Is all spam coming through as "Unsure"?
I thought I read somewhere that, during the beta phase, all spam was being posted as Unsure. Is this still the case?
I get occasional (two a day, a few days a week) spam generated by a human being in China. Their post is, to me, obviously spam. But it always seems to come back as Unsure, which means that they get a CAPTCHA which they can decrypt.
Increasing use of human beings by spam networks was one reason why I wanted to try Mollom over straight CAPTCHA. Humans can crack CAPTCHAs, so I needed a spam solution that means spam posted by a human gets rejected without them even being invited to pass a CAPTCHA test.
If all spam comments are coming through as unsure, when will this extra caution be switched off?
If, OTOH, Mollom really is unsure about these messages, should the passage of time mean that it learns to distinguish them with greater confidence?
Own question answered
Today I finally had a spam comment flagged directly as spam.
More spam than unsure
For most sites, Mollom returns more 'spam' than 'unsure'. It might be the case that Mollom is still learning your site's behavior. See also http://mollom.com/humans-entering-captchas.