Saturday, November 9, 2002, 8:52:35 PM, rick wrote: > after 2 days it is working quite well. It uses probability instead of > static filters. Works wonderful so far.
Are you just using it to filter spam? I was interested in replacing all my bat rules with this, but I don't know how well it would work. Given that my current filters get 90% of my spam anyway, I'm not sure it's worth it. To the original message - I rely solely on filters, and get just a few spam messages in my inbox. I haven't had any false positives yet, but I check every few days. My spam filter is mainly a list of things that almost always occur with spam: "mortgage" "Valued customer" "\$\d{1,}(\.|,)?\d*" (that last regexp catches $100 etc. , I seem to get a lot of casino spam) If spam slips through the net, I look for identifying features to add to the filter. It works so far, but for people who get a lot of spam it might not work as well. Oh, I should say that the rule excludes anyone in my address book. I considered whitelisting any message that includes my name in the 'to' header, but can't figure out the best way to do it. -- Adam [EMAIL PROTECTED] The Bat! 1.62/Beta5 on XP Pro How come abbreviated is such a long word? ________________________________________________ Current version is 1.61 | "Using TBUDL" information: http://www.silverstones.com/thebat/TBUDLInfo.html