For a little "more than a month now":http://revjim.net/item/9933/ I've been using "CRM114″:http://crm114.sourceforge.net/. I figured it was about time for a status report.
I don't track how many incomming mails I receive. It's too difficult since they come from so many places and I often mail myself which inflates the numbers considerably. Most spam catching statistics are supplied as the inverse of the number of spams that made it through versus the number of mails received in total. Mine, however, are calculated based on the number of spams received and the number of spams caught. So, please keep that in mind as you review these statistics.
Additionally, remember that the spam catching ability of any learning mail scanner gets better as you use it. I only train on errors and I train on every error, if that helps you understand my numbers any better.
From Febuary 24th until now, I've received 6,231 spams. 5,972 of them were caught by CRM114. That's a 95.84% catch rate. Additionally, 100 messages were marked as spam that weren't (false positive).
In the last 7 days, however, I've received 1,456 spams. 1,427 of them were caught by CRM114. That's a 98.01% catch rate. Additionally, 4 messages were marked as spam that weren't (false positives).
I should also note that, in the last 7 days, all 4 of the false positives were computer generated messages with heavy use of HTML (which looks a lot like SPAM). As I get more non-spam messages that look like these, it'll get smarter about it.
All in all I'm very happy with the results. A lot better than when I was using SpamAssassin.
