The highest and lowest bayes rules are BAYES_99 (spam probability > 99%) and 
BAYES_00 (spam probability < 1%), but often the confidence is as high as 
0.9999 or 1.0000 (rounded). 0.999 instead of 0.99 means (in theory at least) 
that the FP chance decreases by a factor 10. Conversely at the other end.

This has been mentioned before on this list and I have added such rules 
locally. Question: Wouldn't it be wise to add them to the standard 
distribution as well?

-- 
Magnus Holmgren        [EMAIL PROTECTED]
                       (No Cc of list mail needed, thanks)

Attachment: pgpDC7U1c74Y8.pgp
Description: PGP signature

Reply via email to