SpamAssassin 3

[Mark Nienberg]

On 17 Nov 2004 at 16:26, Kyle Harris wrote:
> For those that have made the upgrade, are you generally seeing more spam
> being caught, less being caught, or about the same?  Opinions?

Initially, I was catching fewer spam messages with 3.01 than I was with 2.64.  But
that is probably because I had tweaked the old version as time went by until it was
very specific to our mail flow. Now that I have been gradually revising scores and
adding a few custom rules, it is getting better.  Also, when I upgraded, the conversion
of my Bayes database failed (yes, I followed the directions carefully).  I was  forced to
abandon it and use the starter database from Fortress.  So it probably hasn't settled
down completely yet.

In particular, my opinion is that the default Bayes scores and SPF scores don't make
much sense and are overall too low.  I realize that they are derived from a
complicated statistical analysis aimed at minimizing false positives etc.  Maybe
statisticians can appreciate that, but as an engineer, I can't abide scores where
BAYES_99 is less than BAYES_95, or SPF_FAIL is less than one point.

I'm pretty confident that with a little more tweaking I'm going to get back to (and
probably improve on) my old detection rate.
--
Mark W. Nienberg, SE
Tipping Mar + associates
1906 Shattuck Ave
Berkeley, CA 94704
510 549-1906 ext 236
http://www.tippingmar.com

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the MAQ (http://www.mailscanner.biz/maq/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).

Support MailScanner development - buy the book off the website!