F1 2026 and the Analytics Problem of an Empty Dataset

In December 2026, in a small studio in Hamburg, an editor slid a sheet of paper across the desk with exactly three lines on it: who wins 2026, which team breaks out, which team collapses. I opened my personal coding sheet — more than four thousand rows spanning nineteen seasons, every formation, every pit window, every race sequence. The last column, the one labelled 2026, was completely empty. Not a single lap of the new rulebook had been run. Not a single car built to the 2026 specification had completed a real lap on real asphalt. I handed the sheet back and said I could talk about 2026 for twenty minutes, but everything I said would be an unverified hypothesis. The defeat at Luzhniki taught me what victory never will.

Context: a new rulebook, an empty dataset

2026 opens the biggest regulatory cycle since 2026. The new power unit splits output almost evenly between the internal combustion engine and the electrical system, with the electric side rising to roughly 350 kW; sustainable fuel becomes mandatory; active aerodynamics replace the role of traditional DRS; cars are significantly shorter, narrower and lighter than the 2026 generation. The manufacturer list changes too: Audi takes over Sauber, Ford partners with Red Bull Powertrains, Honda moves to Aston Martin, Alpine switches to Mercedes power, and Cadillac joins as the eleventh team.

Every one of those changes is a variable that has never been measured on a racetrack. This is the point I want to stress: every 2026 analysis currently in circulation rests on three data sources — simulation, wind tunnel and engine dyno. All three have value, and all three share the same historical flaw: they correlate poorly with real racing in the opening phase of a regulatory cycle.

Look back at recent rule changes. In 2026, almost nobody in the commentary world named the dominant team correctly, even though Mercedes had been quietly preparing its hybrid power unit for years before the regulations were published. In 2026, wider tyres and higher aerodynamic load created a new order that the models of the time failed to predict. In 2026, ground effect returned: Ferrari started stronger than expected, Red Bull won the long development race, Mercedes fought bouncing for half a season.

Three cycles, three scenarios, one common thread: the team that reads the rulebook earliest wins, and reading the rulebook early leaves almost no trace in public data.

What the data can actually say — and what it cannot

Based on my experience tracking and coding race data across multiple regulatory cycles, I split every F1 prediction into two groups. The first group is standardised variables: pit-stop counts, stationary times, tyre compound distribution, top speed on the straights. These carry high confidence, but they only work once a cycle is at least eight to ten rounds old.

The second group is variables that exist only in the interaction between human and machine: how a driver deploys electrical energy on corner exit, how he chooses the moment to trigger active aero without losing stability on a fast entry. This group decides championships, and this group cannot be modelled before the first race.

I learned that distinction from another sport. When Tokyo 2026 brought athletics into my beat, I recorded Marcell Jacobs winning the 100m in 9.80 seconds — a result almost no forecast model had placed among the favourites. At the same time, at the Euros, I was tracking Leonardo Spinazzola and realised that Jacobs's stride model let me quantify the acceleration of a full-back pushing high. From that I built a flank acceleration index shared by both sports. The track and the pitch are not opposites; they are two rhythms of the same heart. But both taught me the same lesson: an index only means something when it is measured on the actual subject, under the actual competitive conditions.

In 2026, when the Bundesliga returned in empty stadiums, I had what amounted to a near-perfect natural experiment. I compared 82 post-lockdown matches with 82 pre-pandemic matches. The home win rate fell from 42.9 percent to 33.3 percent; average goals dropped by roughly 0.4 per match. Clean results, because one variable had been removed from the equation entirely. Empty stands, and home advantage becomes a number that does not round. The 2026 rulebook does the opposite: it adds a whole set of new variables to the equation and leaves no control group behind. There is nothing to compare against. When the stands are empty, sport strips off its shell and exposes its skeleton — and when the rules change, that skeleton is replaced by another one nobody has seen yet.

The contrarian angle: the power unit is not where the fight is decided

There is an almost unanimous consensus in the analysis community: whichever team solves the 2026 engine problem first will win. I think that consensus is technically correct and strategically wrong.

The power unit is the most tightly regulated item in the entire rulebook. Output is capped, fuel flow is capped, mass is capped, upgrade count is capped, and most importantly, the power unit is the one item every manufacturer can simulate almost perfectly on a dyno. It is therefore the item where the gaps between teams will close fastest. Four manufacturers arrive at once with four different philosophies but the same narrow technical corridor, and a narrow corridor always produces convergence.

The real differentiation sits at the interface between active aerodynamics and energy management. That is the zone the rulebook cannot legislate in absolute numbers, because it depends on how a driver makes decisions in thousandths of a second at each corner. A team can have the best engine and still lose the title because its driver cannot read the moment to close the wing. A team can have the third-best engine and still win because its driver turns energy deployment into a repeatable skill.

I have to add something about the cost cap here, because it completely reshapes the contest. When you cannot spend money to fix a wrong concept, being wrong becomes many times more expensive. The 2026 season will not reward the team that spends the most, but the team willing to pick a direction and hold it. Teams are also locking up their lead drivers on multi-year deals, a strategy that shows they are buying access to the new cycle's training data rather than buying past results. The transfer market does not buy the present; it buys promises about the future.

The blind spot of preparation

One notable detail few analyses mention: the 2026 development window opened while the calendar was compressed and wind-tunnel allowances were allocated according to the previous year's results. That means the more a team wins in the current cycle, the less wind-tunnel time it has for the future project. The mechanism is designed to flatten outcomes, but it also creates a paradox: the teams forced to gamble hardest are the teams with the least verification data.

So I set out three scenario branches for the opening phase, with relative probabilities after reviewing the history of the last four rule changes. Branch one: a midfield team reads the rules earliest and leads for the first three to five rounds, roughly 30 percent — the Brawn 2026 scenario. Branch two: a big team finds a small advantage it can repeat and turns it into a long winning streak, roughly 45 percent — the Mercedes 2026 scenario. Branch three: the order flips repeatedly through the first ten rounds because teams are still learning their own rulebook, roughly 25 percent.

I do not believe in luck; I believe in numbers lined up straight. And right now, the 2026 numbers are still scattered across the table.

What to watch

The greatest defeat is learning to read the game before it begins. For 2026, reading the game starts with something very concrete: the first test session in which a complete car runs a full sequence of consecutive laps. Not the fastest time. The consecutive lap count. That is the only indicator that tells you a concept survived the wind tunnel and entered the real world.

The 2026 season will not be decided by who predicts best this winter, but by who dares hold a hypothesis long enough for the data to answer.