For funsies: the actual median score of KZ2 over the 9 week cycle is this week's number of 68,803. Standard deviation is pretty huge on that, but you can still guess at the ballpark without busting out MiniTab. Applying a simple "best fit" line to the range, you'd figure on definitely seeing next week's sales being on the "-" side of the +/- confidence interval for a best guess speculation (i.e.- next week will see 45k or fewer sales, most likely).
Now, compare that to the logic you're trying to apply to PS3's sales... KZ2 averages 166k a week. Current total sales of KZ2 are at 1.5m...
Halo 3 has sold 9.5m. Which model (median and linear regression analysis vs. taking the average) would you use to "best guess" when/if KZ2 would tie Halo 3 if H3 was discontinued?
You're telling me in 48 weeks (April '10), KZ will have sold over 9m copies?
Again, being overly generous, my model proposes we freeze the sales here and extrapolate (basically picking now as the median over an expected continuation of the downward trend), KZ would have to keep making the current numbers well into holiday '11. Taking the regression analysis into account, KZ won't be within pissing distance of the top 50 by then, and will never come close to Halo numbers.
Believing in the PLAYSTATION®3......IS.......S_A_C_R_I_L_E_G_E
