10-16-2015, 01:56 AM
The Prime allows Binomial CDF calculations for high number of trials. Perhaps the Prime should reject number of trials in excess of 1E6 trials. For 1E6, the binomial CDF for 5E5 successes and p of an individual success=.5 yields .500388553668 (seems reasonable). By the time we try 1E7 trials, 5E6 successes, p=.5, the CDF is under .5 (.47+)! With 1E8 trials, 5E7 successes, p=.5, CDF is .03+. Outrageously bizarre is 1E9 trials,5E8 successes, p=.5 CDF is <-1.78!