Here is an example of one: A message was scored with bayes only and it gave it 0.1. So it was auto-learned as ham. Then I did an sa-learn --spam on it. I re-ran spamassassin on it to see the value. Bayes gave it a 5.1 score (I changed it from the default value, I believe) and when AWL came in, it put it somewhere in the middle (2.x or 3.x). So I ran it through a bunch of times and eventually it averaged out to > 5. So when bayes missed it, it cause the ripple effect, whereas if bayes had known it was highly spam, it would have given a higher score and AWL wouldn't have classified it as ham.
Thomas -----Original Message----- From: Matt Kettler [mailto:[EMAIL PROTECTED] Sent: Wednesday, September 29, 2004 6:28 PM To: Thomas Wetmore; users@spamassassin.apache.org Subject: Re: AWL: How to view entries and/or remove a wrong entry? At 05:21 PM 9/29/2004, Thomas Wetmore wrote: >Hi > >AWL sometimes learns a spam incorrectly as ham. Then when another spam >comes in from the same place, it is AWL'd as ham. How can I remove an >entry in the AWL or change its value? Even if I sa-learn --spam the mail, >it still is AWL'd as ham. (I didn't see anything in the FAQ) 1) first make sure you understand how the AWL works.. Just because the AWL subtracts points, doesn't mean it thinks it's ham. http://wiki.apache.org/spamassassin/AwlWrongWay 2) if you MUST change a value, and you fully understand what was said above, you can use the following commands parameters to spamassassin (see man spamassassin for more details) -W, --add-to-whitelist Add addresses in mail to whitelist (AWL) --add-to-blacklist Add addresses in mail to blacklist (AWL) -R, --remove-from-whitelist Remove all addresses found in mail from whitelist (AWL) --add-addr-to-whitelist=addr Add addr to whitelist (AWL) --add-addr-to-blacklist=addr Add addr to blacklist (AWL) --remove-addr-from-whitelist=addr Remove addr from whitelist (AWL) WARNING: make sure you understand what the AWL really is before manually adjusting it. The AWL is NOT a white/blacklist system, it merely has a similar effect due to it's score averaging behaviors.
