F1: When Data Goes Silent, Mistakes Begin With Belief

F1: When Data Goes Silent, Mistakes Begin With Belief

Las Vegas, the night of 17 November 2026. The first free practice session of the Las Vegas Grand Prix stopped after exactly nine minutes. No rain. No engine blew. A water valve cover on the track surface was ripped loose, and when Carlos Sainz's front wheel ran over it, it tore through the floor of his Ferrari. The team had to change the chassis; Sainz took a ten-place grid penalty because the power unit system had been pushed beyond its permitted allocation.

No simulation model anticipated that failure. No sensor measured how solid the track surface was before the wheels turned. All weekend, the organisers believed everything had been checked. The biggest risk in modern F1 usually sits in empty data cells that nobody is willing to read as empty.

Across forty-one years of watching this industry, I remember clearly the era when strategy calls were made with human eyes. An engineer stood beside the car, listened to the engine, read the tyre wear and judged. Today the pit wall is a wall of screens: telemetry streaming in real time, timing loops buried under the asphalt, positioning signals, track temperature, and hundreds of measurement channels on every car.

Parallel to the technical layer sits a governance layer: the cost cap, aerodynamic testing restrictions, administrative investigations. Every major decision in this sport — whether cars go out, whether a team breached a financial ceiling, whether a driver is penalised — passes through a data pipeline.

That pipeline can fall silent.

In 2026, while working within the AC Milan coaching staff, I was assigned to audit the movement dataset of 20 Serie A matches from the 2026-17 season. Expected goals at home at San Siro read 1.85; away it read 1.02. An enormous gap. Yet actual goals scored were level. Cross-referencing the video, I found the culprit: a sensor in the south-west corner lagged by 0.2 seconds, corrupting the coordinates of every goalkeeper-led build-up. A 14-page internal report recommending recalibration went to the board. Head coach Vincenzo Montella used the findings to shift ball circulation to the right flank; the team won 5 of their last 8 matches and secured a Europa League place.

The lesson does not sit in the 0.2 seconds. If nobody cross-checks, the system still runs, still produces reports, still looks correctly formatted — and is still wrong.

Since then I have held one professional rule: every metric must be verified across at least two sources before it goes to air. Three years later, that rule paid me back on a night in Russia.

On 27 June 2026, Germany versus South Korea at the World Cup. On the 70th minute I posted on Twitter that Germany's defensive line was pushing an average of 68 metres high, that pressing had broken down 17 times, that South Korea already had 12 counter-attacks, and that without dropping the block, the goal would come from an aerial situation. In the 93rd minute, Kim Young-gwon scored exactly that goal. Thousands of accounts mocked me for turning emotion into arithmetic, but Gazzetta dello Sport still republished my distorted-trapezoid diagram of Germany's defence.

What I took from it sits elsewhere: numbers must be translated into spatial images before they stick. I stopped writing about a 68-metre high line and started writing about a zip that had burst open at the valve box.

And here the story returns to F1.

The operational data layer

Qatar, October 2026. Losail posed a problem data could not solve on its own. Pirelli analysed running data and found sidewall separation on tyres when cars rode the high kerbs at speed. The conclusion arrived inside the weekend: a maximum of 18 laps per tyre set. An ordinary race became a compulsory pit-stop puzzle, where strategy was locked down by a single safety number.

The same weekend produced another contested data layer: track limits. Dozens of lap times were deleted in qualifying because automated measurement registered a wheel crossing the line. Technically, the system was right. Sportingly, it produced a paradox: the fastest drivers were the ones losing the most time.

A measurement system only answers the question it was designed to answer. It cannot distinguish intent from a wheel pushed wide by turbulent air.

The weather and commercial data layer

Spa-Francorchamps, 29 August 2026. The race was declared underway, the cars ran two laps behind the safety car, then stopped. The result stood and half points were awarded. Max Verstappen was credited with the win.

The argument afterwards centred on television, rights fees and the credibility of the organisers. But what Spa truly lacked was a category of data that has never been standardised: human visibility in heavy rain. Wind speed is measurable, rainfall is measurable, humidity is measurable. The feeling of a driver running behind another car at 200 km/h with water thrown up in a white wall has no channel recording it.

An empty grandstand does not kill a race, but it removes something numbers cannot measure. At Spa that year, what was removed was consent. When spectators cannot react, only the organisers are left judging their own case.

The human data layer

Abu Dhabi, 12 December 2026. The safety car appeared in the closing stages, and the handling of lapped cars decided the title between Max Verstappen and Lewis Hamilton. Mercedes protested; the FIA rejected it. But by February 2026, the FIA itself restructured race operations: the race director role was split, and direct radio lines between team bosses and the race director were ended.

People usually see a legal argument there. I see something else: a volume of data pressing down on one human being. The race director had to watch screens, listen to twenty drivers' radios, handle requests from twenty teams and decide within seconds. That is a cognitive overload problem, and no telemetry board measures it.

The financial data layer

In October 2026, the FIA published its cost cap findings for the 2026 season. Red Bull were fined 7 million US dollars and handed a 10 percent reduction in aerodynamic testing for 2026. The 2026 cost cap stood at 145 million US dollars across a base of 21 races.

The telling part is the shape of the penalty: it hits development time, not money. That is deliberate, because in F1 an hour of wind tunnel time is worth an enormous sum. Aerodynamic testing allowances are allocated inversely to championship position: the weaker the team, the more it may run. The aim is to flatten the gap.

But financial data carries an inherent weakness: it is declared, then audited, then published. Between those three steps lie many months, sometimes more than a year. That lag means the penalty lands later than the season it was meant to correct.

Three layers, one common thread

Placed side by side, a pattern appears. Qatar: an accurate system lacking context. Spa: context present but no measurement channel. Abu Dhabi: enough data but an overloaded processor. The cost cap: correct data arriving late.

Every collapse has a precondition; few people are willing to look before it happens. At Las Vegas 2026, the precondition was a water valve cover forgotten on a checklist. At Spa 2026, the precondition was a commercial agreement signed before anyone asked how the drivers actually felt.

And this is where the instinct of someone who audits data for a living speaks up.

The real blind spot

This industry teaches people to trust numbers. A new engineer joining a team is trained not to argue with data. That is right most of the time. But it breeds a dangerous habit: when there is no data, people assume there is no problem.

That is the actual blind spot.

A team allocates millions of euros to sensors, simulation and laboratories. Very little is allocated to checking where those sensors are silent. The machine only measures what it was designed to measure, and nobody is paid to ask which areas remain unmeasured.

The consequence spills into matters that look outside engineering. A driver returning from major surgery is assessed on reaction time, on neck-load data, on fitness tests. Those metrics can return 98 percent. But the feeling before turning into a corner at 300 km/h cannot be captured by any sensor. That hesitation does not appear in the data table; it appears in the tone of voice on the radio, in the half-second delay before a driver answers the engineer.

Data only tells part of the story; the rest lies with those who know how to listen. That driver does not say he is afraid. That driver says the tyres are not right. A good engineer is the one who hears the difference between those two statements.

It is also why I never place a metric on an altar. Every tracking number belongs on a dissection table, not on an altar. A dissection table means cutting it open, cross-referencing it, establishing the conditions it was born in, which device produced it, at what hour, and who read it first.

This method is slow. It does not suit a ten-minute news cycle. But in a sport where the gap between two cars in qualifying can be under a thousandth of a second, a 0.2-second error is not a small detail. It is the entire story.

What to watch

The solution to this problem does not lie in fitting more sensors.

F1 has proven over more than a decade that it can build the best measurement systems in motorised sport. What is missing is a different discipline: auditing silence. That means, before every weekend, someone must be accountable for answering the question of what has not been measured, rather than only presenting what has.

The last three years show movement. The FIA accepting the human factor into race operations, Pirelli willing to impose a hard 18-lap limit rather than trust a model, teams speaking publicly about driver mental health — all are steps in that direction.

But a large gap remains: nobody is formally accountable for the empty cells.

And when a data pipeline stops speaking, the most dangerous thing is not the noise. It is silence being read as safety. If the sensors go quiet for a weekend, who will be brave enough to say that they are quiet?