ABSTRACT
The importance of Monte Carlo methods as we have discussed before is that they enable us
to construct estimates of features of probability distributions which are impossible or very
difficult to compute analytically. Here “estimate” is not used in the sense we have before of
a quantity depending on data (which is random through the data being used) to estimate a
feature of an unknown probability distribution. Rather it is we who are generating random
quantities which can be used to estimate features of a probability distribution specified in a
completely known way.