Mailing List Archive

Question about hit rate
I'm new to this list and just go started using SpamAssassin a few weeks
ago. I'm currently not set up to use the Bayesian portion. I'm on a
windows system and wanted to start off slow. Plus, I don't quite
understand the logic, though I'm there is some. It seems that most
critical part of the Bayesian portion if for users to communicate (some
how) to the engine false negatives and positives. Is this correct? It
seems without that salearn would mark urls of spams that get through as
possibly good urls.

Which leads me to my major question. I've downloaded most of the rule
files from the Rules Emporium, but even with those rules, 1 out of every
2 or 3 spams still gets through. Is this normal?

Thanks,
Mark DeMichele
RE: Question about hit rate [ In reply to ]
33%-50% of spam making it thru????? No that is not normal. tripwire,
bigevil, evilnumbers, backhair.....ect alone should catch more then that.
Are you sure they are working???

It should be more like 1-4 per 1000 spams slipping thru.

--Chris

-----Original Message-----
From: Mark A. DeMichele [mailto:demi@intellipro.com]
Sent: Monday, February 09, 2004 8:05 AM
To: spamassassin-users@incubator.apache.org
Subject: Question about hit rate



I'm new to this list and just go started using SpamAssassin a few weeks ago.
I'm currently not set up to use the Bayesian portion. I'm on a windows
system and wanted to start off slow. Plus, I don't quite understand the
logic, though I'm there is some. It seems that most critical part of the
Bayesian portion if for users to communicate (some how) to the engine false
negatives and positives. Is this correct? It seems without that salearn
would mark urls of spams that get through as possibly good urls.

Which leads me to my major question. I've downloaded most of the rule files
from the Rules Emporium, but even with those rules, 1 out of every 2 or 3
spams still gets through. Is this normal?

Thanks,

Mark DeMichele
RE: Question about hit rate [ In reply to ]
I was just going to post a list of my rule files and discovered that a
couple of them had a "txt" extension tagged onto them. It must of
happened when I did a Save As from IE and I didn't notice. backhair and
tripwire were ok, the others were not. Maybe my hit rate will improve
now.



Thanks,

Mark



P.S.



I'm assuming when I reply to these messages that I should remove the
previous senders e-mail address and just leave the spamassassin-users
address. Is that correct. Sometimes there is no spamassassin-users
address.



_____

From: Chris Santerre [mailto:csanterre@MerchantsOverseas.com]
Sent: Monday, February 09, 2004 9:27 AM
To: Mark A. DeMichele; spamassassin-users@incubator.apache.org
Subject: RE: Question about hit rate



33%-50% of spam making it thru????? No that is not normal. tripwire,
bigevil, evilnumbers, backhair.....ect alone should catch more then
that. Are you sure they are working???



It should be more like 1-4 per 1000 spams slipping thru.



--Chris

-----Original Message-----
From: Mark A. DeMichele [mailto:demi@intellipro.com]
Sent: Monday, February 09, 2004 8:05 AM
To: spamassassin-users@incubator.apache.org
Subject: Question about hit rate

I'm new to this list and just go started using SpamAssassin a
few weeks ago. I'm currently not set up to use the Bayesian portion.
I'm on a windows system and wanted to start off slow. Plus, I don't
quite understand the logic, though I'm there is some. It seems that
most critical part of the Bayesian portion if for users to communicate
(some how) to the engine false negatives and positives. Is this
correct? It seems without that salearn would mark urls of spams that
get through as possibly good urls.

Which leads me to my major question. I've downloaded most of the
rule files from the Rules Emporium, but even with those rules, 1 out of
every 2 or 3 spams still gets through. Is this normal?

Thanks,

Mark DeMichele
RE: Question about hit rate [ In reply to ]
Sure, no problem. SA only updates its rules on upgrades. Therefore you would
have to wait for the latest SA and upgrade to get new rules. Also New rules
go thru what is called a GA run. Think of it as VERY intense testing. But
the problem with all of this is spammers change way to often. 1 week after a
Spamassassin release, they have changed methods to get around.

What we did was form the SARE (SpamAssassin Rule Emporium) which has about
15 people who constantly write, tweak, and test rules for SA. (Don't ask me
who they are, I'll hurt myself replying!) We have some rulesets that work
great and can be found at :

<http://www.merchantsoverseas.com/wwwroot/gorilla/bigevil.cf>
http://www.merchantsoverseas.com/wwwroot/gorilla/bigevil.cf

<http://www.merchantsoverseas.com/wwwroot/gorilla/99_FVGT_Tripwire.cf>
http://www.merchantsoverseas.com/wwwroot/gorilla/99_FVGT_Tripwire.cf

http://www.emtinc.net/includes/backhair.cf
<http://www.emtinc.net/includes/backhair.cf> (includes popcorn.cf now!)

http://www.emtinc.net/includes/chickenpox.cf
<http://www.emtinc.net/includes/chickenpox.cf>

<http://www.emtinc.net/includes/weeds_2.cf>
http://www.emtinc.net/includes/weeds_2.cf

<http://www.yackley.org/sa-rules/evilnumbers.cf>
http://www.yackley.org/sa-rules/evilnumbers.cf

<http://mywebpages.comcast.net/mkettler/sa/antidrug.cf>
http://mywebpages.comcast.net/mkettler/sa/antidrug.cf

Also check out the link in my sig. There is a small unofficial webring
dealing with rules, wikis, and hampsterdance. :) Basically you just put
these cf files into your /etc/mail/spamassassin directory for sitewide use.
Restart spamd and *boom*!

Some files are updated monthly, weekly, daily, and whenever the hell I feel
like it. :-)

HTH,

Chris Santerre
System Admin and SA Custom Rules Emporium keeper
http://www.merchantsoverseas.com/wwwroot/gorilla/sa_rules.htm
<http://www.merchantsoverseas.com/wwwroot/gorilla/sa_rules.htm>
'It is not the strongest of the species that survives,
not the most intelligent, but the one most responsive to change.'
Charles Darwin

-----Original Message-----
From: Paul Randall [mailto:hostmaster@aztecinternet.co.uk]
Sent: Monday, February 09, 2004 10:29 AM
To: Chris Santerre
Subject: Re: Question about hit rate



Hi Chris,


I wondered if you could tell me. I am pretty new to Spsmassassin. These
rules for tripwire, bigevil ... etc that you mentioned. Are they installed
by default in the curent Spamassassin distro?


If so then could you tell me where they are installed please?

If not then where can I get them and is there a readme on how to install
them?


Thanks loads for your help.


Paul Randall



On 9 Feb 2004, at 14:27, Chris Santerre wrote:


33%-50% of spam making it thru????? No that is not normal. tripwire,
bigevil, evilnumbers, backhair.....ect alone should catch more then that.
Are you sure they are working???



It should be more like 1-4 per 1000 spams slipping thru.



--Chris

-----Original Message-----

From: Mark A. DeMichele [mailto:demi@intellipro.com]

Sent: Monday, February 09, 2004 8:05 AM

To: spamassassin-users@incubator.apache.org

Subject: Question about hit rate



I'm new to this list and just go started using SpamAssassin a few weeks ago.
I'm currently not set up to use the Bayesian portion. I'm on a windows
system and wanted to start off slow. Plus, I don't quite understand the
logic, though I'm there is some. It seems that most critical part of the
Bayesian portion if for users to communicate (some how) to the engine false
negatives and positives. Is this correct? It seems without that salearn
would mark urls of spams that get through as possibly good urls.


Which leads me to my major question. I've downloaded most of the rule files
from the Rules Emporium, but even with those rules, 1 out of every 2 or 3
spams still gets through. Is this normal?


Thanks,


Mark DeMichele
RE: Question about hit rate [ In reply to ]
Thanks. I have all those rules now that I fixed the "txt" extension
problem earlier. However, since I fixed it about an hour ago, I
received 9 spams. 6 were marked correctly and 3 made it through. I
know things are kind of working because when I look at their headers,
they do in fact have scores. Unfortunately the scores are less than 5.



Also, I found set of blacklist files at
http://www.stearns.org/sa-blacklist/



The instructions to install them are kind of weird. It seems to me that
I can just put the sa-blacklist.current.cf

and sa-blacklist.current.uri.cf files in the rules folder and it should
just work.



Does anyone know about these files?



BTW, I'm using No Spam Today! on windows as a SMTP proxy. The way it's
set up, it shell's out to spamassassin each and every time. So when you
change rules you don't have to restart anything. This doesn't seem to
be the most optimal setup, but it works.



Thanks again for your responses.



Mark DeMichele



P.S.



Also, can someone tell me if I can leave the senders e-mail addresses in
my To: when I reply.



Thanks again.



_____

From: Chris Santerre [mailto:csanterre@MerchantsOverseas.com]
Sent: Monday, February 09, 2004 10:58 AM
To: 'Paul Randall'
Cc: Spamassassin-Talk (E-mail)
Subject: RE: Question about hit rate



Sure, no problem. SA only updates its rules on upgrades. Therefore you
would have to wait for the latest SA and upgrade to get new rules. Also
New rules go thru what is called a GA run. Think of it as VERY intense
testing. But the problem with all of this is spammers change way to
often. 1 week after a Spamassassin release, they have changed methods to
get around.



What we did was form the SARE (SpamAssassin Rule Emporium) which has
about 15 people who constantly write, tweak, and test rules for SA.
(Don't ask me who they are, I'll hurt myself replying!) We have some
rulesets that work great and can be found at :



http://www.merchantsoverseas.com/wwwroot/gorilla/bigevil.cf
<http://www.merchantsoverseas.com/wwwroot/gorilla/bigevil.cf>

http://www.merchantsoverseas.com/wwwroot/gorilla/99_FVGT_Tripwire.cf
<http://www.merchantsoverseas.com/wwwroot/gorilla/99_FVGT_Tripwire.cf>

http://www.emtinc.net/includes/backhair.cf (includes popcorn.cf now!)

http://www.emtinc.net/includes/chickenpox.cf

http://www.emtinc.net/includes/weeds_2.cf
<http://www.emtinc.net/includes/weeds_2.cf>

http://www.yackley.org/sa-rules/evilnumbers.cf
<http://www.yackley.org/sa-rules/evilnumbers.cf>

http://mywebpages.comcast.net/mkettler/sa/antidrug.cf
<http://mywebpages.comcast.net/mkettler/sa/antidrug.cf>

Also check out the link in my sig. There is a small unofficial webring
dealing with rules, wikis, and hampsterdance. :) Basically you just put
these cf files into your /etc/mail/spamassassin directory for sitewide
use. Restart spamd and *boom*!

Some files are updated monthly, weekly, daily, and whenever the hell I
feel like it. :-)

HTH,

Chris Santerre
System Admin and SA Custom Rules Emporium keeper
http://www.merchantsoverseas.com/wwwroot/gorilla/sa_rules.htm
'It is not the strongest of the species that survives,
not the most intelligent, but the one most responsive to change.'
Charles Darwin

-----Original Message-----
From: Paul Randall [mailto:hostmaster@aztecinternet.co.uk]
Sent: Monday, February 09, 2004 10:29 AM
To: Chris Santerre
Subject: Re: Question about hit rate

Hi Chris,



I wondered if you could tell me. I am pretty new to
Spsmassassin. These rules for tripwire, bigevil ... etc that you
mentioned. Are they installed by default in the curent Spamassassin
distro?



If so then could you tell me where they are installed please?

If not then where can I get them and is there a readme on how to
install them?



Thanks loads for your help.



Paul Randall



On 9 Feb 2004, at 14:27, Chris Santerre wrote:



33%-50% of spam making it thru????? No that is not normal.
tripwire, bigevil, evilnumbers, backhair.....ect alone should catch more
then that. Are you sure they are working???

It should be more like 1-4 per 1000 spams slipping thru.

--Chris

-----Original Message-----

From: Mark A. DeMichele [mailto:demi@intellipro.com]

Sent: Monday, February 09, 2004 8:05 AM

To: spamassassin-users@incubator.apache.org

Subject: Question about hit rate



I'm new to this list and just go started using SpamAssassin a
few weeks ago. I'm currently not set up to use the Bayesian portion.
I'm on a windows system and wanted to start off slow. Plus, I don't
quite understand the logic, though I'm there is some. It seems that
most critical part of the Bayesian portion if for users to communicate
(some how) to the engine false negatives and positives. Is this
correct? It seems without that salearn would mark urls of spams that
get through as possibly good urls.



Which leads me to my major question. I've downloaded most of the
rule files from the Rules Emporium, but even with those rules, 1 out of
every 2 or 3 spams still gets through. Is this normal?



Thanks,



Mark DeMichele
Re: Question about hit rate [ In reply to ]
On Mon, 9 Feb 2004 11:23:29 -0500, "Mark A. DeMichele"
<demi@intellipro.com> wrote:
>Also, can someone tell me if I can leave the senders e-mail addresses in
>my To: when I reply.

You can, but then the sender will receive two copies: one directly from
you and the other from the list server. It's usually best to reply to
just the list, unless you just want to make a personal reply to the
sender that would be of no interest to the rest of the list. In that
case delete the Cc: to the list.

Hope this helps.
Alan
P.S. It doesn't hurt to your trim your replies. We don't need to see
the complete history of what you're replying to repeated again and again
underneath your reply.
--
Please avoid sending me Word or PowerPoint attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html
RE: Question about hit rate [ In reply to ]
Got it.
(how's that)