ConfigServer Community Forum

I feel like my Bayes is not auto-learning high scoring spam. When I look at my bayes DB info, it shows the # os Spam messages as 685, which has been pretty much me manually training it. In Mailwatch, I see a lot of high-scoring spam go by, but that does not seem to increment the # of spam.

Any ideas on how to troubleshoot this?

Number of Spam Messages: 685
Number of Ham Messages: 10,664
Number of Tokens: 192,097
Oldest Token: Thu, 26 Mar 2015 10:41:10 -0500
Newest Token: Fri, 27 Mar 2015 11:17:11 -0500
Last Journal Sync: Fri, 27 Mar 2015 11:14:43 -0500
Last Expiry: Fri, 27 Mar 2015 04:06:41 -0500
Last Expiry Reduction Count: 65,806 token

I am also interest in this matter (on front end each mail show : Auto learned=no, after mannually clicked on "learn as spam".