At Austin the soft tyre was slower than the medium — Hadjar's 'undriveable' complaint is not the real problem

Hook: the number that was never on the timing screen

From Isack Hadjar's second free practice at Austin came a piece of information far heavier than anything printed in a timing column: the soft compound gained time through sector 1 and then handed back more than it had gained between the end of sector 1 and the chequered flag. Over a full lap, the soft was slower than the medium. That is the kind of finding no team writes into a press release, because writing it down means admitting that the entire economics of qualifying has broken.

Hadjar was fourteenth in FP1, almost a full second behind team-mate Max Verstappen. In FP2 the car was called "undriveable", the front end described as being lost "massively", and the one-lap degradation characterised by the driver himself as "not normal". Any one of those three is a headline. But the headline went to the emotive word, and the actual evidence sat quietly in the inverted compound order.

I have spent twenty-two years reading paddock numbers, and I have carried one habit since 2026, when I built a cost-per-point model of the constructors' championship: an emotional word is never the bottom line. Force India's 187-point season taught the same lesson — the fastest story came from outside the garage and the truest one came from the balance sheet. Here, too. I followed the lap time back to the setup sheet, and the story changed. A lap time is a confession, and the inverted compound order is its clearest sentence.

Context: COTA's geography, the tyre's thermal arithmetic, and a narrowing door

COTA is an awkward combination. Turn 1 is a wide uphill hairpin demanding low-speed rotation; Turns 2 to 9 are the Esses, a sustained high-speed direction change where the front-left accumulates load for several seconds without relief; Turn 11 is a second hairpin; then a long back straight into a braking zone. Low-speed rotation wants a sharper, more aggressive front. The Esses cook exactly that front. That tension is the engineering problem of the circuit, and it is not a Red Bull problem in origin.

Tyre thermal arithmetic sits on top of it. A soft compound carries less thermal mass than a medium and a narrower working window. Cross that window and surface grip leaves, the car slides, and sliding generates more heat — a feedback loop that deepens itself once started. Hadjar's own words describe precisely this loop: tyre temperatures were "sky-high even though they were running the same speeds in those corners." That sentence is technically valuable. If cornering speeds were comparable and thermal load was not, the excess energy is being injected by the car's balance, not by the driver carrying more minimum speed.

Technical director Pierre Waché compressed the mechanism into one line: if the tyres are overcooked there, the lap is more or less finished. That reads like an excuse and is not one. It is a statement of circuit mechanics — a mistake in the Esses cannot be recovered elsewhere, because the lost time is created in corners that will not cool down again.

The weekend's architecture sits on top of all this. This is not a sprint weekend, so two of three practice sessions are spent and only FP3 remains. Parc fermé arrives after qualifying, which means the car's fundamental balance is locked where qualifying leaves it. Whatever protects the front tyres over a Sunday stint cannot be built after Saturday evening. Fifty or sixty minutes is the whole corrective window.

Core: the evidence chain

One. The FP1 baseline was never good, and this is the least reported fact. Hadjar said he was "happy-ish" after FP1 and believed he was heading in the right direction. He then changed a lot, and after FP2 described a "massive step down", saying the car needed to be brought back "at least" to its FP1 level. Read that phrase carefully: it treats FP1 as an acceptable target. In FP1 that target was fourteenth. The paddock had a narrative; the garage's data log had the receipts. The narrative was that the team made a good start and then broke the setup. The receipt says the start was not good either — its weakness simply lay on a different axis, the slower sections, where Waché himself says gains are needed.

FP1 was a morning session. COTA's surface is bumpy and the track carries little rubber early. A low-rubber track always looks fast and always lies about degradation, because the tyre slides, stays cool, and therefore wears slowly. By the afternoon rubber goes down, grip rises, load rises, and temperatures climb faster. Some of FP1's "satisfactory" balance was therefore a gift from track evolution rather than a reading of the car. This is an inference, and I label it one — the source supplies no temperatures, no sector times, no tyre allocation.

Two. The post-FP1 change was a multi-variable operation, and that is the real process error. Hadjar changed "a lot" after a session he had described as nearly satisfactory. On a thermally hostile circuit that is the most expensive mistake available, for a simple reason: change five parameters at once and if the car gets worse you own five suspects and zero sessions. You cannot say which change bit. That is precisely why, after FP2, Hadjar could not identify what had broken — he was reduced to saying the team needed to sit down and see what could be done.

This is not an indictment of the driver's ability. Setup direction is a joint decision between driver and engineering, and approving the scope of a change is the garage's responsibility. Elite teams expect a disciplined single-variable approach from a second driver across sixty minutes on a thermally marginal circuit. The process failed here, and that is more predictive than the raw gap. I do not chase the outlier; I audit the process that produced it.

One inference worth stating: the post-FP1 direction most likely added front-axle load or reduced rear stability — front flap, ride height, or mechanical front stiffness. That is the direction that helps low-speed rotation and pushes the front tyre past its ceiling through the Esses. A "massive step down" is consistent with an over-correction on a single balance axis. Medium confidence; not presented as fact.

Three. The front-tyre death in the Esses came from the car, not the driving load. The driver reported the fronts were finished after the first sector and completely killed in the high-speed corners, while adding that speeds through those corners were the same. If speed is equal and temperature is not, the surplus energy comes from balance: too much front-axle aero load, a front end sliding and generating friction heat, or a mechanical front stiffness combination loading the tyre's edge rather than its face. Waché's line gives the mechanism institutional acknowledgement, and it keeps the problem inside the car, which is where it belongs.

Four. The real headline: the soft was slower than the medium. The medium run felt okay; the soft run collapsed. The soft gave time in sector 1 and lost all of it and more afterwards. In a conventional weekend this is a qualifying-breaking condition. The entire purpose of the soft is to be faster than the medium over one lap. If it is not, the team either sacrifices grid position on the medium or burns a set of softs looking for a workaround. Neither is free. Note that this is far more serious than "undriveable", which is emotive and imprecise, because the compound inversion is a measured, repeatable failure. It also localises the problem geographically: the soft was quick in sector 1, so the failure was not in the launch or the first braking zone. It was in the Esses, where sustained high-frequency direction change drains energy from the tyre — exactly the mechanism Waché described.

Five. Verstappen's "fine-tuning" versus Hadjar's "undriveable" is setup divergence, not concept failure. Same team, same session, same base design. One side of the garage reports roughly a second and the word undriveable; the other reports fine-tuning and a largely positive start, with high-speed versus low-speed balance still needing work. If the RB22's concept were broken, both garages would be complaining. They are not. Verstappen's remark points to a car-level weakness that is controlled; Hadjar's problem is a different, acute, thermal one. The team is managing two distinct problems, and conflating them will misdirect the fix. Hadjar himself drew confidence from Verstappen's day, which is worth recording: the relationship is not yet adversarial.

Six. A genuine structural conflict: low-speed demand versus high-speed thermal ceiling. Waché says gains are needed in the slower sections. Low-speed rotation normally asks for more front-axle load and a more aggressive front. That same front load, through Turns 2 to 9, cooks the tyre. The same medicine is required by one organ and poisonous to another. The RB22's aero-mechanical balance window at COTA is therefore very narrow. That is not evidence of a wrong whole-car concept; it is a circuit-specific compromise in which you choose where to lose time. My read is that the correct answer concedes a little in sector 1 to keep the whole lap together.

A caveat is required. The source names the car RB22, which if accurate points to 2026-regulation machinery — narrower front tyres, reduced overall downforce, active aero. If so, the front-tyre thermal sensitivity observed here may be a category-wide phenomenon rather than a Red Bull defect, because narrower front tyres carry less thermal mass and overheat faster. I hold this under an explicit condition: if the designation is a transcription artefact, the inference lapses. The model stays quiet until a filing, fixture, or ledger contradicts it.

Seven. The session-allocation error: the soft was evaluated too late. The medium felt okay and the soft collapsed, which means the team discovered the soft's thermal ceiling during the qualifying simulation — late. On a thermally hostile circuit with limited sessions, the most fragile compound should be evaluated first so reaction time remains. Here it was evaluated last, in the most expensive session. The cost was an entire qualifying simulation and one soft set.

Eight. FP3's realistic target is asymmetric: durability, not pace. If the soft is slower than the medium over a lap, the optimal programme is not to make the car faster. It is to trade sector-1 headline time for front-tyre survival through the Esses. That will look slower on the stopwatch and be faster in fact, because a lap that completes beats a lap that ends in sector 2. Teams find this decision hard to explain publicly, because it requires admitting they are choosing to be slower.

Nine. Strategy exposure does not end at qualifying. A car that cannot keep front tyres alive across one push lap will suffer in COTA's long first stint, where the front-left decides the race. If Hadjar starts in the pack, expect an early-stop or aggressive undercut bias on his car. That is not strategy; it is damage limitation.

Ten. A process audit: who approved the change? When a driver requests a large change after a subjectively decent session, the race engineer holds a veto and should use it. Approving a multi-parameter change on a thermally marginal circuit is the garage's responsibility, and the post-session review will ask where the approval process loosened before it asks about the driver's performance.

Contrarian: the paddock's reading, steelmanned and then tested

The conventional reading deserves its strongest form. Red Bull's second seat has crushed young drivers; a driver fourteenth in FP1, a second off, publicly calling the car undriveable while the champion calls the same car fine-tuning, is a driver under pressure. All three supporting points are real, and a one-second gap is always a one-second gap.

Now the documents. First: the team-mate gap is F1's only valid reference, but only when the cars are comparable. Here they demonstrably were not — one in a self-inflicted setup dead-end, one in a known-good window. Treating the second as a pure talent signal is a serious analytical error, and it almost certainly overstates the true qualifying delta. A corrected setup should compress it substantially, though a residual gap to Verstappen at COTA remains likely. Second: the genuine concern is not the deficit but the diagnostic process. "A lot" of changes after a session called nearly satisfactory is the inverse of how experienced second drivers behave, and it is more predictive of future outcomes, because a gap moves with setup and a habit does not. Third: "undriveable" carries career risk disproportionate to its technical content. It is imprecise — it names no axis — and it creates a public contrast management will notice. That is narrative risk rather than sporting risk, but at Red Bull the two have historically been coupled. Fourth, and most important: is this Red Bull's problem or the category's? If this really is new-regulation machinery, front-tyre thermal sensitivity may be shared across every team in year one of a reset. That reading is harder because it names no villain, and the paddock wants a villain. Fifth, the correlation line: some of the FP1-to-FP2 regression is track evolution, not setup. Rubber going down and load rising increase degradation independently of any change. I cannot apportion the two without temperature data, and that is the honest position.

Two further points are usually skipped. Driver style is not zero in a thermal failure: a high minimum-speed, rotation-hungry approach amplifies front-axle thermal load. It is an amplification factor, not the root cause — the driver said corner speeds were the same — but it belongs in the ledger. And the evidence base itself is thin: no sector times, no track or air temperature, no tyre allocation, no car spec. What exists is two credible first-party voices and two hard numbers. I have kept my conclusions inside that boundary, and where I inferred, I said so.

Takeaway: what to watch in FP3

I will write these down before the session so I cannot invent them afterwards. One: which baseline Hadjar's car starts from — a return toward FP1, or toward Verstappen's FP2 setup. If the latter, the entire post-FP1 exploration has been written off, and that is the correct call. Two: the first soft push lap, and specifically the delta between sector 1 and sector 3. If sector 1 gets slower while the rest of the lap loses less, the compromise is working. Three: front-left behaviour on the out-lap. If the tyre is already hot on exit, the problem extends beyond setup into brake-duct or rim thermal management, and the whole FP3 agenda changes. Four: whether the team splits programmes. Verstappen's baseline is healthy, so Red Bull holds a natural hedge — a conservative, front-protective setup on Hadjar's car while Verstappen validates the performance-optimal direction. If they do not use it, that is itself a signal.

And the question larger than the weekend, to which I have no number. Red Bull's second seat has been presented for years as an engineering problem: a car built for one driver and not the other. Austin's Friday raised a less frequently asked one. If, on the same afternoon, in the same garage, one driver called the car undriveable and the other called it fine-tuning, are we really talking about a car — or about the process that produced two different cars in a single afternoon? I do not have the data for that. I have only this: the soft tyre was slower than the medium, and that is a receipt no press release can erase.