What are the Chances of Spammers Getting Past Bayesian Spam Filters?

Bayesian spam filters work by analyzing the contents of the email, and then calculating the probability of the message being spam. They analyze the characteristics of both, the legitimate mails as well as the spam mails. For the Bayesian spam filtering process, both, the characteristics of the legitimate mails as well as the spam mails are equally important. As it analyzes the mails, it keeps on building up a list of genuine as well as spam characteristics.

Once you have analyzed and classified the message as genuine or spam, it can be used to train the spam filters. Spammers would have a real bad time if spam filters can be specifically trained for individual users. They would have a hard time hoodwinking most people's spam filters, and the spam filters would have the ability to adapt to whatever the spammers would try.

What Can The Spammers Try

The only reason why an email is a spam is that it is significantly different from the regular and legitimate email. The spammers are likely to succeed and get past the Bayesian spam filters if they make their spam emails look identical to the ordinary mails everyone gets. They probably have the wherewithal to do it today. But spammers are not there to send ordinary looking emails, and the ordinary looking emails are the only way to get past the anti spam filters, such as the Bayesian spam filters.

Ordinary looking emails do not work for spammers as their aim is to make money. They send emails to hundreds of thousands of people and at a miniscule cost to them - and if only a fraction of the recipients respond, they have got it made. If they do decide to switch to ordinary looking mails to get past the Bayesian spam filters, we would be receiving much more spam in our inboxes every day, just as we used to before the Bayesian spam filters came into play. But it will not work for them - commercially.

There is one thing they can try, though, to get past the Bayesian spam filters. The basis of the Bayesian spam filters is that if one word appears very frequently in a legitimate mail, it is so significant that it can make any message appear to be genuine, even if it happens to be spam.

The spammers now need to find a way to find out those good email words to hit the bulls-eye. They can do this by using HTML return receipts to see the messages you opened, and can include one of them in the spam message to get it past the Bayesian spam filter. Such things are time consuming and HTML return receipts do not always work.