THREE DIFFERENT SPAM FILTERING STRATEGIES
 

100% NO SPAM
For people who get little or no email from unknown senders. Simple, brute force, black and white approach to filtering.

99+% NO SPAM        DOWNLOAD
For people who get some but not a lot of email from unknown senders, and want nearly perfect spam control. This is the filter set I use. Currently I get 99.7% spam filtering with this set. (It's optimized for the spam I receive, of course - your results may vary)

95% NO SPAM          DOWNLOAD
Recommended for people who get lots of email from unknown senders or who need to minimize the risk of losing email from unknown senders. 95%-plus spam detection, and only 2% false positives with NO whitelisting in my sample email sets.  Add a whitelist filter and you should get an extremely low rate of false positives.
 

 
100% NO SPAM, GUARANTEED

This simple filtering strategy is a black & white, minimalist approach to filtering, and can be accomplished with as few as two filters. The first filter is a whitelist filter, or set of filters, and intersects a nickname in your address book  containing all of the email addresses you wish to receive mail from. Eudora v5.2 and later also has the options to intersect or not intersect an entire address book.

The second filter checks for nothing, and simply sends every other incoming email to the trash. This blunt-force approach to filtering is not really useful to most of us, as we do expect to get legitimate email from people or organizations  we don't know (and therefore can't whitelist). But if you only want to get email from known sources such as family, friends, and business correspondents, and if you never want to see another piece of spam, then this filtering strategy might be your solution.
 

 
99+% NO SPAM

(This is the filter set I use)

Step One: Identify Dangerous Attached Files
Email with attached files can contain dangerous viruses, and often these files come to us from the computers of people we know. One filter, placed first in the list before the whitelist filters, will clearly identify these potentially dangerous attached files. I send these emails to the inbox for special handling.

Step Two: Catch and Keep
Make a "whitelist" filter to catch ALL incoming email from your friends and family, business correspondents, and any other addresses that you wish to always accept email from. Whitelist filters transfer email to the inbox with no further filtering. In my case this filter is simply the "passlist" nickname in my Eudora address book. It contains all of my Address Book entries as appropriate, plus a few extra's I've added (this is easier to create than it sounds, don't worry). In Eudora v5.2 and later this filter can simply be set to "intersects address book" without having to create a special all-inclusive passlist nickname.

Also, make a filter to catch and keep email from any "friendly" domains that you want to receive all email from - your employer, places you shop online, your internet service provider (unless spammers are forging headers from them), etc.

Make any other special filters you might require to catch and keep specific emails such as newsletters, or lists you may subscribe to. etc

Step three: Catch and Destroy
Any email that passes the "catch and keep" filters is immediately suspect - it's apparently not from anyone you know, or else they'd be listed in your passlist or address book. Set up a highly targeted gauntlet of very specific spam filters for all email that makes it this far, and transfer the snagged emails either to a spam mailbox, or directly to the trash. By adding colored labels, priority tags or using the "make subject" filter Action on the incoming spam, it's easy to quickly identify which filter caught which spam and to review certain types of spam for possible "false positives" before deleting them for good.

Note* As of June 2003 I send all my incoming spam to the trash.mbx instead of a separate spam.mbx, and preview it there. My Eudora is configured to automatically empty the trash on exit. This system works well for me, but is not recommended if you are just starting out with the filters, until you are comfortable that you are not getting false positive hits.

(Optional) Step Four: Auto-Reply Asking for Password
(PLEASE NOTE THAT I DON'T USE OR RECOMMEND THIS OPTION)
The emails that make it past your last "catch and destroy" filter are not from  people you know, but don't contain anything that your filters have identified as spam-like. If no further action is taken they go to your inbox by default. But if you really don't want to get any spam, you might consider setting up a filter at the end of the list that looks for a keyword - a "secret" password such as your first name (as long as that's not your email login). Many of your friends and aquaintances will use your name already in their email to you, while many of the spammers are greeting you by your login name, such as "Hi Cecilw" in my case. This filter  would quarantine or delete the email from unknowns if it doesn't contain this keyword, and send an auto-reply explaining the need for the keyword to make it past your spam filters. No spammer is likely to bother replying to this or even will see it, since real spammers almost always use forged From: headers. Real people on the other hand will have an easy option for getting through your shields. The Auto-reply filter is not included in the downloadable filter file below but is shown here

Download GENERIC 99+% Filter Set:
    Filters99new.zip (updated 2-24-03)
After downloading, unzip this file and copy it to your Eudora data folder - where your mailbox files are located. Rename your existing filter file "filters.pce" to something else, and then rename this file from "filters99new.pce" to "filters.pce". (Eudora must be closed while you do this or it will overwrite this file when it closes). Follow the instructions in the Notes file to change several of the filters from my email or domain name to yours.

95% NO SPAM

For people who get lots of email from unknown senders and/or who need to minimize the risk of losing valid email from unknown senders.

Use the same 3-step strategy as outlined above for the 99+% spam filtering, but use the second, less aggressive set of Eudora spam filters.  In addition to the filters in this set it is strongly recommended that you create  friendly domain, mailing lists, and/or newsletter subscription protection filters as applicable.

If your login name is (hopefully) different than your given name, you can also add a filter to detect spam that greets you by your login name in the body of the message: "Hi CecilW" for example goes straight to my spam bucket.

Download 95% Filter Set:
   Filters_lowpos_latest.zip (updated 2-22-2003)
After downloading, unzip this file and copy it to your Eudora data folder - where your mailbox files are located.  

NOTE* This download now contains a nickname and address book whitelist (passlist) filter. Without the passlist filter I get about a 2% false positive rate with this filter set;  but your email may be different than my 4400 spam and friendly email samples. There are several entries in this filter file that need to be edited with your username and domain - instructions are included in the zip file. Also if you want to send spam to a mailbox other than the one I use (called SPAM), you must edit each spam filter accordingly (or just do a search and replace in your text editor program for "spam.mbx").

 

Notes*

Most Filters Are Optional

Not all of spam filters will be right for everyone. Everyone should definitely use some sort of whitelist filtering - I can't think of any reason not too. But all the other filters can be thought of as modular pieces of a whole system, and can be inserted or removed as needed.

It's possible to filter out all email from unknown sources containing HTML or an Http: link to a web site for example. But since this would include people with links to their homepages or to their web email service provider (Hotmail or Yahoo etc.) in their signatures, I don't filter for HTML. You would not want to apply this filter if you get a lot of valid email from people you don't know. Most of the really offensive spam contains HTML, but can be filtered on other terms. A secret keyword filter to block out all email from unknown persons is probably not good for most people, as it can make getting in touch with you a bother for people not on your whitelist. Not good if you do business over the internet perhaps, and don't wish to aggravate your clients. But for other people who only exchange email with known friends and family, it may be the perfect solution.

Watch the Filter Results, and Update As Needed

Once you have created your list of filters, you may initially get a few "false positives", emails you want to keep but that are inadvertently sent to the spam box. Watch for these, and take appropriate action when you find them - these cases should decrease to background noise levels after a few weeks. If they're from someone you know and love, or at least like, add the email address to your passlist. If they're from someplace you regularly get email from, add their domain to the appropriate inbox filter. If it's a newsletter or mailing list you're on, make an inbox filter for it.

NEXT PAGE - THE FILTERS