Bayesian spam filtering is used in most email server and applications to establish rules for determining what and how messages get delivered to and end users INBOX or SPAM or Junk Folder.

Mathematical calculations are applied on multiple elements of incoming email to prevent “False Positives” (marking mail as spam or junk that is actually good mail).

One of the elements is the machine or server sending the email – or the provider hosting that server. Each machine on the internet is identified by an IP address. If the sender IP address or range has been either abused by senders – or machines on that range of numbers have suffered a virus infection – the IP addresses can end up with a poor reputation or on blacklists of recipient servers.

Depending upon the severity of the bad reputation the receiving server might simply add a couple of pointsto the potential spam score rather than block the message all together.

As a general rule, email security tightens as the organizational structure increases so individual users such as friends or relatives with a private email address will probably accept most forms of email message with little more than anti-virus filtering.

Small businesses may have more advanced anti-virus and some anti-spam and possibly other content filtering.

Large business are more likely to have a multi-tiered filtering system installed that may very well block any message containing embedded images, multiple links, scripts or other non-text elements.

Mail applications that rely on rules in software updates -like Outlook or other mail reader applications adjust security filters according to threat trends.

Below is a typical spam report:

2 points for Comcast Cable customers – was due to a series of virus infections spewing spam and virus – 2 points is not enough to land good email in the bit bucket.

2 points for a link to sharefile – means there has been a reported blast of abusive use of sharefile as a distribution service for unwanted email or potential malware.

Still – 4 points is not enough to land in the bit bucket
However this adds up and links that pile up in your signature line can add to an unwanted score.


X-Spam-Report:
* 2.0 RCVD_IN_COV_SPAMMERS RBL: Spammer blocked by CompanyV RBL
* [Comcast cable customers]
* 2.0 URIBL_COV Message links to a URL listed in the CompanyV blocklist
* [URIs: sharefile.com]
* 0.0 INVESTMENT_ADVICE BODY: Message mentions investment advice
* 1.0 EXTRA_MPART_TYPE Header has extraneous Content-type:...type= entry
* 0.5 HTML_MESSAGE BODY: HTML included in message
* 0.5 BAYES_50 BODY: Bayesian spam probability is 40 to 60%
* [score: 0.4943]
X-Spam-COV-Tests: BAYES_50=0.5,EXTRA_MPART_TYPE=1,HTML_MESSAGE=0.5,
INVESTMENT_ADVICE=0.042,RCVD_IN_COV_SPAMMERS=2,URIBL_COV=2