When an Empty Data Cell Gets Read as 'No Risk'

On my screen, in the eleventh row of a spreadsheet more than two hundred rows long, there was a blank cell. That cell should have held a driver's long-run time from the second free practice session. I mistyped a formula, left it empty, exported the sheet and published it on my personal blog. Three days later a reader in Manchester messaged me: "Did that team skip long runs, or are they hiding something?" I reopened the original file and realised there was something worse than a wrong number: a blank cell. A wrong number makes people suspicious. A blank cell makes people write a story into it. From that day I set one rule for every data sheet I publish: never leave a cell blank. If it hasn't been measured, write "not measured". If the measurement failed, write "measurement failed". Never let the reader fill the gap.

The summer of 2026 taught me that a gap is never empty; it is simply waiting for the right reader. But there is a paradox attached that took me another two years to fully understand: a gap is also waiting for the wrong reader.

That paradox is not a private matter for one blog. It sits at the centre of the biggest regulatory cycle Formula 1 is about to enter since 2026. Formula 1 is a sport that runs on declarations. A car is not directly measured by anyone outside the team; it is described. Teams submit technical documents to the FIA, financial reports to independent auditors, tyre data to Pirelli, and internal briefings to sponsors. Much of what the public calls "F1 information" is in fact a filed declaration. Misunderstand that, and a reader will keep confusing an absence with a denial.

The cost cap is the clearest example. The base figure of 135 million US dollars for a 21-race season, applied from 2026, plus roughly 1.2 million US dollars for each additional race, is not a measurement. It is an accounting figure declared by the team itself and audited afterwards. Between the declarer and the auditor there is always a lag, and inside that lag every interpretation becomes possible.

Then there is the ATR scale, the Aerodynamic Testing Restriction. Since 2026, aerodynamic testing rights have been allocated in reverse order of the final standings: the champion may use only 70 per cent of the allowance, while the last-placed team gets 115 per cent. An allowance of 100 per cent corresponds to roughly 320 wind tunnel runs and 80 hours of aerodynamic testing in a six-week period. The gap between first and last in the same period therefore approaches 150 runs. This is an equalisation mechanism, and every equalisation mechanism depends on the parties honestly declaring the runs they have used.

And on 6 June 2026, the FIA World Motor Sport Council published the 2026 technical regulations: a hybrid power unit splitting output between the internal combustion engine and the electrical system, the MGU-K raised to 350 kW, the MGU-H removed entirely, 100 per cent sustainable synthetic fuel, active aerodynamics with two wing modes, and cars that are smaller and about 30 kilograms lighter. It is the largest reset since the hybrid era began.

Those three systems, the cost cap, the ATR scale and the 2026 rules, share one feature that very few analyses touch. All three run on declared data, and all three share the same blind spot. That blind spot is the gap. An empty cell in a verification system does not mean "no problem". It means "nobody has measured". Those two statements are completely different, and the price of misreading the first as the second is usually paid across an entire regulatory cycle.

On 28 October 2026, the FIA announced an Accepted Breach Agreement with Red Bull Racing concerning its 2026 cost cap report. The penalty was a 7 million US dollar fine and a 10 per cent reduction in aerodynamic testing allowance over twelve months. Public debate focused almost entirely on the 7 million. Based on my experience watching race weekends and technical briefings, that is reading the wrong headline. The 7 million is a one-off cost, comfortably absorbable within a leading team's budget. The 10 per cent aerodynamic allowance cut is a long-running penalty that feeds directly into the car development pipeline over many months. In a technical environment where track testing is not unlimited, the scarcest resource is not money. The scarcest resource is the number of permitted mistakes.

The real question lies in the structure of the system. A cost cap report is not an automatic measurement. It is thousands of lines of declaration covering engineer salaries, material costs, depreciation and costs relating to the three leading teams, with many categories excluded, from driver salaries to marketing to power unit costs for customer teams. When a line is not declared, the system does not raise an alarm. It registers silence. And in a system that reads declared data, silence is the most dangerous grey zone, because it generates no signal to react to.

This is why I tell young editors that analysing Formula 1 is identical to analysing football in one respect: every tactical diagram starts with a shaky hand-drawn line in PowerPoint. That line is not beautiful. It has no presentational value. But it is honest, because it shows where I could not draw. The bad analyst erases the shaky line and replaces it with a perfect straight line that has no data behind it.

The ATR scale is an example of designed asymmetry, and that asymmetry only means something if we read it correctly. At about 320 wind tunnel runs for a 100 per cent allowance, the champion receives roughly 224 runs, while the last-placed team receives roughly 368 per six-week period. A midfield team can have nearly a hundred more runs than the leading team in the same window. But here is where I want everyone to pause. More runs do not automatically become faster lap times. They become more chances to test hypotheses, and those chances only have value if the team has enough people to read the results correctly. A team running 368 runs and misreading all 368 is worse off than a team running 224 and reading 150 of them correctly.

Transition is not a stretch of running. It is the silence between two intentions that few people can read. In this case, that silence sits between the moment data leaves the wind tunnel and the moment an engineer decides to keep or discard an aerodynamic detail. That is where teams invest most, and where the media reaches least, because there is nothing to photograph.

The same pattern appears at the level of the Technical Directive. On 7 July 2026, the FIA issued TD39 on vertical bouncing and floor stiffness, effective from the Belgian Grand Prix. In 2026, TD018 tightened the rules on flexible bodywork. The crucial point many miss: a technical directive does not change the car immediately. It changes the definition of what counts as legal. That means data in the wind tunnel and in CFD simulations must be re-read before any mechanical part is manufactured. Some teams spent two weeks merely checking whether their model remained valid. Those two weeks appear in no timetable, no press conference and no championship table. They appear three races later, in the form of a car lacking downforce at low speed.

Post-race scrutineering delivers a similar lesson, more visibly. On 22 October 2026, at the United States Grand Prix in Austin, both Lewis Hamilton and Charles Leclerc were disqualified for excessive skid block wear. On 28 July 2026, at Spa-Francorchamps, George Russell was disqualified after winning because his car was 1.5 kilograms under the minimum weight. All three had passed earlier inspections. This is not a story about cheating. It is a story about the nature of an inspection. A racing car is a system whose state changes continuously across laps, tyre sets and track temperatures. A technical certificate is a snapshot at one point in time. Reading "passed inspection" as "no risk" is a classic cognitive error, and it happens to people who have worked in the sport for decades.

I verified this in the smallest way available: I built my own Excel sheet recording every scrutineering result and every instance in which that result was overturned after the race. Over the last three seasons the reversal rate is not high. But the rate is not what interests me. What interests me is the interval between the moment a result is published and the moment it is overturned, and how many articles are written inside that interval on a fact that is about to expire.

At the level of the personnel market, the mechanism is even harder to read. On 10 September 2026, Aston Martin announced the signing of Adrian Newey as a senior technical partner, with the contract effective from 1 March 2026 and a stated target of the 2026 car project. Most articles stop at the question of which team gets stronger. Few touch the real mechanism: gardening leave, the mandatory break between two employers. It is a tax on knowledge. It exists because what is bought in the F1 labour market is not a person but access to the way that person reads data.

In that market, an engineer's value does not lie in knowing which aerodynamic detail belonged to the previous team. It lies in knowing which question must be asked before a new dataset. That cannot be written into a contract and cannot be measured by any index. A self-made data architect never waits for someone to hand over the data, but also never believes that absent data means absent problems.

Meanwhile the rumour market works on an entirely inverted logic. It manufactures data from the absence of data. An empty seat, a private meeting, a deleted status update, all get read as signals. The problem is not that rumours exist, but that they are not tiered by source. An unsourced claim and a claim from a veteran paddock reporter must never carry the same weight. A driver manager has an incentive to generate noise, because noise distorts the price level. That is the largest hidden cost of the transfer market, and it appears on no balance sheet.

At this point I have to argue against myself. The assumption above is that the greatest risk comes from missing data. In reality, most serious mistakes in the paddock come from complete data read badly. There are two kinds of error, and they carry different levels of danger. A false positive happens when we see a pattern that does not exist. It is noisy, easy to debunk, and usually dies within a few races. A false negative happens when we look at an empty cell and conclude there is nothing to worry about. It is silent, it triggers no reaction, and it can survive an entire regulatory cycle.

In August 2026, rewatching the entire season's free practice sessions, I noticed something small but memorable. The long-run comparison tables published widely featured only three leading drivers in almost every round, because the other drivers had not completed enough long-run laps. The casual reader looks at the table and concludes the rest of the grid is not worth following. The truth is that their data was not collected adequately, not that they were slow. No data and no problem are two entirely different states, but in a spreadsheet they look identical.

The counterintuitive view I want to propose is this: in the 2026 regulatory cycle, the greatest competitive advantage may not come from collecting more data, but from daring to write empty cells into the sheet. Every major team has access to the same public GPS data, the same FIA documents, the same race calendar. When everyone reads one dataset, the difference is not in the volume read, but in who accepts the part that has not been read. That is why I say that when there is no football, I draw football. And it turns out drawing is also a way of understanding. The same holds for the racetrack: when there is no data, I draw the gap, and the gap forces me to be honest.

One more factor that no spreadsheet measures. After more than three years working in London, I have learned that every technical decision is made by a person who is tired, under pressure from above, and aware of a deadline. The car does not run on a theoretical curve; it runs on a chain of human decisions. Removing that factor from the analysis is volunteering to be blind to half the picture. The shaky hand-drawn line in PowerPoint comes from a human question, not from a dataset.

I have no intention of concluding that everything is unknowable. The opposite. What I want to state is a rule of discipline: before publishing any data sheet, ask yourself which number genuinely changed the conclusion, and which number was only there to thicken the page. Personally, once I finish writing a conclusion, I always try to imagine another dataset scenario that would lead to the opposite conclusion. If I cannot imagine one, my conclusion is not mature enough to publish.

For my part, I will spend the coming month rebuilding a tracking sheet, this time tracking every public claim teams make before the 2026 season and checking them against on-track data from winter testing. Each claim gets its own status column, and there will be three options: verified, unverified, no data available. No cell will be left blank.

Is an empty cell in your spreadsheet waiting for you to write a story into it, or waiting for you to admit that you have not measured it yet?