2026 03 17

Our new paper, Benchmarking Formula 1 results using a normal model, is now available as a preprint. A long-standing question in sport is how to separate the effects of skill and luck. In Formula 1, this often comes down to distinguishing between driver performance and the performance of the car. At present, four elite teams dominate the sport and have won every major race over the past four years. In the paper, we use univariate and bivariate normal models to estimate realistic performance expectations at both the driver and team levels, distinguishing between elite and non-elite teams. This provides a more meaningful benchmark for Formula 1 results. We apply the approach to analyse the most recent fully completed F1 season (2025). You can read the paper here