Up until now I had it configured to only check email address matches, which is the most unique identifier. IP address and username matches are much more likely to be false positive so I disabled those checks. But now that the threshold is configurable I might set a very high threshold of like 30 matches for IP addresses and maybe usernames.
Good point to keep in mind, it took me a long time to figure that out. Although honestly those “toxic domains” are aptly named, and any user registering with one of those emails almost surely has bad intentions of spamming or evading a ban for repeat trolling / harassing. Fortunately this plugin just silences the user, so if they were somehow a false positive they can still contact staff with this other (IMHO essential) plugin enabled: