Portuguese Grand Prix 2023

Error: average positions off, compared with the calibrated grid (the grid turned into expected positions with a simple regression, no model). If the model does not beat it, its edge over the raw grid is only calibration.

Position crossing the line on each lap. The top six are in colour; the rest can be switched on in the legend. When a rider retires, their line ends.

Position crossing the line on each lap. The top six are in colour; the rest can be switched on in the legend. When a rider retires, their line ends.

Actual race pace: median of the clean laps, excluding the first lap, cancelled laps and pit laps. Degradation mixes tyre wear, fuel and traffic.

Race

Comeback of the day: Brad Binder, from P14 to P6.

Longest battle: Joan Mir vs Takaaki Nakagami (20 laps).

Overtakes are counted at the finish line: moves undone within the same lap are not seen. It is a measure of movement, not an official statistic.

Sprint

Comeback of the day: Raul Fernandez, from P20 to P11.

Longest battle: Brad Binder vs Raul Fernandez (11 laps).

Overtakes are counted at the finish line: moves undone within the same lap are not seen. It is a measure of movement, not an official statistic.

Official race direction log (Session.pdf). Times are local circuit time.

Rear compound in the race. Pace is relative to the best of the session; degradation includes fuel and traffic, so only compare within the same race.

Open-Meteo data at the circuit location during the race; not the track temperature.

Official classification of each session (Dorna). Best lap and the tyres used on it: official FIM PDFs. Long-run pace: median of the clean laps in long runs (5 flying laps or more); blank if the rider did none. Every rider runs their own programme and fuel load, so long-run pace only compares within the same session.

Estimated practice pace: median of the clean laps on the medium compound used in practice. It is an estimate: every team runs its own programme and fuel load. n = laps used (at least 3). * = only 3 or 4 laps: pace less reliable.

Sprint forecast based on the starting grid.

Forecast updated after the sprint: grid plus sprint pace.

The sprint is the only weekend data that makes us change our mind: it is real race pace, in traffic, the day before. Since 2023, adding it to the forecast cuts the average error from 2.7 to 2.5 positions.

Actual sprint pace: median of the clean laps, excluding the first lap, cancelled laps and pit laps. n = laps used.

Finish line S1 S2 S3 S4 · 4,546 m per lap

Approximate sectors: estimated from the time spent in each sector (typical error 5 to 10% of a lap) until we have the official split points.

© OpenStreetMap contributors

Fastest per sector: 2023

Median gap to the fastest rider in each sector within each session, across all seasons at this circuit; active riders only. The name on the layout is the sector leader. Sector split points on the map are approximate.

An open track: the grid decides less here and there are more comebacks.

When two teammates disagree between pace and grid, pace wins 100% of the time (between 21% and 100%; 1 cases).

Descriptive, over 6 races (2020–2025). 100 = the track where the grid matters most; 0 = the most open. Within a team, practice pace tells who will finish ahead 7 times in 10, but when the faster teammate qualifies behind, they finish behind 6 times in 10: grid position weighs more than pace.

2022 edition

Comeback of the day: Alex Rins, from P23 to P4.

Longest battle: Darryn Binder vs Fabio Di Giannantonio (16 laps).

Overtakes are counted at the finish line: moves undone within the same lap are not seen. It is a measure of movement, not an official statistic.

2022 edition

Net race pace: median of clean laps, excluding lap 1, cancelled laps and pit laps. Degradation mixes tyre, fuel and traffic.

2022 edition

Rear compound in the race. Pace is relative to the best of the session; degradation includes fuel and traffic, so only compare within the same race.