Robert_Charlton - 10:52 am on Sep 29, 2010 (gmt 0)
The thought did cross my mind that if this wasn't a bug, or wasn't somehow query-related, it might be some sort of rock-and-a-hard place A-B test. The Information Week article Jekko links to leads to some Google articles on studying user behavior triggered by hard search tasks, and those make me feel that this is not likely to be that kind of testing.
Here's the Google Research Blog article that Information Week cites...
Official Google Research Blog: Frowns, Sighs, and Advanced Queries --
How does search behavior change as search becomes more difficult?
And here's a link to page offering a pdf download of the research paper the blog discusses....
How does Search Behavior Change as Search Becomes More Difficult?
Anne Aula, Rehan Khan and Zhiwei Guan
From the paper...
Our study showed that there are signals available online, in real time, that the user is having difficulties in at least closed informational search tasks. Our signals together with the signals related to successful and less successful search strategies discovered in earlier research could be used to build a model that would predict the user satisfaction in a search session. This model, in turn, could be used to gain a better understanding of how often users leave search engines unhappy - or how often they are frustrated and in need of help, and perhaps an intervention, at some point during the search session.
Doesn't sound like that's what's happening here.