[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20031008014534.GE6194@Bragi.private.network>
From: mds at helices.org (Michael D Schleif)
Subject: Spam with PGP
"Jonathan A. Zdziarski" <jonathan@...learelephant.com> [2003:10:07:14:18:15-0400] scribed:
> > My "obsolete" spamassassin catches 90% of the spam that gets through the
> > DNSBLs into my inbox.
>
> 90% is extremely poor considering most probability-based filters run
> from 96% - 99.6%. For the past couple months, I've been at literally
> 100%, and I get 50-60 spams per day. No need for DNSBL.
Just exactly *HOW* do you measure ``literally 100%''?
How do you account for false positives and false negatives?
Do you manually examine every message, whether or not flagged as spam?
Personally, I receive between 2,000 and 3,500 messages each day, of
which ~200 are flagged as spam via procmail rules and spamassassin. I
receive email from a very heterogeneous mix of sources, including
mailing lists, companies whose products are interesting, personal
contacts, technical resources, &c. For me, one or two messages each day
improperly flagged is the cost of doing business on the Internet.
As impressive as that phrase ``literally 100%'' is, I am considerably
skeptical. I have never -- in thirty years -- found any software
utility that can beat five-nines statistics.
How can we get hold of yours?
--
Best Regards,
mds
-
Dare to fix things before they break . . .
-
Our capacity for understanding is inversely proportional to how much
we think we know. The more I know, the more I know I don't know . . .
--
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.grok.org.uk/pipermail/full-disclosure/attachments/20031007/6511fa47/attachment.bin
Powered by blists - more mailing lists