Sorry about the generic subject, but it is the only thing this newbie knows to describe the symptom.
Platform: Debian (Etch?) Latest Spamassassin in apt (version 3.1.7-deb) Invocation comes from KMail, via spamc (presumably) to the spamd daemon- set up using KMail Wizard, and manually checked Spamassassin doesn't seem to be catching much spam at all. I've run thousands of spams through sa-learn, and hundreds of hams (needless to say, I get the ratio of thousands of spams to tens of hams). I can't see where it's even using the Bayes filter. Here is the config file: ############################################################################ # This is the right place to customize your installation of SpamAssassin. # # See 'perldoc Mail::SpamAssassin::Conf' for details of what can be # tweaked. # # Only a small subset of options are listed below # ########################################################################### rewrite_header Subject *****SPAM***** required_score 5.0 use_bayes 1 bayes_auto_learn 1 ############################################################################# And here's the german portion of an example header from a processed email that should have been spam but wasn't: ############################################################################ X-Spam-Checker-Version: SpamAssassin 3.1.7-deb (2006-10-05) on anw-dev.cfl.rr.com X-Spam-Level: X-Spam-Status: No, score=0.9 required=5.0 tests=SUBJ_HAS_UNIQ_ID autolearn=no version=3.1.7-deb X-Virus-Flag: no Return-path: <[EMAIL PROTECTED]> <snip> ########################################################################### It looks like, to this unwashed newbie, that: a) it's not autolearning (perhaps it doesn't on real emails?) and b) even though this email is full of references like "boosting your sexual power" and "high quality medications", and even comes from address "[EMAIL PROTECTED]", you can see it is still getting a low spam score. TIA -- Thanks and regards, Allen Williams Office: +1.321.309.7931 Mobile: +1.321.258.1272