The circuits

Three numbers decide strategy at a circuit, and they age differently under the 2026 regulations. Pit loss is what a stop costs against the field: the in-lap and out-lap measured against the median lap of the cars that stayed out. It is a few tenths to a second longer than the pit-lane time F1 quotes before a race, because it also carries the approach and the rejoin. It is governed by geometry – pit-lane length and speed limit – so it carries over from 2024–25 unchanged. The overtaking threshold (the pace advantage in s/lap that gives a 50% chance of getting past within two laps) changed with the new cars, so we show the 2026 figure first and the pooled 2024–26 value in grey. Degradation (α, s/lap of tyre age) is measured from this season's dry race here. 2026 changed the cars and the tyre range together, so no season before it is used for tyre behaviour at all. Where the 2026 race has not been run yet we show a forecast built from this season's races at circuits of similar character - surface abrasion and whether the tyre is limited by cornering load or by traction.

Long stint (p90) is the 90th percentile of stint length the teams actually ran on that compound, not the single longest one – how far they were willing to go, not how far the tyre could have gone. The model applies a cliff beyond it.

≈ 0 marks a compound whose degradation could not be measured: either the tyre genuinely did not degrade (Madrid hard) or the race was driven to a delta in traffic so the lap times never showed it (Monaco). weak marks a slope with an uncertainty as large as the value. Wet races are excluded. Circuits with a single race in the data are marked.

2026 estimate marks an overtaking threshold for a circuit the season has not visited yet. Passing is easier under the new cars: across the eleven circuits raced both before and during 2026 the threshold fell by a median factor of 0.80, and the share of close encounters that ended in a pass rose by a factor of 1.5. The estimate takes the circuit’s own earlier threshold, scales it by that factor, and shrinks it halfway to the season median, because a single-race fit is noisy – fits with few completed passes land systematically high.