We've been receiving a few emails from customers about the spam they've been receiving recently.
We currently already filter all the email for spam. According to our reports, 9 out of 10 messages we receive are spam and are removed with a 0% false-positive rate. Unfortunately we cannot "machine-filter" more than this without potentially making mistakes.
You can attempt to further filter them yourselves by using the scoring mechanism our anti-spam mechanism uses by reading and using these instructions.
However, whether it would meet with much success these days is a different matter, as the few spam emails that do make it through are doing so by using some clever techniques to make it look (to a machine) similar to legitimate email (at the moment they use normal text unrelated to the spam and an image that contains themessage they are trying to send, usually with imperfections to stop any form of OCR being performed on the image) as well as looking like and acting like a proper email server from the email servers perspective to prevent techniques like greylisting.
Unfortunately the specifications for the email protocol were created in 1982 (even though email was used in various forms before this) and has changed very little since then. Email was originally conceived as a method of sending messages between academics, not as the business medium it has turned into today.
There is a war going on between spam and the anti-spam programs, and just as in nature there is an element of survival of the fittest. A lot of spam will die, but some will always make it due to the sheer volume and the ingenuity of the people that generate it.
So I'm afraid some spam emails will always get through, and you can see this at places like Hotmail and Google mail to see that even with the huge resources that these businesses have to draw on, they cannot solve this problem, and I'm afraid we have no "magic bullet" either.