In both examples, the mean() is called in a loop and the results
are expected to be floats.  In the context of Monte Carlo
estimation, high accuracy rather than perfect accuracy will suffice.