You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Larry Rosenman <le...@lerctr.org> on 2004/08/19 10:09:45 UTC
Why HAM learn on BAYES_99?
Why did the below message get learned as HAM on a bayes_99?
This is with 3.0.0-rc1.
And, how can we trap this one?
I've gotten like 3 of them, and all have made it to my mailbox....
(I've relearned them as SPAM, but BAYES_99 is not sufficient to
get it marked as SPAM).
Ideas?
--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749
---------- Forwarded message ----------
Return-path: <qm...@finchdam.com>
Envelope-to: ler@lerctr.org
Delivery-date: Thu, 19 Aug 2004 02:58:28 -0500
Received: from mail4.finchdam.com ([208.184.186.215])
by lerami.lerctr.org with esmtp (Exim 4.41)
id 1BxhoZ-0003I6-9Y
for ler@lerctr.org; Thu, 19 Aug 2004 02:58:28 -0500
Message-ID:
<80...@finchdam.com.178802680>
Date: Thu, 19 Aug 2004 03:57:55 -0400 (EDT)
From: Urgent Real Estate! <qm...@finchdam.com>
Reply-To: Urgent Real Estate! <rm...@finchdam.com>
To: ler@lerctr.org
Subject: Foreclosure Listing Site Shutting Down!
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Unsubscribe: To stop receiving this, please send a message
to <rm...@finchdam.com> with "bye" in the subject
X-Granda: neo 1 2ij 2k9172 8ng5sl
X-Spam-Checker-Version: SpamAssassin 3.0.0-rc1 (2004-08-15) on
lerami.lerctr.org
X-Spam-Level: *
X-Spam-Status: No, score=1.9 required=5.0 tests=BAYES_99 autolearn=ham
version=3.0.0-rc1
Dear larry,
Where have you been?
Did you get my last message. If not, listen up because I have some startling news to share with you right now! This is so serious that it can't wait until Friday. Don't even think about buying a foreclosure until you read this urgent message.
To get the story simply goto:
http://www.finchdam.com/dfeasder/aqq6zeiIqjl8tnIoznIiy0p4sIy1mI14/1/159?c=742
<a href="http://www.finchdam.com/dfeasder/aqq6zeiIqjl8tnIoznIiy0p4sIy1mI14/1/159?c=742">Story Here</a>
Best Regards,
L.V.
President/CEO
P.S You have to hurry because once I expose this scandal going on right now in the foreclosure industry my site may be shut down. So this may be your last chance to get the inside scoop.
To end contact:
http://www.finchdam.com/kehdieyws/0/aqq6zeiIqjl8tnIoznIiy0p4sIy1mI14
<a href="http://www.finchdam.com/kehdieyws/0/aqq6zeiIqjl8tnIoznIiy0p4sIy1mI14">End Contact</a>
Foreclosure World
P.O.B. 859
Norwalk, CT. 06856
Re: Why HAM learn on BAYES_99?
Posted by Loren Wilton <lw...@earthlink.net>.
Well, just off the top of my head, I think the following might be worth a
point or two:
header FROM_BANG From =~ /!\s/
score FROM_BANG 1.2
header __SUBJ_BANG Subject =~ /!/
meta BANG_BANG FROM_BANG && __SUBJ_BANG
score BANG_BANG 2
describe BANG_BANG You're dead!
header X_HUH exists:X-Granda
score X_HUH 0.5
Untried, of course, and for all I know will hit on every message from AOL.
Loren
----- Original Message -----
From: "Larry Rosenman" <le...@lerctr.org>
To: <sp...@incubator.apache.org>
Sent: Thursday, August 19, 2004 1:09 AM
Subject: Why HAM learn on BAYES_99?
> Why did the below message get learned as HAM on a bayes_99?
>
> This is with 3.0.0-rc1.
>
> And, how can we trap this one?
>
> I've gotten like 3 of them, and all have made it to my mailbox....
>
> (I've relearned them as SPAM, but BAYES_99 is not sufficient to
> get it marked as SPAM).
>
> Ideas?
>
>
>
> --
> Larry Rosenman http://www.lerctr.org/~ler
> Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
> US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749
>
> ---------- Forwarded message ----------
> Return-path: <qm...@finchdam.com>
> Envelope-to: ler@lerctr.org
> Delivery-date: Thu, 19 Aug 2004 02:58:28 -0500
> Received: from mail4.finchdam.com ([208.184.186.215])
> by lerami.lerctr.org with esmtp (Exim 4.41)
> id 1BxhoZ-0003I6-9Y
> for ler@lerctr.org; Thu, 19 Aug 2004 02:58:28 -0500
> Message-ID:
>
<80...@finchdam.com.178802680>
> Date: Thu, 19 Aug 2004 03:57:55 -0400 (EDT)
> From: Urgent Real Estate! <qm...@finchdam.com>
> Reply-To: Urgent Real Estate! <rm...@finchdam.com>
> To: ler@lerctr.org
> Subject: Foreclosure Listing Site Shutting Down!
> Mime-Version: 1.0
> Content-Type: text/plain; charset=us-ascii
> Content-Transfer-Encoding: 7bit
> X-Unsubscribe: To stop receiving this, please send a message
> to <rm...@finchdam.com> with "bye" in the subject
> X-Granda: neo 1 2ij 2k9172 8ng5sl
> X-Spam-Checker-Version: SpamAssassin 3.0.0-rc1 (2004-08-15) on
> lerami.lerctr.org
> X-Spam-Level: *
> X-Spam-Status: No, score=1.9 required=5.0 tests=BAYES_99 autolearn=ham
> version=3.0.0-rc1
>
> Dear larry,
>
> Where have you been?
>
> Did you get my last message. If not, listen up because I have some
startling news to share with you right now! This is so serious that it can't
wait until Friday. Don't even think about buying a foreclosure until you
read this urgent message.
>
> To get the story simply goto:
>
http://www.finchdam.com/dfeasder/aqq6zeiIqjl8tnIoznIiy0p4sIy1mI14/1/159?c=742
>
> <a
href="http://www.finchdam.com/dfeasder/aqq6zeiIqjl8tnIoznIiy0p4sIy1mI14/1/15
9?c=742">Story Here</a>
>
>
> Best Regards,
>
> L.V.
> President/CEO
>
>
> P.S You have to hurry because once I expose this scandal going on right
now in the foreclosure industry my site may be shut down. So this may be
your last chance to get the inside scoop.
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> To end contact:
> http://www.finchdam.com/kehdieyws/0/aqq6zeiIqjl8tnIoznIiy0p4sIy1mI14
>
>
> <a
href="http://www.finchdam.com/kehdieyws/0/aqq6zeiIqjl8tnIoznIiy0p4sIy1mI14">
End Contact</a>
>
>
>
>
>
>
> Foreclosure World
> P.O.B. 859
> Norwalk, CT. 06856
>
>
>
>
>
Re: Why HAM learn on BAYES_99?
Posted by Daniel Quinlan <qu...@pathname.com>.
Matt Kettler <mk...@comcast.net> writes:
> Dan.. AFAIK that's an outright bug in spamassassin. One of the rules of
> autolearning introduced in the 2.61 series is to never autolearn anything
> that strongly contradicts existing learning (ie: never learn a message as
> ham that would be bayes_99)...
I'm not sure we really fixed it to begin with.
> Did that intentional feature somehow get reverted out of 3.x?
>
> http://bugzilla.spamassassin.org/show_bug.cgi?id=2437
*shrug*
Daniel
--
Daniel Quinlan
http://www.pathname.com/~quinlan/
RE: Why HAM learn on BAYES_99?
Posted by Larry Rosenman <le...@lerctr.org>.
Matt Kettler wrote:
> At 02:43 AM 8/19/2004 -0700, Daniel Quinlan wrote:
>>> Why did the below message get learned as HAM on a bayes_99?
>>
>> http://wiki.apache.org/spamassassin/AutolearningNotWorking
>
> Dan.. AFAIK that's an outright bug in spamassassin. One of the rules
> of autolearning introduced in the 2.61 series is to never autolearn
> anything that strongly contradicts existing learning (ie: never learn
> a message as ham that would be bayes_99)...
>
> Did that intentional feature somehow get reverted out of 3.x?
>
> http://bugzilla.spamassassin.org/show_bug.cgi?id=2437
That was my bug back then too.
I really think that bayes_99 needs to be higher score, and that the
Fix to not auto-learn as ham when bayes_99 is present needs to be put back
in.
This one slipped through with JUST the bayes hit. :(
LER
--
Larry Rosenman http://www.lerctr.org/~ler
Phone: +1 972-414-9812 E-Mail: ler@lerctr.org
US Mail: 1905 Steamboat Springs Drive, Garland, TX 75044-6749
Re: Why HAM learn on BAYES_99?
Posted by Matt Kettler <mk...@comcast.net>.
At 02:43 AM 8/19/2004 -0700, Daniel Quinlan wrote:
> > Why did the below message get learned as HAM on a bayes_99?
>
>http://wiki.apache.org/spamassassin/AutolearningNotWorking
Dan.. AFAIK that's an outright bug in spamassassin. One of the rules of
autolearning introduced in the 2.61 series is to never autolearn anything
that strongly contradicts existing learning (ie: never learn a message as
ham that would be bayes_99)...
Did that intentional feature somehow get reverted out of 3.x?
http://bugzilla.spamassassin.org/show_bug.cgi?id=2437
Re: Why HAM learn on BAYES_99?
Posted by Daniel Quinlan <qu...@pathname.com>.
Larry Rosenman <le...@lerctr.org> writes:
> Why did the below message get learned as HAM on a bayes_99?
http://wiki.apache.org/spamassassin/AutolearningNotWorking
--
Daniel Quinlan
http://www.pathname.com/~quinlan/