Leo's pill domains. Feed several to sa-learn (gets you a high BAYES score), make sure that net tests are enabled and do use digests (DCC, Razor and Pyzor); Then these spam will get 30+ point scores. Even with no net tests, your example scores > 4 points without BAYES, so training BAYES will cause it to be flagged. The SARE rules can also help quite a bit.
Some time has probably passed, but your example hits 5 SURBLs, URIBL, the SBL and RAZOR, as well as received in the XBL, SpamCopBL and a handful of standard non-net rules (INVALID_MSGID, FORGED_MUA_OUTLOOK and more). Paul Shupak [EMAIL PROTECTED]