> Individual tests like SPAMCHECK and NOXMAIL generate 
> a number of false positives...

Hi John,

Can you give me some example for false positives created by SpamChk?
Was they triggered from header- link- or keyword-checks?

We've set SpamChk to return "his own opinion" as a weight to declude. So
SpamChk will not say clearly "this is spam" or "this is not spam"
Attached you can see a report from today with over 1300 msgs and the
returncode from spamchk. 

Notes: 
We've a weighting-system with hold on 100. Because we add "[s%weight%]"
to any subject on hold messages we have no need to delete msgs
automaticaly. We simply sort the hold messages in Spamreview by subject
line. Then we can check the hold messages from 100 to ~200. done. 

We mean that for an external test like SpamChk it's very difficult to
say it had created false positives. If he reports more then the hold
value for a legit mail: Yes this we consider a false positive. 

Markus


Attachment: spamchk_returncode.PDF
Description: Adobe PDF document

Reply via email to