Computation of metric 4.6 was stopped because the results could not be easily trusted. They depended on the caching of DS records by authorities, which is not ubiquitous. Before setting up another statistical estimate, we need to establish a ground truth measurement, for example by actually polling a set of reference domains.