# Any good, free, Spam filtering techniques to share?

**URL:** <https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505>\
**Category:** In My Humble Opinion\
**Created:** [May 11, 2003, 4:07pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505 "2003-05-11T16:07:29Z")\
**Posts on this page:** 16\
**Page:** 1

<div class="post-metadata">

**Author:** ![Revtim](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/revtim/32/1042_2.png) [@Revtim](https://boards.straightdope.com/u/Revtim)\
**Post date:** [May 11, 2003, 4:07pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/1 "2003-05-11T16:07:29Z")

</div>

I’ve been pretty happy using [www.cloudmark.com](http://www.cloudmark.com) 's Outlook plug-in, but it’s not going to free forever. I _really_ don’t want to have to spend money to filter spam; when that happens “the spammers have won”.

Any suggestions? I was hoping there was a free, open source type product that works the way Cloudmark does, where users mark spam for the whole Cloudmark community. I found a UNIX thing, but I’m pretty married to Outlook on XP right now.

---

<div class="post-metadata">

**Author:** ![Doomtrain](https://avatars.discourse-cdn.com/v4/letter/d/0ea827/32.png) [@Doomtrain](https://boards.straightdope.com/u/Doomtrain)\
**Post date:** [May 11, 2003, 7:25pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/2 "2003-05-11T19:25:38Z")

</div>

Does Outlook let you set up filters based on the content? My filter catches anything with these words/phrases in the body:

penis  
this is not spam  
mortgage  
septic tank  
%  
penile  
ink

Granted, it catches some of my legit mail and doesn’t catch all the spam, but it cuts about 12 out of my usual 15 spams out. I just check my Possible Spam folder for anything from legit senders and delete the rest.

---

<div class="post-metadata">

**Author:** ![Revtim](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/revtim/32/1042_2.png) [@Revtim](https://boards.straightdope.com/u/Revtim)\
**Post date:** [May 11, 2003, 7:37pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/3 "2003-05-11T19:37:05Z")

</div>

Yes, that’s my last resort, to set up some home-rolled filters. Outlook actually makes it pretty easy to make them. I’d prefer a Cloudmark-like solution though, whereby when one person marks something as spam then everybody else using the product gets it automatically filtered.

---

<div class="post-metadata">

**Author:** ![AHunter3](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/ahunter3/32/368_2.png) [@AHunter3](https://boards.straightdope.com/u/AHunter3)\
**Post date:** [May 11, 2003, 10:28pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/4 "2003-05-11T22:28:36Z")

</div>

I use Eudora and I catch about 995 pieces of spam out of 1000 sent my way. Yes, these are “home-rolled filters”, and it’s an ongoing operation to keep adding new ones, but it’s probably the most effective system in the long run.

I have 771 rules, divided up into categories and with labels applied to the spam messages to indicate which category of spam-rules the message was caught by. (I can also select a trashed message and with the Shift key down go to the filters and see which specific filter snagged the message).

I have Eudora set to actually empty the trash only when I exit the program or manually select “Empty Trash”, and the Trash folder is sorted by spam-category. I usually take quick glance at the subject titles and senders to check for any legitimate mail accidentally snagged.

---

<div class="post-metadata">

**Author:** ![RealityChuck](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/realitychuck/32/195_2.png) [@RealityChuck](https://boards.straightdope.com/u/RealityChuck)\
**Post date:** [May 11, 2003, 11:05pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/5 "2003-05-11T23:05:25Z")

</div>

Replace Cloudmark\* with [Spammunition](http://www.upserve.com/spammunition/default.asp). It works just as well.

Some things to filter on:

$  
!  
free (in either subject or sender)  
opt

\*Have you noticed their plans? Those who have been a part of the system, creating the list of spammers they will be using, are given a whole $1 off a month off the price of the service. Cheapskates.

---

<div class="post-metadata">

**Author:** ![Mangetout](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/mangetout/32/19_2.png) [@Mangetout](https://boards.straightdope.com/u/Mangetout)\
**Post date:** [May 11, 2003, 11:17pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/6 "2003-05-11T23:17:34Z")

</div>

I use SpamPal ([www.spampal.org](http://www.spampal.org)) - it works primarily by the publicly available dodgy sending domain lists, but it has ciustomisable white and black lists.  
It also has a Bayesian analysis plugin that can learn what spam messages look like by statistically analysing the words and their distribution etc.

It marks suspect messages by inserting a word of your choise ( **SPAM** by default) into the subject line; you can then deal with it using the message rules in your email client.

For the first month, I set my rules up to merely divert anything containing **SPAM** to a dedicated folder; this gave me a chance to put a few trusted people on the white lists and to train up the Bayesian filter a bit.  
After that, I set it up to delete **SPAM** messages and reply with a standard message saying that the mail delivery failed; there is also an instruction (to insert one specific code word from a list) in my reply that anyone blocked in error can read and follow to get their message past the filter; spammers don’t read their replies, so they won’t get to see the instructions (and a bot simply would not be able to parse the instruction if it tried).

---

<div class="post-metadata">

**Author:** ![amarone](https://avatars.discourse-cdn.com/v4/letter/a/e0b2c6/32.png) [@amarone](https://boards.straightdope.com/u/amarone)\
**Post date:** [May 11, 2003, 11:27pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/7 "2003-05-11T23:27:20Z")

</div>

I have started an onslaught on SPAM using the ability to set your own filters in Outlook. After one week, I’m now catching 70 - 75% of SPAM. I can’t use the Outlook Adult and Junk filters as they filter out genuine e-mail too often. I work in retail software, so our legit. e-mails often have sales-type of words and phrases that trigger standard junk filters.

You have to look out for the anti-spam tactics they use - one I’ve noticed is to us a zero instead of the letter O, for example p0rn. Another is to put spaces between every letter. I now have a filter to catch “v i a g r a”. And they may use underscores: p\_o\_r\_n.

An awful lot of spam has strings of junk letters at the end of subjects or the e-mail itself. Like idgyee dkfjrjksyh fsddfgh. Presumably this is to try defeat some anti-spam measure, but I don’t know what.

Look out for the fact that Outlook scans for matches within words, not just whole words (maybe there is a way to specify this, but I don’t know it). I discovered this when I added a filter for “cum” and it deleted all my e-mails containing “document”. Actually, I don’t delete the spam - I divert it to a spam folder so I can see the effect of my filters.

The trickiest are those with an innocuous subject, then include nothing but an html picture. This gives no words to filter on.

---

<div class="post-metadata">

**Author:** ![Revtim](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/revtim/32/1042_2.png) [@Revtim](https://boards.straightdope.com/u/Revtim)\
**Post date:** [May 11, 2003, 11:28pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/8 "2003-05-11T23:28:40Z")

</div>

**RealityChuck** : Thanks, I’ll check that out. And, I did notice that Cloudmark is only taking a buck off. It didn’t really matter to me either way, since I knew I wasn’t going pay for it anyway. But it did strike me as cheap, as well.

**Mangetout** : Thanks also. I’m a little wary of the programs that access my mail on the server, though. But, the Bayesian filter sounds interesting; I might check it out just for that.

---

<div class="post-metadata">

**Author:** ![Mangetout](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/mangetout/32/19_2.png) [@Mangetout](https://boards.straightdope.com/u/Mangetout)\
**Post date:** [May 11, 2003, 11:53pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/9 "2003-05-11T23:53:37Z")

</div>

SpamPal works as a proxy between your mail client and server… for the record, it did go through a rather shaky patch (well, the beta did) where it would prevent the client from dopwnloading a message, but still allow it to be cleaned up off the server.

It got better though and I’ve now been using it for a month with no problems (80+ legit email a day and 100+ spam - last time I turned off the filter to look)

---

<div class="post-metadata">

**Author:** ![Doomtrain](https://avatars.discourse-cdn.com/v4/letter/d/0ea827/32.png) [@Doomtrain](https://boards.straightdope.com/u/Doomtrain)\
**Post date:** [May 12, 2003, 5:37am UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/10 "2003-05-12T05:37:36Z")

</div>

You may want to see if your ISP offers anything. A lot of em (AOL and MSN come to mind right off) have started promoting all kinds of spam control stuff.

---

<div class="post-metadata">

**Author:** ![Kalhoun](https://avatars.discourse-cdn.com/v4/letter/k/3bc359/32.png) [@Kalhoun](https://boards.straightdope.com/u/Kalhoun)\
**Post date:** [May 12, 2003, 1:35pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/11 "2003-05-12T13:35:28Z")

</div>

I use MSN and don’t get ANY spam. Ever. It’s truly miraculous. I don’t know how they do it, but it’s the best.

---

<div class="post-metadata">

**Author:** ![Revtim](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/revtim/32/1042_2.png) [@Revtim](https://boards.straightdope.com/u/Revtim)\
**Post date:** [May 12, 2003, 10:44pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/12 "2003-05-12T22:44:28Z")

</div>

> [@](#):
>
> \*Originally posted by GMRyujin \*  
> \*\*You may want to see if your ISP offers anything. A lot of em (AOL and MSN come to mind right off) have started promoting all kinds of spam control stuff. \*\*

My ISP is Adelphia Cable, and I guess they don’t have spam controls (at least I couldn’t find any on their web pages). I could have sworn they said they were going to add some with a recent mail upgrade, but I can’t find any. But, they at least _finally_ implemented web-based email access and increased the size limits of the individual emails and the mailbox as a whole.

Funny, there’s finally something AOL and MSN has over my local provider. I guess I could rub their faces in the fact that lowly AOL has spam controls and they don’t…

---

<div class="post-metadata">

**Author:** ![racekarl](https://avatars.discourse-cdn.com/v4/letter/r/6bbea6/32.png) [@racekarl](https://boards.straightdope.com/u/racekarl)\
**Post date:** [May 13, 2003, 1:22pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/13 "2003-05-13T13:22:39Z")

</div>

I thought that you could continue to use the cloudmark beta software for free, and only had to pay if you upgrade to version 1?

---

<div class="post-metadata">

**Author:** ![Revtim](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/revtim/32/1042_2.png) [@Revtim](https://boards.straightdope.com/u/Revtim)\
**Post date:** [May 13, 2003, 10:18pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/14 "2003-05-13T22:18:47Z")

</div>

That’s possible **racekarl** , I’ll have to look into it. Thanks!

---

<div class="post-metadata">

**Author:** ![Revtim](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/revtim/32/1042_2.png) [@Revtim](https://boards.straightdope.com/u/Revtim)\
**Post date:** [May 17, 2003, 4:34pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/15 "2003-05-17T16:34:16Z")

</div>

Hey **RealityChuck** , Spammunition does work well. Thanks!

I’m not crazy about the fact that you can’t delete the spam, though. Maybe a future release will keep the results of its analyzation, so you can delete it.

---

<div class="post-metadata">

**Author:** ![beagledave](https://avatars.discourse-cdn.com/v4/letter/b/6de8d8/32.png) [@beagledave](https://boards.straightdope.com/u/beagledave)\
**Post date:** [May 17, 2003, 4:43pm UTC](https://boards.straightdope.com/t/any-good-free-spam-filtering-techniques-to-share/174505/16 "2003-05-17T16:43:18Z")

</div>

At home I use Mail for Mac OSX which does Bayesian filtering…probably 90% or so effective.

At work, I have to run OS9, and so use Outlook Express. In addition to the built in Junk Mail filter set to “high”, I’ve followed the suggestions mentioned [here](http://www.macintouch.com/spam03.html) for filtering on the headers. I don’t get **any** legitimate mail routed from Asia or Latin America (or several other countries) and so have filtered on the IPs from those area. I use SpamCop to do the grunt work in figuring out where the mail originated from, and act accordingly. It’s taken a bit of time to set up, but it works well and much better solution than filtering on content (for this particular email clinet)
