I am not a programmer, so I could not think of any "rule" to catch this one. I only hope that someone here can. I have gotten quite a few of these (at least two per day), and it's always from my forged email address from someone named wilfred or arthur.
Thanks.
<snip>
X-Spam-Status: No, hits=-1.2 required=7.3 tests=BAYES_00,HTML_FONTCOLOR_RED,HTML_FONT_BIG,HTML_FONT_INVISIBLE,HTML_MESSAGE,HTTP_ESCAPED_HOST,HTTP_EXCESSIVE_ESCAPES,MIME_HTML_ONLY autolearn=no version=2.60
Ok first thing that jumps out at me, you don't need better rules.. your bayes training is screwed up. BAYES_00 should only hit messages which are very clearly NOT spam based on the training.. Have you been relying entirely on autolearning or something?
The message here isn't full of any kind of bayes poison, and as such it should be very easy for it to be properly classified. Bayes should very easily classify this email at at least BAYES_80, if not BAYES_99... which would be enough to cause a very severe swing in the score. At bare miniumum it should drag it up by 6.5 extra points, if not 10.3.
Either disable bayes, or hand train it.. don't let it just sit there and autolearn (which unfortunately is the default) without hand training it some too.
------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
