[Spambayes] FILTERING

[Tony Meyer]

> Bayes database initialized with 6 spam and 3303 good messages 

This is not good.  SpamBayes works best with roughly even numbers of ham &
spam.  1::3 or even 1::5 might work ok, but you have 1::550!  To SpamBayes,
everything will look like ham, because almost everything it has ever seen is
ham.  You should retrain from scratch.  See
<http://entrian.com/sbwiki/TrainingIdeas> for more information about
training.

It's suspicious that the number of trained spam is so low, if you have been
dragging messages out of the Inbox (or whereever) into the spam folder.  If
the count (it's displayed on the main tab of the SpamBayes Manager dialog)
doesn't change when you do that, could you send us a log that includes you
doing that?  (The one you sent only included moving in the other direction).
Alternatively, you can use the "Delete as Spam" and "Recover from Spam"
buttons.

=Tony.Meyer

-- 
Please always include the list (spambayes at python.org) in your replies
(reply-all), and please don't send me personal mail about SpamBayes.
http://www.massey.ac.nz/~tameyer/writing/reply_all.html explains this.