On Jan 24, 2005, at 9:37 AM, Damir Cosic wrote:
Hello Everybody,
I started using spamassassin a while ago with version 2.X.X and it was
working great. It quickly decreased level of spam in my inbox to 1-3% of
total spam that was being sent to my address. I use it on my Mandrake 10
workstation by piping all messages from Evolution through spamassassin
-e, and throwing everything that doesn't return 0 away. I also have Spam
and Ham folders for further educating spamassassin and I periodically
execute sa-learn --spam --mbox
~/evolution/local/SA-learn/subfolders/Spam/mbox. I used default
configuration and there was never any need to tweak it.
Soon after version 3.0.0 became available, I did an upgrade. I don't
remember all the steps I executed, but I know I was following
instructions and it left me under an impression that everything went
smoothly. Next day, I realized it didn't -- my inbox was full of spam
again. Since the upgrade included conversion of Bayesian rules database,
I couldn't explain that by the need of the spamassassin to be retrained,
and accordingly, things never got any better. Today I found 103 spam
emails in my inbox and 154 recognized by spamassassin as spam -- not the
most satisfying ratio. I would appreciate any help or advice on how to
debug this problem. Apologies if this was already asked. I was not able
to find it.
Damir Cosic
I found the same drastic and disappointing lack of efficacy of 3.0 from 2.63. My solution (thus far yielding much better results) has been to boost the BAYES, EXCUSE, and URI scores.
--Paul
--------------------------------------------------------------------------------------------------------------------- This email may contain material that is confidential and privileged and is for the sole use of the intended recipient. Any review, reliance or distribution by others or forwarding without express permission is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.
