Advantages: The heart of SpamBully; trained properly the Bayesian filter can be almost as effective as you are in filtering spam messages.
How it works: Bayesian filters work by looking at the words that are used in spam emails and good email and then classifying the email based on the words that these emails contain. There are certain words that tend to comprise spam emails and also certain words that tend to comprise your non spam or good emails. By analyzing these patterns SpamBully can effectively route spam messages to the Spam folder and good emails to your Inbox. It can adapt itself to your email as you correct its mistakes.
Limitations: Bayesian filters that are improperly trained can be inaccurate. It can also take up to a few weeks for Bayesian filters to become properly trained. It is recommended to closely monitor SpamBully during the first few weeks of operation to make sure it is categorizing your email properly. You may need to mark emails as needed with the Spam and Not Spam buttons in the toolbar so SpamBully better learns your email habits.
Recommendations: We recommend enabling this filter for almost all users. Properly trained Bayesian filters can become well over 99% accurate in many cases. One other recommendation we have is to NOT delete all of your old spam messages. If you ever need to retrain your Bayesian filter, your old spam messages are a gold mine of knowledge to rebuild your Bayesian filter from and can help make your filter even more accurate.