My mail server handles mail for a number of addresses and domains, so it
gets some spam. Only about 1% of it gets through though. I use spamhaus
and spamcop's distributed blacklists in conjunction with a hand-maintained
IP-address-based block list. Once something goes into the block list, it's
unlikely to ever come out. This turns out to be pretty effective because
a lot of spammers appear to return to the same permissive ISPs to host their
spam runs. If I block an IP range from what appears to be a highly permissive
spammer ISP, there's a good chance that one of the class C IP blocks belonging
to that ISP will block a spam run a month or two later.
Here are some blocklist entries, ordered by the month that I added them, along with how many (usually class C ranges, but occasional class B) entries are in each month's file. These are in the form of a sendmail access file. Feel free to use them however you'd like.