[Spambayes] Can spammers easily sneak through the SpamBayes filter ?

Jesse Pelton jsp at PKC.com
Wed Dec 19 20:12:29 CET 2007

I don't know the mathematics behind it, but I can say that I've had very
good results with SpamBayes for years, during which time spammers'
techniques have evolved considerably.  One of the strengths of a
Bayesian filter is that it gets trained on the messages you actually
receive, so it tends to adapt to such changes.  (If I don't think it's
well-trained for the messages I'm currently receiving, I tend to delete
my training database and start fresh, since training is easy and quick.)
SpamBayes looks at message headers as well as the message body, and this
may help it to correctly classify messages that use tricks in the
message bodies.

Bottom line: in my experience, spammers cannot easily get by SpamBayes.
Of course, there's no guarantee that this is everyone's experience, or
that it will remain true for me forever.  One of the odd things about
SpamBayes is that it works much better than it seems it should, though,
and as long as that's the case, I plan to keep using it.

-----Original Message-----
From: spambayes-bounces at python.org [mailto:spambayes-bounces at python.org]
On Behalf Of gpr
Sent: Wednesday, December 19, 2007 1:18 PM
To: spambayes at python.org
Subject: [Spambayes] Can spammers easily sneak through the SpamBayes
filter ?

Can spammers easily sneak through the SpamBayes statistical filters
easily -
for example how about a spam message with good words camouflaged (in
something like white color) and only
spam content visible to the user. When the filter comes across the
the spam score 
would lean towards lower value and would sneak through the filter.

How does SpamBayes address such spam techniques?

Thank you,
View this message in context:
Sent from the Spambayes - General mailing list archive at Nabble.com.

SpamBayes at python.org
Info/Unsubscribe: http://mail.python.org/mailman/listinfo/spambayes
Check the FAQ before asking: http://spambayes.sf.net/faq.html

More information about the SpamBayes mailing list