Hamilton leads FP3 at the United States Grand Prix: the most important gap was never published

At five in the afternoon on Saturday in Austin, the timing screen for the third practice session at Circuit of the Americas showed a familiar name at the top: Lewis Hamilton, in Ferrari red, with a fastest lap of 1:34.893. Behind him sat Max Verstappen in the Red Bull. Charles Leclerc, in the second Ferrari, was third at a gap of 0.234 seconds. The two Mercedes followed, with Andrea Kimi Antonelli fourth and George Russell fifth.

That is the entire data set. The gap between Hamilton and Verstappen was not published. Neither were the gaps for Antonelli or Russell. There is no tyre compound, no fuel load, no engine mode, no track or air temperature, no indication of when the fast laps were set, and no sector times. A headline says Hamilton edged Verstappen, but the source itself supplies no number to quantify the word edged.

The first task is to state plainly that this is not enough to conclude anything about the true competitive order at Austin.

COTA is the worst circuit for reading a practice session

Circuit of the Americas runs 5.513 km over 20 corners, and its structure seems designed to defeat inference from a single timing sheet. Turn 1 climbs roughly 41 metres, braking from over 300 km/h into the 80 km/h range. The run from Turn 12 back to Turn 1 is nearly 1.2 km of almost full throttle, meaning straight-line aerodynamic efficiency dictates much of the lap time. Immediately after comes the esses from Turn 2 to Turn 9, where the car changes direction repeatedly under high aerodynamic load, demanding stable downforce and suspension soft enough to keep the tyres in contact with the road.

Those two requirements conflict. A car set up for the long straight will struggle in the esses; a car set up for the esses will lose time on the straight. A quick lap at COTA therefore carries very little information without sector times. None were provided.

The surface is also famously bumpy. This is the most under-discussed technical variable in any Austin report, and it has a direct consequence: teams that run the car too low in search of downforce pay for it in plank wear. At the 2026 United States Grand Prix, both Hamilton and Leclerc were disqualified after the race for excessive plank wear despite finishing near the front on the road. That is the clearest precedent that at COTA the binding constraint is not finding more downforce but keeping it without breaching ride-height rules.

A Saturday FP3 also tells us something about the weekend format. Under a Sprint format there would be no Saturday FP3 and the report's premise would collapse. Its existence implies a standard weekend, meaning teams still have time to adjust setups before qualifying. The value of the session lies there, not in the order.

Why practice order is the weakest signal of the weekend

A practice lap is governed by at least seven variables at once: fuel load, engine mode, compound, tyre age, track temperature, track evolution, and each team's run plan. None of the seven appears in the source.

Consider two scenarios. In the first, Hamilton sets 1:34.893 on new softs, low fuel, maximum power, late in the session on a rubbered-in track, while Verstappen is second on scrubbed softs carrying race fuel. In that case the real gap could exceed half a second and mean nothing for qualifying. In the second, both set their times in identical conditions and are separated by 0.05 seconds, in which case the sheet reflects genuine pace.

These scenarios produce opposite conclusions and the source offers no tool to distinguish them. That is why I always read a practice sheet by run structure first and order second.

The absolute time deserves a second look. At 209 km/h average across 5.513 km, 1:34.893 sits well below the usual dry pole pace at COTA. That points to a damp or drying track, a red-flag interruption, or a deliberately conservative run plan across the field with heavy fuel and low power. All three lead to the same conclusion: do not extrapolate this order to qualifying.

Ferrari's upgrade: present, unvalidated

The only confirmed technical fact is that Ferrari brought an upgrade to COTA. No component, no design concept, no claimed wind tunnel or CFD gain, no before-and-after comparison.

P1 and P3 are consistent with a working package, but consistency is not evidence. The same result is produced by a team running low fuel on new softs while rivals run race-fuel simulations. With the available data, both hypotheses carry equal weight. Numbers never lie, but the people reading the reports do.

There is also opportunity cost. A late-season package consumes cost-cap room and aerodynamic testing restriction allowance that cannot be recovered within the season. If this weekend sits late in the current regulation cycle, the marginal return on any package is low while the opportunity cost for the 2026 transition is high. That Ferrari is still spending now says something about how the team reads its own championship chance. A leading team usually stops developing.

Ride height and plank wear: the forgotten variable

At COTA the key mechanical variable is not total downforce but the balance between ride height and permitted plank wear. The regulations set a minimum plank thickness and a maximum wear allowance. Austin's bumps cause repeated contact at low-speed corners and under braking into Turn 1. Every contact removes material. A team choosing a low ride height gains lap time in practice and risks disqualification on Sunday.

Practice order can therefore reflect a setup choice the team will have to abandon. A car fastest in FP3 on a slammed floor is not necessarily fastest in the race, because it will have to be lifted to pass scrutineering.

Mercedes: an internal order that should not become a story

Antonelli fourth, Russell fifth. A rookie ahead of the team leader is the most tempting headline of the day and the easiest to misread. No gaps were published, no tyre data, no run timing. A one-position gap between P4 and P5 sits entirely within the noise floor of a practice session. Mercedes has habitually run its qualifying simulation later than rivals, meaning its times are often set when the track is at its best. If Ferrari ran early and Mercedes late, P3 and P4 can reverse in qualifying without any technical change.

The more meaningful signal is that both Mercedes are in the top five. Two-car consistency matters more than intra-team order, because the constructors' title is decided by the combined score.

Red Bull and the missing car

Red Bull is represented by a single car: Verstappen, second. The other car is not mentioned. If it sits outside the top ten, that is a serious problem for a season-long constructors' campaign. If it is sixth or seventh, the omission is merely editorial. The source does not allow the two cases to be separated, and the correct response is to record the gap rather than fill it with speculation.

The biggest hole: McLaren

No McLaren appears in the top five. That is the most important fact in the article and the least discussed. McLaren is the benchmark of the current cycle. When the benchmark team is entirely absent from the front, there are two readings: a genuine collapse of up to a second per lap, or a completely different programme, whether heavy-fuel long runs, low engine modes to protect the power unit, or a disrupted session. The first is near-impossible in seven days. The second fits every other signal, including the unusually slow best time and the compressed front group.

What actually matters in qualifying

If FP3 order has little predictive value, the signal lies in run structure. First: which team set its best time on new softs and when. Track evolution at COTA across an hour is significant. Second: sector times. A car strong in sector one but weak in sector two has an aerodynamic balance problem, not a power problem. Third: plank wear across long runs, which dictates race ride height and therefore true Sunday pace.

The contrarian angle: narrative inflation

The original headline uses the word edges. But no gap is published, so the word carries no quantity. It is a hierarchy claim based on display order, not data. Sports analysis has a common failure mode I call narrative inflation: a high-uncertainty event retold as a low-uncertainty one so the story sells better. A practice session with an anomalous time and no tyre data becomes evidence of an upset. A rookie ahead of a team leader becomes a generational shift. A car missing from the front becomes a collapse.

I do not believe in luck. I believe in numbers verified three times. Verified three times, this Austin FP3 order does not pass once.

The opposite overreaction should also be avoided. Practice order is low-value data, not worthless data, with one exception: when the same team puts both cars at the front across several sessions, that is real. Ferrari at P1 and P3 is the strongest two-car signal here. It does not prove the upgrade works, but it is a fair starting point for the next question.

Takeaway

At Austin this weekend every judgement about the competitive hierarchy must be suspended until qualifying ends. Until then, the only things worth recording are an anomalously slow practice session, an unquantified upgrade, and an unexplained absence of the benchmark team. Motorsport is emotion, but a team exists on algorithms. And at Austin, the algorithm is missing inputs. The question for Sunday is not who was fastest in practice, but which team dares to keep its floor low for 56 laps and still pass scrutineering when the flag falls.