--nextPart3883082.MlNqDE5l1y
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

On onsdagen den 25 mars 2009, Bowie Bailey wrote:
> BAYES_50 means Bayes has no opinion, the score for that should be 0.

I've set the score for BAYES_50 to 0.7 (I could probably increase that)=20
because in practice, almost all my ham is BAYES_00 or BAYES_01, so if a=20
message scores 4.3 from other rules it's almost certainly spam unless it=20
looks like previous ham. Conversely, if a message hits no other rules, a=20
point or even two from BAYES_50 won't bring it anywhere close to the=20
threshold.

But this is on a personal mail server with a well-trained bayes database, a=
nd=20
raising the score for BAYES_50 is basically equivalent to lowering the=20
threshold, which is usually not recommended.

=2D-=20
Magnus Holmgren        holmgren@lysator.liu.se
                       (No Cc of list mail needed, thanks)

  "Exim is better at being younger, whereas sendmail is better for=20
   Scrabble (50 point bonus for clearing your rack)" -- Dave Evans

--nextPart3883082.MlNqDE5l1y
Content-Type: application/pgp-signature; name=signature.asc 
Content-Description: This is a digitally signed message part.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEABECAAYFAkn95xAACgkQk7mRNn1h4+aYGQCdHyYkwrwbNxRPjBsUcMEC7Uco
SRcAoJ2bQNpxyAt85BnQ2PE/kLPiQXGF
=7t44
-----END PGP SIGNATURE-----

--nextPart3883082.MlNqDE5l1y--