when is Bayes scoring used?

[Mark Nienberg]

If you are counting on SpamAssassin to "auto-learn", rather than manually
feeding messages through it, then it only uses messages with very high or
low scores.  The defaults are

auto_learn_threshold_nonspam  -2.0
auto_learn_threshold_spam  15.0

I imagine that 200 spams with scores greater than 15 will accumulate fairly
quickly, but I don't see many messages with scores less than negative
two, so how will it ever auto-learn any ham?  Even messages internal to
my network score zero.  I'm thinking about changing the default to positive
one or thereabouts.
--
Mark W. Nienberg, SE
Tipping Mar + associates
1906 Shattuck Ave, Berkeley, CA  94704
visit our website at http://www.tippingmar.com