Why does IP address *ALWAYS* match?

I'm trying to implement an IP-address-based spam filter recipie
in addition to the filter I normally use.  I'm harvesting spam headers
in a log file, and I've written a DOS QBASIC program to parse the log
and generate a large filter recipie with one condition for each IP
address (actually the first 3 of the 4 blocks.  Here's a sample
from the spam log file...

Received: from mail.tampabay.rr.com (tampabay.rr.com [24.92.0.1])
        by gold.interlog.com (8.8.5/8.8.5) with ESMTP id QAA08018;
        Fri, 24 Oct 1997 16:46:17 -0400 (EDT)

    I want to test all headers for the string "[24.92.0.".  My
program escape's the bracket and dots, and generates the condition
 * !\[24\.92\.0\.
According to the log with VERBOSE=ON, this test (do NOT match)
returns false, i.e. the pattern has matched.  I don't see a backslash
immediately preceding a digit, so the pattern should *NOT* be a
repetition count.  Any ideas?  If I get really desparate, I could
always try...
 * !\[[2][4]\.[9][2]\.[0]\.
but it does seem a bit extreme.

-- 
Walter Dnes
waltdnes(_at_)interlog(_dot_)com