Though now thanks to AMP, it might make more sense to de-dupe on known prefixes rather than Host particularly. Sadly, for the BBC at least that won't work, since the URLs look like they do /amp instead of their usual sections:
https://www.google.com/amp/www.bbc.co.uk/news/amp/36788782?c...
So now the dupe filter is weaker.
((Obviously I'm not a mod, but I'll try to find a post from dang or sctb explaining this))
EDIT: Here's one: https://news.ycombinator.com/item?id=10223645
> We've adjusted the dupe detector to reject fewer URLs. If a story hasn't had significant attention in about the last year, reposts are ok. That's been the policy for a while, but we've brought the software closer to it. It will still reject reposts for a few hours, though, to avoid stampedes. Allowing reposts is a way of giving high-quality stories multiple chances at making the front page. Please do this tastefully and don't overdo it.
I think the dupe detector only ever identified identical URLs. (And you might notice some sites add some random junk at the end of the url to evade dupe filters on various sites).
/Not an HN mod and I do not play one on TV