on Thu, Apr 25, 2002, Craig Dickson (crdic@pacbell.net) wrote:
> begin Peter Ross quotation:
>
> > Why don't you look into http://www.spambouncer.org/
>
> Why? Spamassassin, in my experience, is vastly more accurate and
> effective than SpamBouncer. I used SpamBouncer for several months up
> until March 2002. I could never get it to block even almost all spam
> without also having a lot of false positives, even after extensive
> tweaking of variables and even some customization of SpamBouncer's
> procmail recipes. (Some of SpamBouncer's tests are utterly mad --
> block all mail from Telstra? That's most of Australia!) I switched to
> Spamassassin plus Razor and found that even without customizing my
> configuration at all, it did a much, much better job. It's quite rare
> now for me to see spam other than in the "junk" folder to which I
> redirect such things, and false-positives are even more rare.
I've been keeping tabs on SA's specificity and sensitivity. On ~40k
mails, ~3300 spams, adjusting for some exceptional cases (one
mailbombing of 300+ items handled with a separate rule), since Feb 1:
True positive: 95.5%
False negative: 4.5%
True negative: 99.82%
False positive: 0.18%
Damned good tool.
Peace.
--
Karsten M. Self <kmself@ix.netcom.com> http://kmself.home.netcom.com/
What Part of "Gestalt" don't you understand?
A guide to GNU/Linux backups:
http://kmself.home.netcom.com/Linux/FAQs/backups.html
Attachment:
pgpilE6STW182.pgp
Description: PGP signature