In the results presented below we first look at the benefits of using a threshold to decide whether to run the validity checking program, on the “small” problem with 10,008 members. As we had hoped, using a threshold in this fashion was of significant benefit. The next set of experiments presented here look at a simple technique for dynamically setting the threshold, again with very encouraging results. Finally we present a few results on the “medium” problem with 44,928 members, and a “large” problem with 79,848 members.