1. google is generally better test site cause it use large amount of javascript, client end, so it relies less on connection speed. but of course, I don't think one test can judge the whole thing anyway.
2. 3 browsers were obviously operated in same network conditions, Im not sure if the connection speed would make much difference
3. Im glad ppl trying to analyze the testing process, I would only HOPE the guy who so thoroughly analyzed wired's test also did some analysis of apple'e test, maybe their test have many problems too?