Simple and efficient algorithms for computing exact cumulative discrete probabilities and its inverses (Q1775983)

The goal of this paper is to introduce a rather universal approach to the calculation the cumulative probabilities for a wide range of important discrete distributions (e.g. the binomial, negative binomial, Poisson and hypergeometric distributions and their truncated versions). The proposed algorithm uses only the recurrence relations of the type \(P(X=k-1)=a_kP(X=k)\) and the mode of the random variable \(X\), where \(a_k\) are increasing. No starting value \(P(X_0=k_0)\) for any \(k_0\) is required. With a slight modification the second algorithm for calculation the inverse distributions can be obtained. Both proposed algorithms are simple (15 or 20 lines of code, respectively), very fast and numerically stable even if the sample size is of one billion.