Pirelli were right about one stop – but the window was eight laps too early

Silverstone loads the tyres harder than any other circuit, above all the front-left through the fast right-handers – over 5 g through Maggotts, Becketts and Chapel. Pirelli therefore brought the three hardest compounds: C1 as hard, C2 as medium, C3 as soft. But the asphalt is not abrasive, and after the sprint Pirelli saw lower degradation than expected. Their simulation said one stop was around 13 seconds quicker than two, and that medium → hard with a stop on laps 24–30 was fastest, per F1.com.

They were right about one stop. Our optimiser, run afterwards with the degradation measured from the race, says the same: two stops 12 seconds slower. But it also says something Pirelli did not know on Sunday morning: the medium (C2) degraded less than the hard (C1) – 0.024 against 0.039 s/lap. With that degradation the optimal stop is lap 33, with a window within one second from lap 28 to 38. Pirelli's window was the right direction but eight laps too early.

We should be honest about our own model: it had no pre-race forecast at all for Silverstone. A sprint weekend means a single practice session, and nobody ran a long run there on anything but the hard. That is a gap we need to close.

Antonelli had pole, Leclerc took the lead at the start. Ferrari stopped Leclerc on lap 25 – inside Pirelli's window, but by the race's degradation 2.2 seconds from optimum. Antonelli kept going. Lap 26, 27, 28 … to lap 35, on a stint whose slope was zero: the C2 medium did not degrade at all. When he came in he was 0.1 seconds from optimum. What looked like a bold departure from the plan was in fact the time-optimal strategy. It was Ferrari who departed, on the early side.

Antonelli rejoined 7.0 seconds behind Leclerc with ten laps' fresher hards. In five laps he took back 3.3: 7.0 → 3.7 on lap 40. Then the left wheel shield broke, and Mercedes had to make three stops in six laps. That was the end of the fight for the win.

Would he have won? The data says no. 3.7 seconds on lap 40 with 0.66 seconds a lap in hand catches Leclerc around lap 46 – and to pass at Silverstone you need 1.14 s/lap, the second-highest threshold on the calendar. The Safety Car came on lap 48. Second place, not the win. Click a scenario to load it in Simulate below.

Red Bull went the other way: Verstappen in on lap 17 for hards, an undercut attempt on Russell. By the race's degradation it cost 8.4 seconds against optimum the moment it was made – and five laps later the VSC came on lap 22 and gave everyone who had not stopped a cheaper stop. The call was logical with Pirelli's window in hand; the race made it expensive. Verstappen went off at Stowe on lap 47.

Hamilton had a five-second penalty for a false start. Ferrari brought him in on lap 23, before Leclerc, although the tyres felt fine: penalty served, hards on, and then a strong drive back through the field, past Verstappen to P3 and P2 when Antonelli hit trouble.

Lap 48: Safety Car. Leclerc leads, Hamilton second, Russell third. Ferrari brought both in for softs; Mercedes left Russell out. The order became Leclerc – Russell – Hamilton, and Ferrari's thinking was clear: at the restart Hamilton would have fresh softs against Russell's older tyres. Fred Vasseur said afterwards that they were fairly convinced of a restart and that he would make the same call again, per F1.com.

The race never restarted. The simulator says the obvious: had Hamilton stayed out he would have been second. It is still a good decision with a bad outcome, not a bad decision – with two or three green laps the tyre offset would have been large. But it is also the only decision in the race that cost a place.

Silverstone adds a strategy type to the 2026 map: tyre offset through a long first stint. Spa was waiting for the neutralisation, Hungary the overcut in clean air, Zandvoort flexibility when degradation changed, Monza the alternative start plus VSC. The same principle does not work everywhere. But Silverstone also shows something about forecasts: Pirelli's fastest theoretical strategy can be right in kind and count and still have the wrong window – and whoever reads the degradation in real time, as Mercedes did with Antonelli, finds the right window while the others follow the plan.

How we did it: lap data from FastF1 (F1 Live Timing). Degradation = within-stint slope on green laps, fuel-corrected at 0.035 s/lap; the optimiser run afterwards with the race's degradation and a 20.0 s pit loss. The simulator runs everyone else on their actual lap times; under the Safety Car the order is frozen and a stop costs the SC loss in the ranking. Sources: F1.com (Pirelli's pre-race assessment and post-race note, Vasseur's comment), FIA timing.