Custom Scores

[Peter Bonivart]

Sorry for answering two persons (Ugo and Pete) in one mail...

>>Just installed DCC on one of my servers today and is working nicely -
>>made me think that, if some messages are listed with checks like DCC or
>>certain RBLs, then they must be alsmot %100 spam, or
>>undesirable emails?

RBL:s sometimes list legit servers for a while for several reasons and
DCC doesn't even try to decide if a message is ham or spam, it just
assumes that if really many of the same message circulate it's spam.
That sounds crazy but it works really well. But you can't depend on any
single one source, that's why SA adds them up.

>>Has anyone heard of DCC or the best RBLs listing legit senders or
>>emails? is it worth giving these a much higher score so these message
>>score as High Spam and are deleted on the spot?

Read above comment. Don't bump the score excessively.

>>OR am i am missing the central reaosns why this likes DCC only
>>score 1.81 ?

It scores 1.81 because you don't use Bayes, if you did you would get
2.91. Bayes helps a lot, a BAYES99 adds 5.4 points.

> If that can help you, I got many DCC_CHECK score with 1.81, but also one with 2.91, like the one below:

Read above comment. Look in /usr/share/spamassassin/50_scores.cf (or
/usr/local/share/spamassassin/50_scores.cf), the last column is used
when you have net tests and Bayes enabled.

--
/Peter Bonivart

--Unix lovers do it in the Sun

Sun Fire V210, Solaris 9, Sendmail 8.12.10, MailScanner 4.25-14,
SpamAssassin 2.63 + DCC 1.2.30, ClamAV 0.67 + GMP 4.1.2