Part 1 of a series in which I implement standard normal distribution random number generation (ZRNG) and verification algorithms. Intended for a future project where I test the most efficient way to determine the best player in a race with the fewest trials. Currently I have implemented the Box-Muller transform and am verifying by comparing the mean, std, skewness and kurtosis with that of a Z distribution. I also calculated the RSME of those 4 attributes to aggregate that into a single error value. Does it make sense to do that? I don't know, but I did it anyway. Due to the simplicity of the Box-Muller transform and the way Scratch works it might be possible it already is the fastest algorithm that can be implemented. Nevertheless I will continue. ToDo: ZRNG algorithms: Box-Muller polar form Central limit theorem Inverse transform sampling Marsaglia polar method Ziggurat algorithm Verification methods: Shapiro–Wilk test Kolmogorov–Smirnov test P–P Plot Q–Q Plot Additional: Generation timing Speed improvements regarding screen refresh
