SpamSieve identifies spam and colors it according to how “spammy” it is. Is there any benefit to going in and manually identifying emails as spam that are not blue (the most spammy) to improve the algorithm?
I generally recommend only correcting the mistakes.
Thanks. It rarely (never?) identifies a good email as spam. It still misses some spam that I manually identify as spam
Does the Log window say that these were mistakes?
I went back through a week of the log and found 4 that were not identified as spam that I manually identified as spam.
The log says
2 with the note: SpamSieve had predicted this message to be Good, and you corrected the mistake. SpamSieve’s Bayesian classifier predicted this message to be Good based on a statistical analysis of its content.
and
2-with the note SpamSieve had predicted this message to be Good, and you corrected the mistake. SpamSieve classified this message as Good because it matched an allowlist rule.
Thas is out of about 1145 entries in the spam log. Does not miss many actual spam. There were no good messages identified as spam
It’s good that the log seems to agree with you about which messages were SpamSieve mistakes. If it didn’t that could indicate a setup problem or a damaged file. With more than 99% of the spams being caught automatically, you seem to be in good shape, but if there are problems with more spams or with certain types of spams, please send in a diagnostic report and we can investigate.
Thank you, I find SpamSieve to be quite reliable. I have been using it for a long time. It looks like I started with ver. 2.6.4 in 2007 
1 Like