Using this approach (and a binary decomposition of n - 1), it is possible to compute the p-value with O(log2(n) × L2) memory complexity and O(log2(n) × L3) time complexity. As L usually grows very fast when we consider more complex events En, these complexities are a huge drawback of the method. Moreover, numerical precision considerations prevent this approach to give accurate results when using the relation ℙ() = 1 - ℙ(En) to compute the p-value of the complementary event (as the absolute error is then equal to the relative precision of the computations).