there's a spider behaving badly out there: limelight networks inc. is apparently launching a new search engine, and their bot (1) doesn't identify itself as such; (2) doesn't obey robots.txt; (3) identifies its ua as "mozilla" (version 1.8.1.11); (4) and runs your pages' javascript - !
that last one is the real kicker - many websites are instrumented with javascript-launched analytics code, and this thing (almost) appears as a legitimate person to google analytics and other programs. so counts are way off in some places. i say "almost", since the bot isn't handling cookies, so each pageview records as a unique vist.
now, it could be said that that is a cool effort at overcoming the problem that spiders have with javascript navigation, instead of relying on proper anchor tags, but they didn't consider all the other consequences.
recommended: add an exclusion for the bank of ip addresses used by that bot: 208\.111\.154\.[0-9]+
note: i'm talking about exclusions from analytics results, not actually blocking such traffic from the website. for that, the ip would need to be blocked, and since the robot exclusion seems shaky at this time, one would just have to accept the continued (refused) page requests. assuming the bot knows what to ask for next, which it shouldn't if it never even got to your root page.
so, sorry guys, get your act together and then we'll talk, hmm ?