F1 2026: The Biggest Advantage This Season Is Reading the Empty Data Cells Correctly

There is an empty cell in the winter 2026 dataset.

Eleven teams. Twenty-two cars. Three days of testing in Bahrain. Thousands of telemetry channels recording every millisecond of throttle travel, every degree of active-aero flap movement, every joule released by the MGU-K. Across that entire body of data, not a single line answers the one question that decides the championship order: what is the correlation coefficient between wind tunnel and racetrack for a 2026 aerodynamic concept?

No team publishes it. No team genuinely knows it. A new aerodynamic concept is only validated once it has run enough laps across enough circuit types, and that takes six or seven rounds — well into summer.

I have sat in enough timing rooms to know the lap chart always shouts louder than the data. At 60, I no longer believe in luck, only in the numbers that have not yet had time to speak. But there is one category of data that costs me more sleep than a wrong number: data that does not exist, quietly filled in with a plausible story.

Data is never in a hurry. People always are.

A season with no baseline

To see why that empty cell matters, place two parameters of the new rule cycle side by side. The 2026 power unit lifts electrical output from 120 kW to 350 kW, removes the MGU-H entirely, and splits energy roughly 50/50 between combustion and electric. Active aerodynamics replaces DRS with two configurations — high downforce and low drag — cutting downforce by around 30% and drag by up to 55%. Cars are close to 30 kg lighter, tyres are narrower, fuel is 100% sustainable.

Each of those changes severs a thread to the past. Lower downforce means the entire underfloor flow map that teams spent four years refining since 2026 becomes scratch paper. Active aero means the car has two aerodynamic personalities on the same lap, each needing its own dataset. Losing the MGU-H means the thermal management and energy deployment problem is rewritten from zero, and it feeds directly into how a team builds a strategy for the final twenty minutes of a race.

The 2026 season ran 24 rounds. So does 2026. But the whole dataset of those previous 24 rounds now has the value of a dictionary for a different language. Engineers can still read it, still understand it, but they cannot translate it onto the car sitting in the garage.

My method has not changed since 2026: every argument passes three gates. First the hypothesis — what happens if this concept is right. Second, cross-check against historical data from comparable cycles. Third, and only then, write it as narrative. The second gate is lethal this year, because no historical sample maps cleanly onto 2026.

The biggest empty cell is the correlation coefficient

The aerodynamic testing restriction — ATR — was designed to pull teams closer together. The reigning champion gets 70% of the reference allowance; the last-placed team gets 115%; a new entrant such as Cadillac gets additional allocation on top. Read that table and you see a fair flow of resources: weak teams get more swings, strong teams are forced to economise.

But runs are only an input. The value of each run depends on the correlation coefficient — the ratio between what the wind tunnel predicts and what the track delivers. A team correlating at 0.95 can use 70% of its allowance and still out-develop a team correlating at 0.70 using 115%. In the first year of a rule cycle that coefficient does not yet exist. It only forms once the car has run enough circuit types to separate the aerodynamic variable from the mechanical variable and the tyre variable.

This is where I strip noise away from structure. Broadcast coverage will read the Bahrain timesheet and declare that a team has "found" the concept. But the Bahrain timesheet is dominated by four things with nothing to do with the concept: fuel load, engine mode, tyre compound and track temperature. A quick lap in testing can come from 15 kg less fuel and an engine map that survives three laps. It predicts nothing about a race.

In the first year of a rule cycle, there is more data and less information — and the market still prices that pile of data as if it were information.

History gives me two clean comparison samples. In 2026, Mercedes entered the hybrid era with an advantage built from 2026: a split-turbo layout and a particular MGU-H packaging. Their rivals were not short of data — they were short of time. In 2026, when ground effect returned, teams that misread the floor concept lost most of a season fixing it, because every fix required a new upgrade package and a new measurement. Both cases reach the same conclusion: a wrong concept in year one cannot be fixed inside that year.

The cost cap turns every upgrade into an investment decision

The cost cap changed the nature of the game. With the ceiling set around USD 135 million per year in the 2026 phase and rising gradually afterwards, teams can no longer run two concepts in parallel to see which one wins. Before 2026, a big team could bring two different floor specifications to consecutive races and keep the faster one. Now every package is a capital allocation decision, and it must be made while the data is still empty.

I call it the crowding-out problem. A team that spends 20% of its development budget on a new floor concept between rounds four and nine has no room left for a rear-wing package. If the concept is wrong, they lose the season rather than a race. The cruellest part is that they will not know it is wrong until the money is gone, because a valid measurement needs the right circuit and the right temperature window.

There is an irony few mention: ATR gives backmarkers more wind tunnel runs, but the cost cap limits their ability to turn those runs into real parts. Running 45% more tunnel time without the money to manufacture and validate means the extra allocation stays inside a file. Two mechanisms designed to compress the field have quietly produced a new inequality: inequality in the speed at which data becomes metal.

Brentford and the discipline of the empty cell

In 2026, at 51, I spent three months profiling 1,247 players across 15 European leagues for a London sports consultancy, filtering to 38 targets on xG, PPDA and chance creation. I built a 12-indicator framework covering everything from high pressing to transition capacity. When Brentford signed Ollie Watkins from Exeter for GBP 1.8 million and later sold him to Aston Villa for GBP 28 million, the lesson was not that they found a good player. The lesson was that they had a process for handling missing data.

Within that dataset, some players had only 11 appearances in a lower division, no pressing metrics, no physical data. The easy option is to discard them. The second option is to fill the gap with instinct — "he looks sharp". Brentford took a third route: label every empty cell explicitly, record a confidence level, then decide whether it is worth paying to acquire more data. They sent people to watch exactly the matches that were missing.

Brentford do not read the future. They simply read data more carefully than everyone else.

I carried that principle into Formula 1. A team in 2026 holds millions of data cells, and a portion of them are empty: no tyre data at 45 degrees Celsius on a new asphalt, no race data in damp conditions on the low-drag configuration, no barrier-impact data for the new energy-absorbing structure. The team that labels those cells "unknown" and plans to fill them moves ahead. The team that fills them with an assumption and forgets about it pays with an upgrade pointing the wrong way.

The 2026 grid is locked; the real empty cells are in 2027

One detail goes largely unnoticed: the 2026 grid is essentially full, which is precisely why the transfer market has turned its lens to next year.

McLaren retain Lando Norris and Oscar Piastri. Ferrari retain Charles Leclerc and Lewis Hamilton. Red Bull retain Max Verstappen, contracted to 2028, and promote Isack Hadjar to the second car. Mercedes retain George Russell and Andrea Kimi Antonelli. Aston Martin retain Fernando Alonso and Lance Stroll. Audi, taking over Sauber, run Nico Hülkenberg and Gabriel Bortoleto. Alpine, on customer Mercedes power, field Pierre Gasly and Franco Colapinto. Williams retain Alex Albon and Carlos Sainz. Racing Bulls have Liam Lawson and Arvid Lindblad. Haas have Esteban Ocon and Oliver Bearman. Cadillac F1 Team, the eleventh entry, debut with Sergio Pérez and Valtteri Bottas on Ferrari power before a planned switch to GM engines from 2029.

Twenty-two seats. No genuinely open seat. And this is where the transfer market reveals its nature: when there is no vacant seat, people still manufacture rumours about vacant seats, because a rumour about a vacancy carries more commercial value than the fact that nothing has changed.

The transfer market is a contest in which whoever prices correctly wins.

Across four years as a transfer market administrator, I learned one rule: rank sources by motive, not by prominence. An agent saying his client is wanted by three teams has an obvious motive — to raise negotiating value. A team saying it is happy with its current line-up has an equally obvious motive — to relieve pressure from sponsors. Both are data, but data about motive, not about seats. Blending the two is the most basic error in reading transfer news.

The empty stands of 2026 and the lesson about collective noise

The empty stands of 2026 exposed something uncomfortable: much of what we called character was simply noise.

With no crowd, some drivers praised for "rising to the occasion" suddenly lost an advantage that had never lived inside the car. At the same time, some underrated drivers looked markedly more consistent. The lesson was not about who was better. The lesson was that an environmental variable had been counted alongside a performance variable for years, and only once it was removed could anyone see what remained.

2026 creates a similar test at the technical level. When every team loses its baseline at once, advantages accumulated over previous years — flow maps, simulation libraries, supplier relationships — stop showing up on the timesheet. For the first six rounds, pundits will praise a midfield team for "finding the concept". The better question runs the other way: or are they simply running well at one circuit type, in one temperature window, with a car 8 kg lighter than its rivals?

Every football cycle imitates the data of the previous cycle, and nobody learns. I wrote that about football, and it transfers to Formula 1 intact. In 2026, Brawn GP won by reading a loophole on diffusers more carefully. In 2026, Mercedes won by starting earlier. In 2026, Red Bull won by understanding ground-effect floors faster. Three different mechanisms, one structure: the winner of a cycle is whoever reads documents and data better, not whoever spends more.

The contrarian view: a reset does not democratise, it amplifies process

The most popular prediction for 2026 is that new rules will flatten the field. Big teams lose accumulated advantage, small teams get a chance, results become unpredictable, the title opens to five or six teams.

I take the opposite position, and it is arithmetic rather than posture.

First, a rule cycle cannot erase institutional memory. Mercedes began preparing its hybrid power unit in 2026, three years before it raced. Red Bull started studying ground effect before the 2026 rules were finalised. A team with a good process does not need more time — it simply starts earlier, and it always starts earlier.

Second, ATR and the cost cap do not permit fast correction. Under the previous cycle, a wealthy team could buy its way out of a development delay. Now a wrong concept in the first quarter means losing two-thirds of the season.

Third, and least discussed: 2026 has five power unit manufacturers launching new products in the same year — Ferrari, Mercedes, Red Bull Ford, Honda and Audi. Five manufacturers, five packaging philosophies, and nobody holds reliability data on a rival. The probability that one manufacturer has a durability problem in the first half of the season is very high, and when it happens, grid penalties arrive earlier than usual. Races decided by technical sanctions produce a distorted championship table, and a distorted table is the worst possible dataset for analysis, because it looks like a performance signal.

My conclusion: the structure of the 2026 title fight will be set before round eight. After the summer break the gaps will close, but closing once the season is decided carries statistical value only.

I may be wrong. But if I am, I want to be wrong because of data, not because I listened to the crowd.

Signals for the next cycle

Four things I will track over the next eight rounds, none of which appear on a timesheet.

One, the timing of the first major upgrade package. A team bringing meaningful aerodynamic changes before round six is confident in its correlation coefficient. A team waiting until after round ten is doubting itself, and that doubt is usually correct.

Two, the gap between the published claim and the measured gain. When a team says a package is worth "two tenths", compare it with the improvement in a dry session, stable temperature, same compound. The distance between those two numbers is the honest measure of a technical department.

Three, technical penalties. With five new power unit manufacturers, the count of permitted-component overruns will be a better durability indicator than any bench test.

Four, and most important, how teams handle their own empty cells. A team that says publicly "we do not know yet" is holding discipline. A team that claims to understand everything after three days in Bahrain is filling the cell with a story, and that story will appear on the timesheet as a 0.6-second gap at round twelve.

Forty years ago I learned this trade by writing every number into a paper notebook and comparing them by hand. Today I can receive ten thousand data points per second from a single car. The tools have changed completely. What has not changed is the temptation to fill a blank with something that sounds reasonable.

The 2026 season will not hand the title to the team that collects the most data. It will hand it to the team with the clearest definition of what it does not yet know — and the patience to go and collect that data before anyone else has thought to name it.