SpamAssassin Without Bayes

["Gustafson, Tim" <tj...@meitech.com>]

Hi Everyone!

I know that Bayes is "the" defacto best way to fight SPAM right now, but
I wonder if anyone out there is running SA without Bayes turned on and
what their experience with it is?

I run SA with Bayes enabled right now site-wide for more than 500 e-mail
accounts, and my problem is that some of my clients who deal in
industries that are very SPAMmy in nature (like a recruiting firm or an
insurance agency) are getting e-mails tagged as SPAM that aren't SPAM
(at least to my client) at all.  Most of these e-mails are being tagged
because of their Bayes score, and I'm curious how many false negatives
I'll get if I just turn of Bayes altogether.

Is there any log analysis tool that will tell me how many SPAMs would
NOT have been marked as SPAM if a given rule was turned off when the
e-mail was received?

Can anyone comment on the effectiveness of SA without Bayes activated?

Tim Gustafson
MEI Technology Consulting, Inc
tjg@meitech.com
(516) 379-0001 Office
(516) 480-1870 Mobile/Emergencies
(516) 908-4185 Fax
http://www.meitech.com/ 

Re: SpamAssassin Without Bayes

["Eric A. Hall" <eh...@ehsco.com>]

On 4/4/2005 12:28 PM, Gustafson, Tim wrote:

> I know that Bayes is "the" defacto best way to fight SPAM right now, but
> I wonder if anyone out there is running SA without Bayes turned on and
> what their experience with it is?

I have it turned off and don't miss it. Tweaking your rules works just as
well, and you don't have to maintain a bunch of user-specific databases.

-- 
Eric A. Hall                                        http://www.ehsco.com/
Internet Core Protocols          http://www.oreilly.com/catalog/coreprot/

Re: SpamAssassin Without Bayes

[Thomas Cameron <th...@camerontech.com>]

On Mon, 2005-04-04 at 12:28 -0400, Gustafson, Tim wrote:
> Hi Everyone!
> 
> I know that Bayes is "the" defacto best way to fight SPAM right now, but
> I wonder if anyone out there is running SA without Bayes turned on and
> what their experience with it is?
> 
> I run SA with Bayes enabled right now site-wide for more than 500 e-mail
> accounts, and my problem is that some of my clients who deal in
> industries that are very SPAMmy in nature (like a recruiting firm or an
> insurance agency) are getting e-mails tagged as SPAM that aren't SPAM
> (at least to my client) at all.  Most of these e-mails are being tagged
> because of their Bayes score, and I'm curious how many false negatives
> I'll get if I just turn of Bayes altogether.
> 
> Is there any log analysis tool that will tell me how many SPAMs would
> NOT have been marked as SPAM if a given rule was turned off when the
> e-mail was received?
> 
> Can anyone comment on the effectiveness of SA without Bayes activated?
> 
> Tim Gustafson
> MEI Technology Consulting, Inc
> tjg@meitech.com
> (516) 379-0001 Office
> (516) 480-1870 Mobile/Emergencies
> (516) 908-4185 Fax
> http://www.meitech.com/ 

I use SA with all the network tests on and spamass-milter to relay mail
from our external MX to the internal Exchange mail server.  We don't use
Bayes at all as there is no good way to train it.  We still have a
phenomenal success rate - very close to 100% accuracy.

I love Bayes for small environments, but even in big ones where it might
not be practical SA still kicks butt.

Thomas

Re: SpamAssassin Without Bayes

[Kevin Peuhkurinen <ke...@meridiancu.ca>]

Gustafson, Tim wrote:

>Hi Everyone!
>
>I know that Bayes is "the" defacto best way to fight SPAM right now, but
>I wonder if anyone out there is running SA without Bayes turned on and
>what their experience with it is?
>
>I run SA with Bayes enabled right now site-wide for more than 500 e-mail
>accounts, and my problem is that some of my clients who deal in
>industries that are very SPAMmy in nature (like a recruiting firm or an
>insurance agency) are getting e-mails tagged as SPAM that aren't SPAM
>(at least to my client) at all.  Most of these e-mails are being tagged
>because of their Bayes score, and I'm curious how many false negatives
>I'll get if I just turn of Bayes altogether.
>
>Is there any log analysis tool that will tell me how many SPAMs would
>NOT have been marked as SPAM if a given rule was turned off when the
>e-mail was received?
>
>Can anyone comment on the effectiveness of SA without Bayes activated?
>
>Tim Gustafson
>MEI Technology Consulting, Inc
>tjg@meitech.com
>(516) 379-0001 Office
>(516) 480-1870 Mobile/Emergencies
>(516) 908-4185 Fax
>http://www.meitech.com/ 
>
>  
>
I can't comment on the effectiveness of SA without Bayes, but I can
suggest a couple of things.   First, try just lowering the scores given
to Bayes a bit so that emails won't be marked as spam solely because of
Bayes.    Secondly, try using per-user Bayes instead of global.   It is
a little more work, but if you are serving email for a varied group of
email accounts, it will be much more effective.