gregbo - 1:12 am on Oct 31, 2006 (gmt 0)

As always, the more the "press" covers the issue of click fraud, the more G, Y!, etc. will be asked to provide reasonable explanations of what they consider to be fraudulent, and how good they are at screening it out.

WRT the "one person is counted as two when using both a work and home computer," this type of measurement "error" is inevitable when the criteria used are cookies, useragents, etc., rather than statistical sampling of a user base.