Our Email filtering system has many components. Probably the most interesting is the spam recognition system. For this we use the highly regarded SpamAssassin product. SpamAssassin uses a sophisticated set of rules to identify and score spam. Each rule compares a specific criteria to the Email message and assigns either a positive or negative score for each comparison. Some blatant spam components such as ADV: in the subject will assign a large value to the score which may alone be enough to determine that the message is, in fact, spam. Other, less obvious yet common spam components, may be scanned for and assigned lesser values since these elements could be found in normal Email as well. In the end SpamAssassin will assign a total value to the message and that value will be compared to your preferences to determine the disposition of the message. To see how SpamAssassin has rated any particular message you need to view the "headers" of the message. Most Email programs will let you view the headers or at least view the source of an Email message. In the section where the Subject and Email routing information is found you will see headers labeled "X-MailScanner-SpamCheck:" If the message was scored as spam you will see a header entry similar to this:
X-MailScanner-SpamCheck: SpamAssassin (score=13.5, required 5, INVALID_DATE_TZ_ABSURD, NO_REAL_NAME, PLING, TO_MALFORMED, US_DOLLARS_3, BIG_FONT, CTYPE_JUST_HTML, FORGED_YAHOO_RCVD)
If the message was not seen as spam it might look something like:
X-MailScanner-SpamCheck: not spam, SpamAssassin (score=0.9, required 5, FROM_AND_TO_SAME)
Each of the tests that matched your message will be iterated in this header entry. Note that some of these matches may yield negative scoring - meaning that they are tests that indicate the message may NOT be spam. To learn more about the tests performed by SpamAssassin, see the Tests Performed page on the SpamAssassin web site.