full-disclosure - spam with anti-bayesian parts

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1073923116.3653.9.camel@tantor.nuclearelephant.com>
From: jonathan at nuclearelephant.com (Jonathan A. Zdziarski)
Subject: spam with anti-bayesian parts

> What I'm wondering is:
> Why do the spammers even go to the length of using random words?

It still works on poorly written bayesian filters and some heuristic
based filters.  But spammers are stupid: they don't realize that they
are also providing a much easier way to identify them for everyone using
the more modern filters.

> . Why don't they grab some real text, say from a news site? There's an endless
> supply of new, proper text out there.

Some do, but it still ends up having the same effect.  The unknown
tokens are not paid particular attention to until the filter learns that
these now words are used mostly in spams.

Some filters are now paying attention only to the HTML portion of a
message, if it exists, as many spammers are putting their Bayesian noise
in the text segment.  This reportedly helps accuracy.