Mailing List Archive: Assigning score to custom rules (and rules in general?)

Assigning score to custom rules (and rules in general?)

Feb 10, 2004, 1:13 PM

Post #1 of 5 (921 views)

I was wondering how more experienced spamassassin admins have been
assigning scores to rules that are relevant to their site (and also per
user if need be). I've installed a few "aftermarket" rulesets now, and
some of the values are (IMO) either too low or two high based on how
many show up in the spams and hams i examine.

Is there a good methodology to figuring out appropriate scores? I was
thinking of writing a script that would look at the proportion of
ham/spam that this rule applies to, but that gives me a percentage, but
not a value.

How are you all doing it?

Blaise. <blaise@geeky.net>
--------------------------------
Random IRC quote:
<BombScare> i beat the internet
<BombScare> the end guy is hard

Re: Assigning score to custom rules (and rules in general?) [ In reply to ]

bah at webmedic

Feb 10, 2004, 1:26 PM

Post #2 of 5 (896 views)

Permalink

On Tuesday 10 February 2004 12:13, Blaise St-Laurent wrote:
> I was wondering how more experienced spamassassin admins have been
> assigning scores to rules that are relevant to their site (and also per
> user if need be). I've installed a few "aftermarket" rulesets now, and
> some of the values are (IMO) either too low or two high based on how
> many show up in the spams and hams i examine.
>
> Is there a good methodology to figuring out appropriate scores? I was
> thinking of writing a script that would look at the proportion of
> ham/spam that this rule applies to, but that gives me a percentage, but
> not a value.
>
> How are you all doing it?
>
>
Wow I wish I could answer you right on this but I'm not sure how. The way I do
my scoreing is very heavily weighted on bayes. With high thresholds. After I
have added quite a few mails to the bayes db I rarely get fp's.

--
-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-
Brook Humphrey
Mobile PC Medic, 420 1st, Cheney, WA 99004, 509-235-9107
http://www.webmedic.net, bah@webmedic.net, bah@linux-mandrake.com
Holiness unto the Lord
-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-~`'~-

Re: Assigning score to custom rules (and rules in general?) [ In reply to ]

mkettler at evi-inc

Feb 10, 2004, 1:41 PM

Post #3 of 5 (901 views)

Permalink

At 03:13 PM 2/10/2004, Blaise St-Laurent wrote:
>Is there a good methodology to figuring out appropriate scores? I was
>thinking of writing a script that would look at the proportion of
>ham/spam that this rule applies to, but that gives me a percentage, but
>not a value.

Don't bother writing your own script for this.. mass-check already comes
with SA and you can feed it mailboxes full of spam and nonspam. From there,
use hit-frequencies to make a pretty report..

As for score assignment, I personally like the "start low and tune up until
it works but doesn't hurt" approach.

RE: Assigning score to custom rules (and rules in general?) [ In reply to ]

csanterre at MerchantsOverseas

Feb 10, 2004, 2:26 PM

Post #4 of 5 (901 views)

Permalink

> -----Original Message-----
> From: Blaise St-Laurent [mailto:blaise@geeky.net]
> Sent: Tuesday, February 10, 2004 3:14 PM
> To: spamassassin-users@incubator.apache.org
> Subject: Assigning score to custom rules (and rules in general?)
>
>
> I was wondering how more experienced spamassassin admins have been
> assigning scores to rules that are relevant to their site
> (and also per
> user if need be). I've installed a few "aftermarket" rulesets
> now, and
> some of the values are (IMO) either too low or two high based on how
> many show up in the spams and hams i examine.
>
> Is there a good methodology to figuring out appropriate scores? I was
> thinking of writing a script that would look at the proportion of
> ham/spam that this rule applies to, but that gives me a
> percentage, but
> not a value.
>
> How are you all doing it?
>
>
> Blaise. <blaise@geeky.net>
> --------------------------------
> Random IRC quote:
> <BombScare> i beat the internet
> <BombScare> the end guy is hard
>

First, ahahahahah I like that sig quote!

Second, Yeah scoring is an art. Depends on your threshold. Rule number 1 is:
One rule should NEVER make the email a spam. I also prefer to score low.
Most of my rules are below 0.66! Bigevil is the only ones I score higher at
3.0. But those are a special case. It is better to let a spam slip by then
score an FP.

--Chris

Re: Assigning score to custom rules (and rules in general?) [ In reply to ]

Robert at Menschel

Feb 11, 2004, 10:20 AM

Post #5 of 5 (905 views)

Permalink

Hello Blaise,

Tuesday, February 10, 2004, 12:13:31 PM, you wrote:

BSL> I was wondering how more experienced spamassassin admins have been
BSL> assigning scores to rules that are relevant to their site (and also per
BSL> user if need be). I've installed a few "aftermarket" rulesets now, and
BSL> some of the values are (IMO) either too low or two high based on how
BSL> many show up in the spams and hams i examine.

Head for the rules wiki at http://www.exit0.us and find the link to my
BobMenschel pages there. One of the first pages in that collection
documents my scoring methods as of a couple of months ago. (I'm working
on enhancements to that, but the documented methods are quite workable as
a startig point.)

Bob Menschel