The three-stop wasn't faster – it was the only strategy with a stop left
Pirelli went a step softer to force more stops, and got the highest degradation we have measured this year. Ferrari answered with a soft start, a stop on lap 11 well before the window, fresh mediums while Mercedes sat on old hards – and Hamilton ran 2.2 seconds a lap faster than Russell. Yet the simulator says the three-stop finishes P3 without Alonso stopping on lap 38. What won Barcelona was that Hamilton, alone at the front, still had a pit stop to make when the VSC came. That was not luck. It was the price Ferrari had paid to keep the option open.
After five races in which the one-stop had become automatic, Pirelli went a step softer in Barcelona: C2 as hard, C3 as medium, C4 as soft, one step down from the year before, explicitly to increase the number of stops, according to F1.com. The circuit does the rest: fast long corners, old abrasive asphalt, big lateral loads on the left side – and the race had moved later in the calendar, into higher temperatures.
The teams had an advantage they lacked in Miami, Canada and Monaco: five days of winter testing at this very circuit in January. But January is not June, and old data could not simply be copied, according to F1.com.
Pirelli saw high thermal degradation on all three compounds, especially the front-left and rear-left. Soft and medium degraded at about the same rate, the soft 0.5–0.6 seconds quicker from new. The hard was the question mark: the few who tried the C2 – Haas and Verstappen – reported more sliding and surface overheating. Yet almost everyone else did the same thing: saved both hard sets for Sunday, to the surprise of Pirelli's Simone Berra, according to F1.com. The teams were saying, indirectly, that the C2 would decide the race.
After Saturday, with 53 degrees in the asphalt, Pirelli were firm: at least two stops. The main line was medium to hard to medium or hard, the alternative soft to medium to hard. Three stops – soft, soft, soft, medium – was surprisingly close in theoretical time; the drawback was traffic, three occasions to find clear track, according to F1.com.
Our model measured afterwards: soft 0.21 seconds per lap, medium 0.15, hard 0.11 – double Monaco, the highest so far in 2026. Tyre life: hard 33 laps, medium 28, soft 23.
Russell had pole, Hamilton P2, Antonelli P3, Norris P4. Mercedes had track position and could choose the medium. Ferrari needed to take the position and put Hamilton on the soft; Red Bull did the same with Verstappen further back, according to F1.com. The start gave nothing: Russell kept the lead, Hamilton P2.
Now Hamilton sat behind Russell on the tyre that degraded fastest. On lap 5 he was 2.4 seconds behind, on lap 10 3.5. The natural thing would have been to wait. Ferrari did the opposite: lap 11, soft to hard, 22.48 seconds in the pit lane, according to F1.com. Pirelli described it afterwards as a heavily advanced stop to maximise the undercut. Russell responded on lap 12, Verstappen the same lap, Norris 13, Antonelli and Piastri 14, Leclerc 16.
The undercut gave Hamilton two seconds but not the place: 1.4 behind Russell once both had stopped. Russell's medium was dying – a slope of 0.29 seconds per lap, the worst at the front – so one more lap on it cost him exactly what Hamilton's one-lap-fresher hard gained. On hard against hard Hamilton then had no chance: 2.1 seconds behind on lap 20, 2.4 on lap 25.
This is the strategic dividing line of the race. Mercedes ran two stops – medium to hard to hard; so did Norris and Piastri. Ferrari brought Hamilton in on lap 27, hard to medium, according to F1.com. The second of three stops, and the only medium at the front.
The effect is immediate: 23.4 seconds behind Russell on lap 28, 18.1 on lap 30, 11.1 on lap 33, 7.6 on lap 35 – 2.2 seconds a lap against a Russell on 15–23-lap-old hards. Mercedes knew Hamilton was on three stops and could have responded. Their model still said two stops was optimal, Andrew Shovlin confirmed afterwards, according to F1.com. The rational thing was to keep track position. The simulator partly agrees: had Russell mirrored Ferrari with mediums on lap 27 and a green third stop on lap 39 he would have finished P3, eleven seconds slower. Three stops was not faster than two. Not in itself.
Norris stopped on lap 35, Russell responded on lap 36, Piastri 36, Antonelli on lap 37, according to F1.com. Mercedes said afterwards that Norris's stop forced them to cover him. That was where Mercedes closed their strategy: two stops done, hard to the flag. McLaren likewise. After the stops Hamilton led Russell by 16.2 seconds on lap 39 – with 23 seconds still to pay. Net he was seven seconds behind, his medium was twelve laps old and fading: 81.9 against Russell's 81.2 on fresh hards. Ferrari were on course to lose.
The simulator is clear about what a green third stop would have given: lap 39 P3, 25 seconds back; lap 45 P3, after thirteen laps stuck behind the Mercedes. The three-stop on its own merits was a podium, not a win.
Fernando Alonso parked on lap 38. VSC laps 40–41. Ferrari brought Hamilton in at once: lap 41, medium to hard, the third stop, for 11.5 seconds instead of 23, according to F1.com. He emerged in the lead, 2.9 seconds ahead of Russell, on fresh hards against Russell's four-lap-old ones, with 25 laps to go. Then 0.6 seconds a lap: 19.6 at the flag. Norris third, Verstappen fourth.
It is easy to call that luck. It is wrong. Mercedes, McLaren – and Leclerc, who made his second stop on lap 39, one lap before the neutralisation – had all closed their strategies when the VSC came. Hamilton was the only one at the front who still had a stop to make, and he had it because Ferrari had planned three stops from lap 11. It is exactly the mechanism that gave Antonelli Suzuka: whoever can extend the stint holds an option on the next neutralisation. At Suzuka Mercedes got the option for free, from tyres that did not wear. In Barcelona Ferrari bought it for an extra stop – and paid in advance.
The simulator shows what the option was worth to the others: had Russell held his second stop until the VSC he would have won by sixteen seconds; Norris likewise. But that would have required knowing Alonso was about to stop, or choosing not to cover each other on laps 35–37. None of them did. Ferrari did not need to know.
Medium instead of hard at the VSC stop would have given the same result – Ferrari picked the right tyre for 25 laps. And the C2, which Pirelli had doubted on Friday, lasted 30 laps for Russell, if 3.5 seconds a lap slower at the end than at the start.
Verstappen also ran three stops – soft, medium, hard, medium on laps 12, 29 and 40, the last also under the VSC – and finished P4, forty seconds back, according to F1.com. Red Bull said the strategy was right, Verstappen that the car lacked the pace. The data agrees with both: his base pace on the hard was 0.6 seconds a lap worse than Hamilton's. A bad result is not a bad strategy.
Ferrari split: Hamilton three stops, Leclerc from P10 two – laps 16 and 39, according to F1.com. Pirelli praised the differentiation. But the second stop on lap 39 came one lap before the VSC: Leclerc was P2 on the road on lap 38, and the same stop a lap later would have been twelve seconds cheaper. He then retired with a power-steering problem on lap 62, so his result says nothing about the two-stop – but his lap 39 is the most expensive single lap of the race.
Antonelli had by then just passed Russell for P2 on lap 61 – and retired a lap later with a power-unit problem, according to F1.com. Second VSC, laps 61–63. Norris inherited the podium.
Judged on the information available when the decision was made:
- Pirelli – a step softer, at least two stops, three stops close in time: all right. The C2 worked better on Sunday than Friday had suggested.
- Ferrari/Hamilton – soft start, stop on lap 11, mediums on lap 27, three stops: not faster on paper, but the only strategy at the front that still had a move left. The race's best decision.
- Mercedes/Russell – two stops hard–hard and covering Norris on lap 36: rational by the model, and what closed the door four laps before the VSC.
- McLaren/Norris – the lap-35 stop that forced Mercedes: good, and the same closed door.
- Red Bull/Verstappen – Ferrari's strategy in a slower car: right, P4.
- Ferrari/Leclerc – lap 39: one lap.
- Friday's hard-saving – the whole paddock was right: the C2 was the race tyre.
Barcelona differs from everything we have analysed so far. China: make the hard last. Suzuka: one stop and track position. Miami: undercut timing. Canada: warm-up. Monaco: out-laps. Here Ferrari accepted the degradation instead of avoiding it, and used the extra stop not as a cost but as a weapon – and the weapon's real value was not the 2.2 seconds a lap on fresh mediums. It was that there was a stop left when everyone else was done.
Sources: F1.com (preview, Pirelli's Friday and Saturday comments, race report, pit-stop summary, quotes from Shovlin, Verstappen and Leclerc), press.pirelli.com (post-race analysis), FastF1 data for lap times, stint slopes, pit times and running order per lap. Simulator pit loss: 23.0 s under green (F1: 22.48 in the pit lane), 11.5 under the VSC. Lap numbers refer to the lap on which the driver entered the pits.
Simulate
Change one driver's pit stops and let everyone else run exactly as they did. The altered driver is modelled with his own pace per compound from the race, degradation, fuel, traffic and the VSC discount – cars that get stuck behind him lose time.