Re: [Xnews] How to score out...

Reality folded in on itself, and somewhere the following words from
"Ohmster" apeared in history:

> Mine was like this, pretty crude I guess:
<snip>

Quick trick, groups can be combined:

[^(comp|linux)\.]
% Kill all articles cross posted to...
Score: -9999
Xref: advocacy

Second trick, you can score multiple entries using double ":"

[.]
Score:: 9999
From: Ohmster|Girls|ffstockings|zoli
From: rypost|Andrea|Ariana|Lacie|~|gloop|deacon|Julia's
Subject: zoli|zoligirls|fucked hard|girrls|KISSING|zg

> Then I had to add my own name in there when I tried to use the
> dollardollardollar killer because now my own posts were blocked
> from this post, which is ridiculous because now *all* of my posts
> get scored high, even when there is no reply.

Try this instead (place before the other entries):

Score: =0
Message-ID: MyBigKitty@
Score: 3000
References: MyBigKitty@\S+>$
Score: 1000
References: MyBigKitty@([^>]*>){2,4}$

It will lock any posts you make so any further scoring will not affect
it, and flag the # of replies you set at a cutom levels... And yes,
that can be somewhat set within Xnews, but it's not as flexable.

[re-arranged]
> I kept putting words to kill the from line to get rid of the stupid
> porn spammers that keep posting to the image newsgroups.

> There really, really must be a better way, these From and Subject
> line killers get to be really, really long and unmanageable after
> a while.

I think I ended up using a "kill all but" filter when I was in those
groups... Lets see, where did I put those entries... Here we are:

[bin.*pictures\.erotica]
% Intrest Tracking...
Score: =1000
Subject: Engl?ish|Translat(ion|ed)
~Xref: (.*:)+
Score: =-3000
Subject: \b(dutch|german|italian|french|spanish)\b
~Subject: Translat(ion|ed)

% Content Pattern Filter...
Score:: =-7000
Subject: ([a-z])\1{10,}|\s{10,}|(\w\W\s?){5,}
Subject: cosplay|guarantee|incest|loli(ta|con)?|p[o0]rn|scat
Score: =0
Subject: ATTN|dou?jin|ecchi|fill|flood|hentai|manga|miss(ed|ing)|<new>|retry|req(uire|uest)?|seeking
~Xref: (.*:)+
Score: -9000
~Subject: ^(Re:)?\W{0,2}(ABPE\w|HCG|FUR)\W|\b(n|r(\d|br|p|pt)?)\b

The princaple was really simple, I just keep adding what I wanted to
allow until only the posts I wanted where left without a score.

The importaint entries are:

Score: =0
Subject: ATTN|dou?jin|ecchi ...
~Xref: (.*:)+
Score: -9000
~Subject: ^(Re:)?\W{0,2}(ABPE\w ...

They play off each other. The first "locking" what I consider a valid
post, and the second "downscoring" anything that dosn't use that groups
particular posting flags. Most binary groups have posting/tagging rules
that spammmers don't often follow making it's easier to implement a
"negative" score, with less trouble then the "Marko/Pollo" method. }:8P

> Wow Max, your score files are so freaking complex it will take
> a long time to figure them out but this is a good thing, I will
> study them.

Hey, if you have any questions I'll be here. Things are a little more
complicated since much of my key scoring has shifted into hamster now,
but I can still convert entries "on the fly" as needed.

--
Maxx "What's That Score" Pollare...

* Score-File Dump: http://www.dragonfur.ca/nowhere *
.