I like the analysis, but I wonder if the criteria is week enough to detect the dupes from medium.com , because medium adds some tracking crap to the URL that confuses the dupe detector of HN. For example see: https://hn.algolia.com/?query=I%20Peeked%20into%20My%20Node_... (this list doesn't include many dupes that were detected and marked).
A problem with this analysis is that it doesn't count the dupes that never had a sibling that get to the front page. Counting it would modify the distribution of some domains and submitters.