Peter McGarvey Sent: Friday, December 12, 2003 8:46 AM

> I've seen several instanaces where the AWL mechanism has cause non-spam
> email to be classed as spam.  Here is example report from one of these:
>
> -4.9 BAYES_00               BODY: Bayesian spam probability is 0 to 1%
>   56 AWL                    AWL: Auto-whitelist adjustment
>
> My understansing of autowhitelisting is that it flattens score spikes
> towards a long term average based on the sender.  Which is obviously not

AWL is working exactly the way it is designed. The idea here is that this
person has in the past sent you email that scored high and now sent you an
email that scored low. AWL averages that out... it doesn't matter whether
it's making the score higher or lower. This email could just as easily be an
email from someone who has sent you spam before and is now sending you a
message that appears to be legit to Bayes. IMHO, if someone has sent me spam
before then we should weight their other email accordingly.

I have found that AWL works quite well and I keep it enabled. The only time
that it is a pain is if I send myself or someone else a test using GTUBE.
That f's up the AWL until I send a few hams. There's a way to remove the
sender from the AWL but I could never be bothered.

As to how to turn it off, check the docs.

cheers,
Colin

Colin A. Bartlett
Kinetic Web Solutions
www.kineticweb.biz




-------------------------------------------------------
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to