(Sorry about the extra 0). That explains better, I had thought from
what you had said that it was a pre-filtering so you never looked at the
filtered emails at all. I use spamassassin with heuristic rules and
Bayesian "learning" based on the spam/non-spam emails I've already
received (but I have it's black hole lists disabled) and for the last
month it's been 99.7% correct (1 false positive and 3 false negative),
but I still feel I have to scan every subject line just in case (the
filtered spam being directed to its own mailbox). And I still do
everything I can, such as munging, to keep it down since any spam at all
is an annoyance so the fewer I get the better, as much for my mental
peace of mind as for the actual work involved. I'm at 40 spams/day
which is tolerable, considering that I have 3 different mailboxes plus
the "catch all" mailbox from my domain name.