MailFrontier Desktop uses a number of techniques to combat spam.
Firstly, it has a Bayesian filter. Bayesian filters calculate the probability that an email is spam based on its content and are adaptive and self-learning. For example, if your ham messages never contain the word "Viagra," then the filter will learn that emails containing that word are probably spam. On the other hand, if you work in a drug store and regularly receive ham messages that contain that word, the filter will learn that emails containing that word are probably not spam. Because Bayesian filters are self-learning, they're able to block a high percentage of spam while blocking only a minimal amount of ham.
Secondly, it has a community-based filter. When a spam email slips through MailFrontier Desktop's net and arrives in your inbox, you have the option to report it. Once an email has been reported by a number of users, MailFrontier Desktop will begin to treat that email as spam.
Thirdly, MailFrontier Desktop has a challenge/response filter that is disbaled by default. If you enable the filter, MailFrontier Desktop will notify senders that the delivery of their email has been delayed until they complete a challenge such as counting puppies.