Firefox never beat chrome, and I like both of them the same amount.
When I first did the data dive, I thought Safari was worse than IE, which I thought made a much better blog post. That would have been way more surprising. It was after due diligence that I was forced to discard that result.
Firefox 4 and 5 in the first graph are significantly lower than Chrome. Obviously it's a pretty dodgy comparison to make since Chrome isn't split by versions, but I agree with hammock that it seems statistically unsound to arbitrarily drop some versions of some browsers, especially when they can't do the same for the Chrome results.
When I first did the data dive, I thought Safari was worse than IE, which I thought made a much better blog post. That would have been way more surprising. It was after due diligence that I was forced to discard that result.