At Mon Feb 2 17:29:29 2004, Sandy S wrote: > > just like everybody does I receive lots of pseudo-random spam mails that > are > > trying to pollute the spamfilters: > > > > --------- > > agreeable anglican hydrochloric distaff freeboot conservative awesome > estes > > hollow mediocrity pray railbird inoffensive drum highland ballard bauble > > arccosine bock heartfelt burn conveyor > > --------- > > > > What can I do against this mails with my SpamAssassin? Are there new > rules?
> I think the general consensus is to go ahead and let Bayes learn them. > After all, how many legitimate email messages do you see with words like > distaff, estes, railbird and arccosine in them? I've found that these > bayes poison words are actually as good a sign of spam as the V**** word! It's not so much "letting Bayes learn them". In many cases, you have to teach these messages to Bayes. Spammers are increasingly tailoring their spam to avoid SA. One of the effects of this is that messages packed full of Bayes poison are not hitting any rules at all, and hence their score of 0 means they get autolearned as *ham*. If you're using Bayes, you really do need to look out for these and re-learn them as spam. Martin -- Martin Radford | "Only wimps use tape backup: _real_ [EMAIL PROTECTED] | men just upload their important stuff -o) Registered Linux user #9257 | on ftp and let the rest of the world /\\ - see http://counter.li.org | mirror it ;)" - Linus Torvalds _\_V
