Scott,
HS = Test says ham, final result was spam. This is an inaccurate ham result.
'False negative'
How are you auto determining that an email that was ham was really spam? Are
you keying in this info into your stats based on your viewing of the email or
by user complaint? Obviously, if
If the test fails, but the message does not hit the hold or delete weight.
Not a perfect measurement, as it does not capture all ham (ham that hits the
hold or delete weight), and misses some spam (spam that does not hit the
hold or delete weight), but it is the most accurate and least subjective