Well, the possibility I mentioned in another thread has come to pass: the PS3 has sold more this month than 360 has all year. (Yes, I know that says more about Microsoft's utter failure in Japan than Sony's success; that was the original context I mentioned it in.)
Seriously, 50k for PS3 this week is too high. Unless we buy the shaky argument about Silver Week, it seems somewhat at odds with standard decay patterns. I certainly expect PS3 to fall off heavily next week, but then I expected it this week too.
Stuff like this makes me question the possibility of ever developing strong prediction tools for sales. It seems like more of an art than even meteorology. Props to the people who are actually good at it (though pretty much all of them seem surprised this week, too!).