Zwift Ladder Race Predictor

Number of independent Monte-Carlo races. More runs give smoother win probabilities but take longer. 500 is a good balance; 100 is quick, 2000+ is very precise.

Per-race “day form” variation. Each rider's power is multiplied by a random factor drawn from N(1, σ) once per simulation (clamped to ±20%). 0 = identical races, 0.5 = realistic day-to-day variance.