Connect. Communicate. Collaborate. Securely.

Home » Kerio User Forums » Kerio Connect » Spam score affected by "BAYES_00: -1.665"?
  •  
tedxas

Messages: 25
Karma: 0
Send a private message to this user
Gang --

My Kerio's been getting HAMMERED by silly viagra/cialis/YouNameIt emails and because of the varied ways the spam spells the words, I can't perform any relatively-accurate content filtering on it.

I've managed to tighten up my kill mark to 2.8.....but somehow these always seem to barely pass.

Does it have anything to do with BAYES giving it a value of "-1.665" as in the header below? And if so, why does BAYES grant it a "negative" value when it is so obviously spam?

Thanks for your counsel!

Ted

==========================
Return-Path: <beahm<_at_>basiltech.com>
X-Spam-Status: No, hits=2.4 required=2.8
tests=BAYES_00: -1.665,HTML_90_100: 0.043,HTML_MESSAGE: 0.001,
UPPERCASE_25_50: 0.039
X-Spam-Level: **
Received: from basiltech.com ([125.179.159.137])
by myserver.mydomain.com (Kerio MailServer 6.1.2)
for me<_at_>mydomain.org;
Wed, 1 Mar 2006 02:18:30 -0600
Message-ID: <000001c63d08$ab0fcd10$58eba8c0<_at_>ijn32>
Reply-To: "Olivette Beahm" <beahm<_at_>basiltech.com>
From: "Olivette Beahm" <beahm<_at_>basiltech.com>
To: me<_at_>mydomain.org
Subject: declinat or news
Date: Wed, 1 Mar 2006 03:18:06 -0500
MIME-Version: 1.0
Content-Type: multipart/alternative;
boundary="----=_NextPart_000_0001_01C63CDE.C239C510"
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 6.00.2800.1106
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106
  •  
Kerio_ktrumbull

Messages: 597
Karma: 2
Send a private message to this user
This indicates that you have not been marking messages as spam. The Bayes_XX score is given based on how you have trained the Bayes filter. The numbers range from 00 to 99, which is the percentage at which the Bayes filter considers the message to be spam. According to how you have trained the Bayes filter, the Bayes filter thinks the message has a 00% chance of being spam, and adjusts the spam score accordingly.

Please ensure that you are using the "Spam" buttons on every spam message you get.

Kevin Trumbull
Kerio Technical Support Team Leader
http://support.kerio.com
  •  
tedxas

Messages: 25
Karma: 0
Send a private message to this user
Ugh.....my users and I have been marking such messages religiously......using the "Kerio Thumbs" in Outlook 2003. Server stats even show that my users are marking spam manually.

How can I determine if Bayes is actually "learning" anything....or at least why it keeps missing the boat on these?

Thanks for your help!

Ted
  •  
Kerio_ktrumbull

Messages: 597
Karma: 2
Send a private message to this user
You can look at your spam log file in the Admin Console. You should see a bunch of entries that indicate that user XXXX marked a message as spam. You should also go through there and see if somebody accidentally marked a bunch of spam messages as non-spam, which I have a feeling may be part of the issue.

Kevin Trumbull
Kerio Technical Support Team Leader
http://support.kerio.com
  •  
tedxas

Messages: 25
Karma: 0
Send a private message to this user
Yes, I can see where some folks have manually marked spam as indicated in the spam log....but here's my current stats:

=====================
Spam filter statistics

Messages checked:
31,659

Spams detected
24,470

Messages marked as spam
278

Messages marked as not spam
28
=====================

As you can see, they're marking a good number as spam.....while marking only a few (comparatively) as "not-spam."

I think Bayes is being "taught".....but it's not "learning" anything, no?

  •  
Kerio_ktrumbull

Messages: 597
Karma: 2
Send a private message to this user
Those stats are only for the current "session", which is since the last time the server was restarted.

If you are recieving a lot of false negatives (messages not being identifed as spam) then this would indicate that somebody at some point marked a large number of messages that would normally be considered spam as not-spam. In this situation you would need to "reset" the Bayes filter by renaming the spamassassin folder in the store directory to something else, then the MailServer will regenerate the folder with a fresh (untrained) Bayes filter.

Kevin Trumbull
Kerio Technical Support Team Leader
http://support.kerio.com
  •  
Lyle M

Messages: 410

Karma: 7
Send a private message to this user
On a side note... are you using any of the other anti-spam features of KMS?

I find that a good balance of blacklist server subscriptions with reasonable settings (blocking or increasing spam score) has a noticable positive effect on our spam control.

Good luck,
Lyle
  •  
tedxas

Messages: 25
Karma: 0
Send a private message to this user
Hi Lyle --

I have 15 blacklists defined with "Increase score 1.0", CallerID is enabled as is SPF and Spam Repellant. So, I've got all the guns blazin'....and I'm still missing the target on a few.

Not a huge trajedy.....but somewhat annoying. As you can tell from the stats, Kerio is still killing off far more than it's missing.
  •  
Lyle M

Messages: 410

Karma: 7
Send a private message to this user
Although it's a bit time consuming, you might want to use your logs to evaluate the effectivnes of your chosen blacklists. If you can identify the lists that give you the most reliable positives, try giving them higher spam scores. I have set our most consistent blacklists to go so far as to drop connections from incoming listed servers.

A more aggressive stance with your blacklists may pay off in your scenario. I'm also curious to know what effect Kevin's suggestion has in your situation.

Cheers,
Lyle
Previous Topic: email stuck in queue timing out
Next Topic: can't reach kerio from anywhere but local
Goto Forum:
  


Disclaimer:
Kerio discussion forums are intended for open communication between forum members and may contain information and material posted by members which may be useful in learning about Kerio products. The discussion forums are not intended to provide technical support for any specific product. Any information implied or expressed in the discussion forums is that of the posting member. Kerio is in no way responsible for the information posted in the forums, or its accuracy. Kerio employees may participate in the discussions, but their postings do not represent an offical position of the company on any issues raised or discussed. Kerio reserves the right to monitor and maintain the forums to promote free and accurate exchange of information.

Current Time: Thu Nov 23 10:38:13 CET 2017

Total time taken to generate the page: 0.00439 seconds
.:: Contact :: Home ::.
Powered by: FUDforum 3.0.4.