[EMAIL PROTECTED] wrote: > On Tue, 09 Nov 2004 15:23:08 -0500, "Kris Deugau" <[EMAIL PROTECTED]> > said: > > Snag mine from http://www.deepnet.cx/~kdeugau/spamtools/ > > Nice meta rules (BAYES_vs_SURBL) -- I like those a lot!!!
Yep. I added them after a few customers reported FPs on legit mailings with properly low Bayes scores, but multiple SURBL hits bumped them *just* over the threshold. Since stabilizing SURBL usage, I've had *far* fewer FNs reported, and I've rarely had FPs reported to begin with- there's one every other month or so. I don't have real accuracy stats overall, but my own account shows ~200-250 spams per day, and I see maybe two to three a week showing up in my inbox. I filter listmail with procmail before it hits SA, but whitelisting has proven equally effective in the past. Overall SA stats for that system have been pretty constant at around 9:1 spam:ham, around 13K messages/day, for close to a year now. Ugh. :/ -kgd -- Get your mouse off of there! You don't know where that email has been!
