We repeat this procedure 10,000 times. The value of 10,000 was selected because 9604 is the minimum size of samples required to estimate an error of 1 % with 95 % confidence [this is according to a conservative method; other methods also require <10,000 samples size (Newcombe 1998)]