F1 testing restrictions: the rules that ration track time

Sakhir, the second week of February, and a Formula 1 car leaves the pit lane in anger for the first time. Three days later it will be crated up and will not turn a wheel again until the first grand prix weekend. That is the entirety of a modern team's on-track preparation, and it is not a quirk of scheduling. The F1 testing restrictions written into the Sporting Regulations govern, day by day and kilometre by kilometre, how much any competitor is permitted to drive, and the permitted amount is a fraction of what any engineer would choose.

The short answer, before the detail. Running the current car anywhere other than a grand prix weekend is prohibited. Pre-season running is capped at a small number of published days that every team shares at the same circuit. Cars several seasons old may be run under a separate regime with its own day and mileage ceilings. Promotional running is allowed on a tight mileage limit and on tyres made for the purpose. Wind tunnel and computer simulation work is rationed by championship position, with the leader receiving the smallest allocation. All of it is paid for out of a capped budget, which means a team can sit inside every sporting limit and still be unable to afford the running.

The activity the ban actually removes

The prohibition is narrower than most people assume, and understanding its edges explains almost everything teams do around it.

What is banned is track running of a car built to the current technical regulations, by a competitor, outside the events on the calendar. The regulations define this as testing of current cars and then permit it only in a short list of named circumstances: the pre-season test itself, any in-season test the governing body specifically organises, tyre supplier development running conducted under the supplier's own arrangements, and a small number of demonstration and promotional activities that carry their own restrictions.

Everything else is out. A team may not take last month's car to a private circuit to try a new floor. It may not rent a runway to check a cooling change. It may not run a straight-line aerodynamic session, which was once a standard part of a development programme and is now simply gone. The rule reaches the activity rather than the venue, so there is no jurisdiction to escape to and no arrangement with a friendly circuit that makes it legal.

That completeness matters, because the previous generation of testing rules tried to cap rather than forbid, and caps invite argument. A limit expressed in kilometres invites a discussion about whether an installation lap counts. A limit expressed in days invites a discussion about how long a day is. A prohibition on the activity itself has one question to answer, which is whether the car ran, and the answer is never in doubt.

The arms race that made a prohibition look reasonable

For most of the sport's history, testing was where the season was actually won.

A well-funded team would maintain a second operation whose entire job was to drive. It had its own cars, its own trucks, its own engineers and its own drivers, and it lived at a European circuit for most of the year. A part conceived on Monday could be on a car on Wednesday and understood by Friday. A team without that second operation waited until a Friday practice session at a grand prix, where the track was busy, the programme was compressed and every mistake was public.

The gap this produced was not small, and it was not a gap in cleverness. It was a gap in the number of experiments per unit of time, which is the variable that decides how fast any engineering organisation improves. A team running four hundred laps a week and a team running none are not competing at the same rate, whatever the relative quality of their engineers.

By the late 2000s the cost of maintaining that second operation had become one of the largest single lines in a top team's accounts, and the smaller entrants were arguing, correctly, that it was pricing them out of the sport. The teams themselves agreed to stop, and in-season testing with current cars disappeared from the start of the 2009 season as part of a broader package of cost reductions. It has never returned in any meaningful form.

The reason it has not returned is worth stating plainly. The ban did not merely save money; it removed an advantage that scaled with money more directly than almost anything else in the sport. A budget ceiling limits what a team may spend. A testing ban limits what spending can buy. The two instruments do different work, which is why the modern rulebook uses both at once.

Pre-season testing is now a handful of shared days

What remains before the first race is a small, public, communal exercise.

In a settled regulatory year, pre-season testing has been reduced to a single test at one circuit lasting three days, held at a venue with reliable February weather, with every team present at the same time. Track time is not allocated per team; the pit lane opens and the circuit belongs to whoever is on it. A team with two drivers therefore has to divide three days between them, against a programme that includes systems checks, aerodynamic correlation runs, long runs for reliability, pit stop practice and tyre work, in whatever weather the week provides.

A year with new technical regulations is treated differently, because the risk of arriving at a first race with cars that do not run is real. The transition to the 2026 rules was given three separate tests rather than one: a first at Barcelona in late January, where the test ran across five days but each team was permitted to use only three of them, then two further tests in Bahrain during February. That is a substantially larger allocation than a normal winter, and it is still less running than a single team would once have completed in a fortnight.

• Days the test ran
• Days one team could use

Barcelona, late January53

Bahrain, mid February33

Bahrain, late February33

Days of testing approved for the 2026 pre-season programme. At the Barcelona test the circuit was booked for five days and each team was limited to three of them; the two Bahrain tests were three days each and open throughout. The figures describe the schedule, not any team's running.

Show the numbers

The 2026 pre-season tests, scheduled days against usable days

Item	Days the test ran	Days one team could use
Barcelona, late January	5	3
Bahrain, mid February	3	3
Bahrain, late February	3	3

The communal format has a second effect that is easy to miss. Because everybody is on the same circuit in the same conditions, every team's data collection includes its rivals. Timing screens, tyre choices, fuel loads inferred from lap time traces, the sound of an engine on a long run: all of it is public information gathered from the same pit wall. Testing in Formula 1 is not private any more, and teams behave accordingly, running programmes designed partly to be unreadable.

Testing of Previous Cars, and why it needed its own rulebook

The ban covers cars built to the current technical regulations. It does not cover an old one, and that gap became a regime in its own right.

Running an older chassis lets a team do things it genuinely needs to do and that a grand prix weekend cannot provide. A reserve driver can accumulate real mileage so that a mid-season call-up is not a first drive. A junior can learn how a car with this much downforce actually behaves before being asked to qualify one. A new race engineer can practise the radio discipline of a stint. A pit crew can rehearse. None of that is development work, and all of it is legitimate.

The difficulty is that a car three seasons old is still a Formula 1 car, and a team that runs one for thousands of kilometres learns things. Correlation between the tunnel and the track, tyre behaviour, brake material performance, validation of the driver-in-the-loop simulator: an old chassis is a usable test bed for all of them if you are careful about what you conclude. So the regime around previous-car testing has been tightened repeatedly, and it now looks less like a loophole and more like a licensed activity.

The current shape of it runs roughly as follows. The car must be genuinely old, drawn from a window of seasons preceding the current championship rather than merely being last year's machine. The team's total is capped at twenty testing days across the season. A driver entered in the championship may take only a small slice of that, limited both in days and in distance, so a race driver cannot spend the summer accumulating mileage while a rival sits at home. Running is permitted only at circuits holding an FIA Grade 1 or Grade 1T licence, which restricts the venue list to real racing circuits with the appropriate safety infrastructure. There are also calendar restrictions that stop a team turning up at a track shortly before the grand prix held there.

• 20Testing days a team may run in a season
• 4Of those days, the most a race driver may take
• 1000Kilometres a race driver may cover in total
• 2Circuit licence grades permitted

Structural limits from the FIA Formula 1 Sporting Regulations governing testing of previous cars. The day and distance caps are per season; the licence grade figure counts the two circuit classifications the rules accept.

The four-day, thousand-kilometre limit on a championship driver shapes team behaviour most. It is generous enough to be useful after a mid-season change of car concept and far too small to constitute a development programme. A driver can use it to relearn a braking technique or to bed in after an injury. A driver cannot use it to test parts.

One further wrinkle deserves mentioning, because it produced an unusual sight. When the tyre supplier needed to develop compounds for a completely new generation of car, no current chassis existed to develop them on. The solution was a set of modified older cars, built on chassis from several seasons before, ballasted and adjusted to approximate the loads the new regulations would generate. Running these was restricted to drivers with substantial recent experience of a current car at genuine racing speed, on the sensible grounds that tyre development data gathered by someone still learning the car is worthless. Those cars were testing devices in the literal sense, and the rules treated them as a separate category rather than pretending they were something else.

Filming days, and the difference between running and testing

The other permitted activity is promotional, and it exists because the sport is a commercial enterprise that needs footage of cars moving.

A promotional event lets a team run a car for a strictly limited distance, using tyres supplied specifically for demonstration purposes rather than the compounds raced at a grand prix. The distance available across a season is measured in hundreds of kilometres rather than thousands, which is well under two grand prix distances and typically split across a small number of separate outings. The demonstration tyre requirement does the real work: with a tyre that behaves nothing like a race compound, lap times mean nothing and no useful setup conclusion can be drawn from them.

Teams still value these days enormously, because what they provide is not lap time but confidence that the car works. A new Formula 1 car is an assembly of thousands of parts that have never been together before. The first time it runs, things are wrong. A loom is chafing, a sensor reads nonsense, a fuel pick-up unports under braking, a radiator will not bleed. Discovering that on a filming day in January, in a garage with the whole team present and no clock running, is worth vastly more than a couple of hundred kilometres would suggest. Discovering it on the first morning of a three-day pre-season test costs a quarter of the winter.

The line between promotion and testing is policed by the mileage limit and the tyre specification rather than by intent, which is the right way round. Nobody has to prove what a team was thinking. The car either ran within the distance on the correct rubber, or it did not.

Aerodynamic testing restrictions: rationing experiments rather than kilometres

Track testing is only half of the picture, and arguably the smaller half. The dominant development tool in modern Formula 1 is not a circuit. It is a wind tunnel and a computer cluster, and both are rationed.

The mechanism sits in an appendix to the Sporting Regulations and operates on a completely different logic from the track rules. Rather than banning an activity, it meters it. The calendar year is divided into aerodynamic testing periods, and within each period a team is entitled to a defined quantity of wind tunnel work and a defined quantity of computational fluid dynamics work. Exceed the entitlement and you have breached the regulations, in the same way that running a car on a Tuesday would be a breach.

Writing that rule in an enforceable form is harder than it looks, and the definitions are where the effort has gone. Wind tunnel activity is counted in more than one currency at once: the number of individual runs, the time the tunnel spends with the wind on, and the total occupancy of the facility. Counting runs alone would invite one enormous continuous run. Counting time alone would invite a team to keep a model in the tunnel indefinitely between measurements. Counting all three closes the obvious routes.

Computational work is harder still, because the natural response to a limit on simulation is to buy a bigger computer. The regulations therefore express the CFD entitlement in a unit that combines the size of the problem being solved with the compute time spent solving it, so that larger hardware does not buy more allowance. A team with twice the cluster finishes its allocated work sooner and then stops. There is no way to spend money into more solving, which is exactly what was intended, and anyone who has tried to write a rule that survives contact with a hardware upgrade will recognise the difficulty.

What follows from all this is the restriction that decides how competitive the field becomes, because downforce is where the largest lap time gains in this generation of car have come from.

The sliding scale, and the date that decides it

The entitlement is not the same for everyone, and that is the whole point.

Each team receives a percentage of a common baseline, and the percentage is determined by its position in the constructors' championship. The team at the top receives the smallest share. The team at the bottom receives the largest. The steps between positions are defined in the published table, and the spread between first and last is wide enough to matter across a season rather than being a token gesture.

The allocation is not fixed for the year either. It is set initially from the previous season's final classification, then reset partway through the season on the standings at a defined reference date. A team that improves during the first half of the year therefore loses tunnel time for the second half as a direct consequence of improving. A team that falls back gains it.

This is a handicap system, and it is unusual in a sport that otherwise treats handicapping as anathema. The justification is that aerodynamic development compounds: a team that is ahead learns faster, which puts it further ahead, which lets it learn faster still. Left alone, that loop produces a field that spreads out over a rules cycle rather than converging. The sliding scale attacks the loop at the input rather than at the output, which is why it changes the shape of a season more than any points adjustment could.

The obvious objection is equally structural. At the reference date, a team sitting on the boundary between two steps has a measurable reason to prefer the lower position. No team will ever admit to it and no stewards' investigation will ever prove it, but the incentive is printed in the regulations. Any rule that pays for failure in a resource has that shape, and this one is no exception. The counter-argument is that the reward for a step down the table is smaller than the reward for a place in the championship, and that teams behave accordingly. It is a reasonable argument. It is not a proof.

1. 1An idea enters the queueAn aerodynamicist proposes a change. Because the number of experiments is capped rather than the number of ideas, the first constraint is internal: a team generates far more candidates than it can afford to evaluate, so a triage process decides what is even worth testing.
2. 2Computational screeningMost candidates are killed here, in CFD, because a simulation costs less than a tunnel run. The allocation is measured in a unit that combines problem size with compute time, so a team cannot buy its way to more screening by adding hardware.
3. 3Wind tunnel validationSurvivors are made as scale models and run in the tunnel, where runs, wind-on time and facility occupancy are all counted against the entitlement. The percentage of the baseline a team may use is set by championship position, so the leading team reaches the end of its allocation first.
4. 4Correlation against realityTunnel and CFD results are checked against measurements taken from the real car, which under the testing ban means data gathered during a grand prix weekend rather than at a test. A team whose correlation is poor is spending its restricted allocation on answers it cannot trust.
5. 5Costing against the capThe part is priced. Design hours, tooling, manufacture and the operating cost of the facilities that produced it are all relevant costs, so an upgrade competes for budget against every other upgrade and against the spares needed to finish the season.
6. 6Manufacture and first runningThe part appears on a car in a Friday practice session, now the only place a new aerodynamic component can be evaluated on track. An hour of practice therefore carries development weight that used to be spread across a week of private running.
7. 7Keep, revise or discardThe item is compared against the baseline configuration, usually run on the other car, and either adopted, sent back for revision or abandoned. Whatever the answer, the allocation it consumed is gone and the next idea in the queue moves up.

The sequence a development item passes through under the current restrictions. Each stage is metered by a different rule, which is why the pipeline is narrow at several points rather than at one.

Where the cost cap meets the testing rules

The two regimes are separate documents with separate enforcement, and they interlock in three specific places.

The first is straightforward. Testing costs money, and the money is capped. Running a previous car requires a truck, a crew, a set of spares, an engine with life left in it and a fortnight of somebody's attention, and every pound of it is a relevant cost under the Financial Regulations. The same applies to a promotional event and to the operating expenditure of a wind tunnel. A team may therefore hold a full entitlement of testing days and decline to use them, because the running would come out of the same budget as the upgrade it wants for the next race. That trade is real and it is made every season, and it is the sense in which a budget ceiling reaches into activities it never explicitly mentions.

The second is the sanction. An aerodynamic testing restriction is one of the penalties available when a team is found to have breached the financial rules, applied as a further cut to the percentage its championship position would otherwise earn. That makes the two regimes formally linked: a financial offence is punished in tunnel time. It is a well-designed penalty, because it takes performance from a team that acquired performance, in something close to the same currency, and because it hurts for longer than a fine. A team that loses a chunk of a season's aerodynamic development does not merely have a slower car that year; it arrives at the following winter understanding its own car less well.

The third is the treatment of facilities themselves. A wind tunnel is a capital asset, and capital expenditure is handled separately from operating costs under the financial rules, with its own ceiling. That combination is deliberate. A team cannot escape the running-cost cap by building a bigger facility, because the build is capped too, and it cannot escape the aerodynamic entitlement by building two tunnels, because the entitlement follows the team rather than the building.

Taken together, the three interlocks close the routes that money used to take. Spend more on running, and you hit the cost cap. Spend more on facilities, and you hit the capital expenditure limit. Run more experiments in the facilities you already have, and you hit the aerodynamic entitlement. Drive more, and you hit the testing ban.

What teams do instead, and what that changed

Restrictions of this severity do not reduce the amount of work a team does. They redirect it.

The largest single consequence has been the rise of the driver-in-the-loop simulator. A modern team runs its simulator through the night of a race weekend, with a driver on a different continent working through setup options and feeding conclusions to the pit wall before the next session. That activity is not track testing and is not restricted as such, though it is expensive and therefore capped in the ordinary way. It has become the place where most setup exploration happens, which is why the phrase about finding something overnight appears so often in team radio and press conferences.

The second consequence is a change in what a Friday practice session is for. With no private testing, the two hours before qualifying carry the load that a week at a private circuit used to carry: aerodynamic rake runs, back-to-back comparisons with a rival specification bolted to the other car, tyre baselines, fuel load calibration. This is why teams sometimes appear to waste a practice session on runs that produce no representative lap time, and why an upgrade is often fitted to one car rather than both. The session is a test, and it is very nearly the only one they have.

Days a team may run in a season20days

Days one race driver may take4days

Per-season caps from the Sporting Regulations. The team figure is a total across all drivers; the race-driver figure is the maximum any one championship entrant may take out of it, and carries a separate distance limit as well.

Show the numbers

Previous-car testing days, team allowance against race-driver share

Item	Value
Days a team may run in a season	20days
Days one race driver may take	4days

The third is a change in who drives. A team can no longer give a promising junior a hundred laps at a quiet circuit to see whether they can hold a car on the edge. The regulations respond by requiring teams to hand practice sessions to drivers with limited grand prix experience a set number of times per season per car, which forces the opportunity into existence rather than leaving it to a team that would rather not lose the running. The previous-car regime does the rest of the work, and a modern reserve driver's mileage comes almost entirely from an old chassis and a simulator.

The fourth is a change in risk appetite. A team that cannot test a radical concept before committing to it will be less likely to attempt one. Some of the more conservative design behaviour of the current era follows directly from the fact that a wrong answer cannot be discovered quietly in March and fixed by May. It has to be discovered in public, at a race, in front of the timing screens, and then unpicked across the remaining allocation. Teams that get a concept wrong under these rules take much longer to recover than they once did, and the way a weekend's tyre and setup work now compresses into three sessions belongs to the same story.

What the restrictions are not

Three things that are commonly said about F1 testing restrictions are not accurate, and correcting them makes the rest easier to read.

They are not primarily a safety measure. The mileage limits exist for cost and competitive reasons, and the safety requirements attached to testing, such as the circuit grade rules and marshalling provision, are conditions that apply to running rather than reasons for the limits.

They are not a limit on total engineering effort. A team's headcount is constrained by the cost cap, not by the testing rules, and the design office continues to work at full speed whether the car is running or not. What the restrictions ration is validation, which is the ability to find out whether the work was right.

They are not uniform across the sport, either. The aerodynamic entitlement varies by championship position, the pre-season allocation varies by year and can be adjusted for a manufacturer or team in unusual circumstances, and the previous-car rules distinguish sharply between a race driver and everybody else. Anyone reading a story about testing needs to know which of these regimes it concerns before the story means anything, and the answer is not always stated.

The engine question, which is separate again

One further limit interacts with all of the above, and it is administered under a different article entirely. Every kilometre a car covers, at a test or at a race, is charged against the life of the power unit elements a driver is permitted for the season. Running consumes mileage from a fixed allocation of engines, turbochargers and electrical components, and exceeding that allocation brings a grid penalty.

This is why teams are careful about how they spend even the running they are permitted. A filming day burns engine life. So does a previous-car test if it uses a current-specification unit, which is one reason those programmes run on older hardware where they can. The interaction is not exotic, but it explains an otherwise puzzling pattern: teams frequently decline available running, and the reason is often that the mileage costs more elsewhere than the data is worth. The full arithmetic of that ledger, and the way component allowances and their penalties are managed across a season, is a separate calculation with its own logic.

How to read a testing story properly

A few checks will tell you whether a piece of testing news means anything.

Ask which regime it falls under. A team running a car in March is doing one of three quite different things: a promotional event on demonstration tyres, a previous-car test on an old chassis, or a tyre supplier programme run under the supplier's arrangements. Each has different limits and different value, and a headline that says a team was testing rarely says which.

Ask who was driving. A race driver at a previous-car test is spending a very small annual allowance and is doing so for a reason. A junior or a reserve driving is routine.

Ask what tyres were on the car. Demonstration rubber means promotional running and means the lap times are meaningless. Anything else means the running falls under a regime with tighter conditions attached.

Ask where the team sits in the championship. The aerodynamic entitlement is set by position, and a team that has climbed the table since the last reference date is running on borrowed generosity that is about to be withdrawn. Improvement in the second half of a season is harder than improvement in the first, and the reason is written in the appendix rather than in the car.

Ask whether the cost cap makes the activity affordable. A team can be entitled to running it cannot pay for, and the reverse is never true.

Finally, watch the first hour of the first practice session of a race weekend, which is the closest thing to a test that now exists. Cars carrying arrays of pressure-tapped rakes behind the front wheels are doing aerodynamic correlation work, checking the real car against the tunnel. A team doing a great deal of that has a correlation problem, and a correlation problem under these restrictions takes months to fix rather than weeks. More on how the rest of the rulebook fits together sits in our motorsport section, and the wider archive of explainers is on the blog.