Various extensions to the usual Mathematica random number generators are provided:

This muddles a set:

Here is a random sample from the probability distribution, given as a list of 10 integers:

This generates 1000 random numbers drawn from a specified cumulative distribution, CDF[PoissonDistribution[1.5],#]&:

Here is a sample of 100 from the multinomial distribution, with probabilities {0.1,0.3,0.6}

This adds NegativeHypergeometricDistribution[n,p,α] to the set of statistical distributions:

This tests for deviations between observed numbers {7,25,8} and expected numbers {10,20,10}: