the acceptance rate for non legacies is 3/90 or 3.33%, and the legacy rate of 30% is actually 9 times the non-legacy in your example. So two things, one is that the Harvard case should show the actual defacto acceptance for unhooked to be in the 2-3%, which your example shows. Second, why do you need to accept 3 legacies and 3 non-legacies, why not, you know, make it more fair and admit 1 legacy and 5 non-legacy? In that case, the legacy is 10% and the non-legacy is 5.5%, a smaller difference.