What data is there on false positives vs junk score and number of messages processed? That is, what fraction of false positives are <95, <90, <85, … after 10,000, …, 50,000 messages (assuming false positives and false negatives have been properly identified along the way)?
SpamSieve does not calculate this statistic. Speaking from my own experience, I don’t recall ever having a false positive with a score above 80.
Thanks for your response. A follow-up question:
On page 57, the manual says “Message with scores above 90 are almost certainly spam.” Was that statement based on (unvalidated?) reports from the field or was that simply a more conservative assessment based on your own experience?
It’s based on my experience looking at lots of log files from SpamSieve users.