With the massive size of their infrastructure, I'm guessing it would be very difficult to conclusively identify a single query, let alone attribute it to Bing.

With the massive size of their infrastructure, I'm guessing it would be very difficult to conclusively identify a single query, let alone attribute it to Bing. Also, the process you describe isn't conclusive, since it's possible that someone else randomly searched for "hiybbprqag". Finally, it's a stronger argument the way it is. Otherwise, it's just Google's word against MS. With their method, they could show that they could predict the search results by sealing a prediction before running the search on Bing. It's independently verifiable.