Hi,
I looked at your filter and its dfinitely got a huge amount in there. If you built it up manually, it must be frustrating that it catches too much.
Mine is based on someone elses ideas, but the basic idea of scoring is very good. It seems to be that you are spending a huge amount of time of details while big things never get mentioned. Look at the scoring example I sent. It catches mainly by spotting that the mailto: in spams is usually different from the From address, or the To: header is left blank, or its html mail or from a dial-up account. None are fatal by themselves but if 3 are together its spam.
The advantage of this is a short(ish) .forward that keeps catching the poor sods because, bless them, a lot of these trailer park folk make the same basic pitch again and again.
For a really good page of hints on this, look at:
http://colondot.net/mbm/mailfilter.shtml
and if confused try:
http://www.exim.org/exim-html-3.30/doc/html/filter_toc.html
Hope this helps.
Patrick
