Procedure used to increase the precision of the estimates that can be obtained for a given number of iterations