AFAIK, it's using specific IP addresses to prevent users. Blocking an IP range, especially in a residential sense, wouldn't be helpful imho, too much room for error, maybe blocking IP ranges of datacentres.
It doesn't block anything. It adds evil "points" to the IP and a smaller amount of evil "points" to every IP in the same range. The more evil your IP has when creating an account the more BTC you have to pay to sign up. Known proxies and Tor already have the max amount of evil too.
It is a very smart approach! I would like so much to use it for one of my web sites...
Any chance to know more details about it?
Is there a public list of datacenter IPs vs residential IPs?