Spam Report Scores

[Chris Trudeau]

Got a question, I'm using MaiLScanner and MailWatch with SA 2.63.

Have the system set to store spam and non-spam.  this works nicely.

I have a seasoned Bayes DB and the accuracy is really pretty
good....however....

I noticed a smattering of messages that got through, no problem, till I dug
a little deeper.

if I run a manual spamassassin on a message, it produces a score that is
different than what was recorded for the message.  Wouldn't be a big
problem...but the score is WAY OFF...

here is an example.

message that recorded a score of  2.58:

<SNIP>
not spam, SpamAssassin (score=3.045, required 6, BAYES_90 2.10,
HTML_IMAGE_ONLY_10 0.02, HTML_MESSAGE 0.10, HTML_TITLE_EMPTY 0.12,
RCVD_IN_DSBL 0.71)
</SNIP>

in my mailwatch database returns:

<SNIP>
X-Spam-Status: Yes, hits=7.8 required=5.0 tests=BAYES_99,HTML_IMAGE_ONLY_10,
        HTML_MESSAGE,HTML_TITLE_EMPTY,RCVD_IN_BL_SPAMCOP_NET,RCVD_IN_DSBL
        autolearn=no version=2.63
</SNIP>

I ran spamassassin with the -p pointing to my
/etc/MailScanner/spam.assassin.prefs file too....


Any ideas would be helpful!

CT

-------------------------- MailScanner list ----------------------
To leave, send    leave mailscanner    to jiscmail at jiscmail.ac.uk
Before posting, please see the Most Asked Questions at
http://www.mailscanner.biz/maq/     and the archives at
http://www.jiscmail.ac.uk/lists/mailscanner.html