Up: Expected Values
Previous: Discrete Case
Answer
One can show analytically, by summing the series with k-th term
k*(1-p)^(k-1)p, that the expected value
of a geometric(p) random variable (the S definition of the distribution!)
is equal to ((1/p) - 1) or (1-p)/p. Thus, for p=1/N, the expected value is
N-1.
Albyn Jones
Tue Sep 12 11:16:10 PDT 1995