For n > 3 fZn (z) can be approximated analytically or numerically. An analytical approximation for fZn (z) can be based on the Central Limit Theorem which roughly states that Zn approaches a Gaussian random variable with mean 0 and variance 1 as n → [infinity] (see p. 369-370 in the text for a precise statement and discussion). Hence a Gaussian pdf with mean 0 and variance 1 can be used to estimate the pdf of Zn. A numerical approximation for fZn (z) can be based on a simulation. Generate n × K realizations of a uniform random variable on (0, 1), and denote them by {xi, k : i = 1, . . . , n; k = 1, . . . , K} . {xi, k : k = 1,