technical centre, and by Sunday evening the whole world already has a summary. The distance between those two moments — when the data is still hot and when the headline is already on the front page — has shrunk from days to a few dozen minutes. That is the space my profession is squeezed into, and it is also where most of the industry's mistakes are born.

I joined F1 reporting in 2026, when knowing whether a car was fast meant standing at the barrier and timing it by hand. No xG, no PPDA, no tyre graphs. Only eyes, a notebook, and a pay-per-minute phone. The discipline of that era was simple: if you did not see it yourself, you did not write it. By 2026 I had set a professional record by covering 406 consecutive Grands Prix live, and more than 500 across my career. People usually cite that record as a feat of stamina. It was really a feat of discipline: 406 consecutive times I had to remind myself that what I was seeing on screen was only a fraction of the event.

In 2026 I served as an editor for an industry award. That role pushed me off the racetrack and forced me to read other people's files instead of gathering my own. The lesson arrived quickly and painfully: a beautifully presented file, with tables and clear headings, can contain not a single verifiable fact. That tidy appearance is the most dangerous thing in this trade, because it lowers the reader's guard.

Anatomy of a null result

Back to the nine-page file from Friday morning. What matters is that its structure is entirely normal. There is a title field, a source field, a summary field, a viewpoint field, an entity field. Every frame exists. Every frame is empty.

A layperson might look at it and think: "The original article must have had nothing special in it." Wrong. A professional looks at it and knows immediately: this is a fault at the data-extraction stage, and it is entirely different from the original article genuinely containing no information. Those two situations require opposite responses. One is a system fault that must be re-run. The other is a valid conclusion that must be accepted, with no further analysis.

The distinguishing signal lies in the internal inconsistency of the file itself. The domain label reads F1, meaning a classifier ran and issued a verdict. But the body is empty. A classifier only assigns an F1 label when it sees F1-related content. If the content does not exist, the label should not exist either. Both appearing together is the trace of a process broken in the middle, not of an empty article.

This is the kind of reasoning I developed after years of working with tracking data. A data gap is never neutral; it always carries information about the process that produced it. When an indicator disappears, the right question is not "what is its value" but "why did it disappear, and at which stage".

I learned this lesson through a concrete mistake, and I still retell it whenever I hire someone new.

The San Siro sensor and 0.2 seconds

In 2026, when I was 48, I was a member of the coaching staff at AC Milan. The board tasked me with validating the motion dataset from 20 Serie A matches in the 2026-17 season. The purpose was practical: the club needed to know why home and away performance diverged so sharply.

The initial result was beautiful, like a fairy tale. Milan's expected goals (xG) at San Siro was 1.85. Away from home, it was 1.02. Nearly double. Anyone reading the table would conclude: Milan is a team that only knows how to play in front of its own crowd, and the problem lies in away-match mentality.

Actual goals scored were identical.

That was the breaking point. A team creating nearly twice as many chances at home but scoring exactly the same number of goals away either has a brutally inefficient attack at San Siro, or the data is wrong. I chose the second hypothesis and went to check. After cross-referencing video with sensor data, I found the culprit: a sensor in the south-west corner of the stand was lagging by 0.2 seconds. Every goalkeeper-initiated build-up — the situation type Milan used most at home — was recorded out of sync. The error was too small for anyone to spot from summary tables alone.

I wrote a 14-page internal report, recommending recalibration and — more importantly — re-reading all historical data with a compensation factor. The practical result: the home-away gap narrowed substantially after correction. Coach Vincenzo Montella used those findings to increase right-side ball circulation, and the team won 5 of its last 8 matches, securing a Europa League place.

The lesson was not "xG is useful". It was: a wrong indicator still produces a highly convincing conclusion, and that conclusion will never incriminate itself. Only cross-referencing multiple sources exposes it. Since then I have set a hard rule for everything I write: every cited indicator carries a note on measurement conditions, and no value is used unless cross-checked against at least two independent sources.

The transmission chain: from sensor to headline

To see why an empty analysis is more dangerous than a wrong one, you have to look at how information travels through this sport.

Upstream, teams collect raw data: sensors, telemetry, force measurements, tyre data. Midstream, analysts turn raw data into models — computation, interpretation, meaning. Downstream, the public receives conclusions: a headline, a judgement, a figure repeated on air.

At every link, some information is lost and some interpretation is added. By the end of the chain, readers no longer see the raw data; they see only the conclusion. And a conclusion has no way to defend itself against scrutiny, because it was severed from its evidence long ago.

An empty analysis breaks that chain at the very first link while keeping the exact shape of a complete one. It has a title, subheadings, tables, conclusions. Readers downstream have no way to know the first link snapped. They will read "risk level: low" and understand "checked, found safe", when the truth is "not checked, unknown".

The confusion between "low" and "undetermined" is the most expensive error class in sports analysis, and it happens daily, in every league, in every market.

Four pillars of verification

After thirty-eight years I have distilled four minimum conditions for an analysis to be considered grounded. If any one is missing, the analysis must be returned, not edited.

The first is provenance. A claim with no named publishing body and no publication date cannot be graded for credibility. In the transfer market, this is the line between a checkable report and an inflated rumour. A contract only looks beautiful on paper until someone tries to fit it into a running system.

The second is a temporal anchor. Without a date, a race, or a session, every observation about trend is meaningless. An indicator only means something when you know which race it came from, at what temperature, on what track surface.

The third is an entity set. You need at least two objects to compare. One team, one driver, one technical director standing alone produces description, not analysis.

The fourth is a minimum quantitative floor. For me, that floor is three information points, one entity set, and one date. Below that threshold, analysing is not doing the same job harder; it is doing a fundamentally different job.

The nine-page file from Friday meets none of the four. No source, no date, no entities, no information points. It is not a weak analysis. It is an analysis that does not exist, presented in the shape of one.

Germany v South Korea 2026 and the distorted trapezoid

Experience taught me that a correct conclusion can still be rejected if it is not presented correctly. That was the lesson of the 2026 World Cup.

That year, thanks to the previous year's internal report, Sky Sport Italia invited me as a technical commentator. In the Germany-South Korea match, on 70 minutes, the score was still 0-0 and Germany was pushing forward for a winner. I posted on Twitter: Germany's defensive line is averaging 68 metres high, pressing has failed 17 times, South Korea already has 12 counterattacks, and if the block is not dropped deeper, the goal will come from an aerial situation.

In the 90+3rd minute, Kim Young-gwon scored exactly to that script. The match ended with Germany eliminated in the group stage.

The reaction did not follow the script. Thousands of accounts piled in to mock me for "turning emotion into arithmetic". Nobody argued with the content. They argued with the presentation.

But Gazzetta dello Sport republished the piece, with the distorted trapezoid diagram I had drawn of Germany's back line. That diagram was what made the story stick. Numbers stay in the report; images stay in the reader's head. Since then I dropped "68 metres high" in favour of "the zip has burst open to the valve box". I describe a defence as a vertical rectangle stretched between centre-back and goalkeeper, rather than listing distances.

That was a methodological adjustment, not a data one. The data was right. The delivery was wrong.

Fabrication risk tops the risk table

In every risk assessment I have built for a football club or a racing programme, one item always sits last and is always underrated: the risk to analytical integrity.

It is underrated because it does not concede a goal, does not break an engine, does not cost a driver a seat. Its consequences are slow and indirect. But once it happens, it destroys the one thing every team needs and no team can buy: the ability to trust its own data.

A team making strategic decisions on a wrong indicator loses one race. A team making decisions on a report filled with guesswork loses a whole season — and by the end of it, does not know why.

Every collapse has a precursor; few people bother to look beforehand. That precursor usually sits in a blank data field, a misaligned sensor, a late report, a postponed meeting. Things too small for a headline, and therefore never fixed.

I have watched technical meetings run three hours just to argue which of two barely different values to use. Outsiders see waste. Those arguments are precisely what stops a project collapsing at the fifteenth race of the season.

The reward goes to whoever dares to assert

Here is the counter-intuitive part, stated plainly.

In today's content industry, whoever dares to assert is rewarded more than whoever dares to stay silent. A decisive claim, even a wrong one, generates engagement. A cautious claim, even a right one, generates silence. Nobody designed that mechanism, but it operates very reliably, and it rewards exactly the behaviour my profession exists to resist.

The result is that the whole ecosystem is pulled to one side. The volume of claims rises, the quality of the foundation falls. Readers encounter more conclusions and less evidence. Gradually they lose the ability to distinguish a conclusion drawn from data from a conclusion written because a conclusion was needed.

The biggest blind spot in the sports public is not a lack of technical knowledge. It is equating volume with accuracy. When a piece is dense with numbers, readers assume it is serious analysis. The more tables, the higher the credibility. But tables, like the nine-page file from Friday, can be full of empty cells and still look thoroughly persuasive.

There is another layer. Teams publish data selectively. They do not lie, but they choose the favourable part to say. A team with power-unit reliability problems will emphasise aerodynamic progress. A team that just lost a technical director will emphasise organisational stability. Readers outside, already trained to trust volume, absorb that curated slice as the full picture.

In this phase of the season, as the standings begin to separate and points pressure rises, that mechanism works even harder. Midfield teams need a story to keep sponsors. Front-running teams need a story to hold psychological distance. Backmarkers need a story to hold hope. Each group tells a different version of the same data, and all three versions travel with the same credibility.

An empty grandstand does not kill the race, but it removes something the data cannot measure. That holds for races without spectators, and equally for analyses without verifiers: the structure survives, but what was inside it has lost its weight.

Not every gap is a fault. Some gaps are conclusions. Some gaps are incidents. Telling the two apart is the entire substance of an analyst's work.

A test for the next race

Based on my experience watching these races, I propose a simple test, applicable from the next round.

Before accepting any claim — mine or anyone else's — check four boxes: source, date, entities, information points. If one is empty, pause. If two are empty, discard. If three are empty, that is not analysis. If all four are empty and the prose still flows, what you are reading has the shape of analysis and the substance of fiction.

Every tracking number belongs on the operating table, not on the altar. This test demands no technical knowledge, only patience. And in a season where ten teams tell ten different stories about the same round every weekend, patience is the cheapest and most undervalued competitive advantage.

Next round, when the data sheets open, watch for the first empty cell. It will be there. The only question is whether anyone has the nerve to point at it and say: there is nothing here yet, go get the source again.

Because what decides who wins a season is not who has the most data, but who knows clearly which of their data is not yet enough. And in a paddock where everyone is talking, the person staying silent to re-check their own numbers is often the only one still holding on to the truth.

I close the nine-page file, clip it back with the black plastic clip, and place it exactly where it belongs: the drawer for things that must be re-run. There is always another race. And there is always another source to go and find.