My understanding however from the "Windows 8 Doesn’t Want Your App. Try Again Later" post was that as part of the approval process, MSFT does automatically test the apps in a low performance environment, and if they don't meed specific benchmarks there (example in the post was startup or shutdown time of 2 second, but I can't remember exactly and am on mobile), they will reject the app and tel you to work on it.

So if that's really the case, maybe they are selective with it. Maybe they don't apply the same rigor to their own apps, which would be a shame since you'd think they have to set an example.