OscarTangoEcho wrote:
Dignified Rube wrote:
His winning final speed estimate from simulation was 1.3%.
By the way, what the fuck does this mean. He had a 1% chance of winning? Where do you get these numbers?
It was less than a 5% total chance of winning for Whitmore. The win probability calculation comes from a simulation of speed figures (SPD + E1, E2 and LP) against those of the other horses and historical winning percentage, adjusted for stakes. I simulate the race 1000x times using BRIS speed figures with Monte Carlo simulation and exponential smoothing. The 1.3% winning final speed probability came from the simulation.
The person that helped me write the program is a former graduate professor of business statistics at Indiana U. and MIT grad in mathematics. He wrote the exponential smoothing part that I incorporated into the Monte Carlo simulation.