Suppose that Y1; Y2; : : : is an output process with steady-state mean and that Y (n) is the
usual sample mean based on n observations. Consider plotting Y (n) as a function of n and
let l0 be the point beyond which Y (m) does not change appreciably. Is l0 a good warmup
period in the sense that E[Yi] for i > l0 and also that l0 is not excessively large? Why?