Pit Stops, Laps and Lap Speed: What Actually Predicts Race Time

Working with 225 Grand Prix winners across 18 variables, the interesting result was how few predictors survived selection. regsubsets() minimized BIC at three: laps, number of pit stops and fastest lap speed.

Variance inflation factors stayed low across all three, which meant the model was not leaning on redundant information. Laps came in at 1.33, pit stops at 1.06 and fastest lap speed at 1.30.

After transforming the response with a negative fourth power and taking logs on laps and pit stops, adjusted R squared settled at 0.667. Feeding in 55 laps, 2 stops and a 202 km/h fastest lap produced a predicted race time of 100.13 minutes.

The strategic reading: circuit length and driver age add color, but race duration is dominated by how many times you stop and how fast the car can go when it is clear.