Making bayesian filter detect some new spam

[Scott Silva]

on 6-23-2010 8:22 AM Mauricio Tavares spake the following:
> I have been getting some spam that has not been triggering the
> bayesian filter. At least it seems that no matter how many times I
> feed them (sa-learn), I still get nothing on bayesian. So, let's take
> a look at the header of one of the latest offending emails (the full
> email is hiding at http://pastebin.com/S8Hfid2V):
> 
> Return-Path: <errores at energysistemonline.com>
> Delivered-To: rt at domain.com
> Received: from correo.energysistemonline.com
> (correo.energysistemonline.com [94.127.190.73])
> 	by mail.domain.com (Postfix) with ESMTP id 1D50A84402F
> 	for <support at domain.com>; Tue, 22 Jun 2010 21:46:30 -0400 (EDT)
> From: =?iso-8859-1?Q?Beatriz_G=F3mez_Esparza_-_Energy_Sis?=
> 	=?iso-8859-1?Q?tem?= <beatriz.gomez.esparza at energysistemonline.com>
> To: =?iso-8859-1?Q?Canal_de_distribuci?=
> 	=?iso-8859-1?Q?=F3n?= <support at domain.com>
> Subject: =?iso-8859-1?Q?Nuevo_T5850_HDTV_-_Canales_de_televisi=F3n_en_alta_definic?=
> 	=?iso-8859-1?Q?i=F3n?=
> Sender: =?iso-8859-1?Q?Beatriz_G=F3mez_Esparza_-_Energy_Sis?=
> 	=?iso-8859-1?Q?tem?= <beatriz.gomez.esparza at energysistemonline.com>
> Mime-Version: 1.0
> Content-Type: text/html; charset="iso-8859-1"
> Content-Transfer-Encoding: quoted-printable
> Date: Wed, 23 Jun 2010 03:46:28 +0200
> Message-ID: <20100623014628zzz.4668292A2AF307E8 at sionemailing>
> X-Mailer: ICS SMTP Component V2.53
> X-domain-com-MailScanner-Information: Please contact the ISP for more
> information
> X-domain-com-MailScanner-ID: 1D50A84402F.AFF1A
> X-domain-com-MailScanner: Found to be clean
> X-domain-com-MailScanner-SpamCheck: not spam, SpamAssassin (not cached,
> 	score=4.658, required 4.7, BAYES_50 0.00, FUZZY_AMBIEN 2.30,
> 	HTML_MESSAGE 0.90, MIME_HTML_ONLY 1.46)
> X-domain-com-MailScanner-SpamScore: ****
> X-domain-com-MailScanner-From: errores at energysistemonline.com
> X-Spam-Status: No
> 
> As you can see, BAYES completely ignored it. Since this is not the
> first time we received an email like this, and I do run sa-learn every
> night on the spam folders (I do move these emails to the said
> directories), I feel like I must be missing something. Suggestions?
It hit on bayes_50. Your system has that score set to 0.00