Quote Originally Posted by SKYMTL View Post
How many processors were tested? Without a huge sample size, there is no way to know if "C3" indeed overclocks better or even has that much better power consumption. We all know that both of those factors can change from chip to chip even within the same revision. ESPECIALLY when the power consumption figure is well within the margin of error (2W).

in this case, he needs roughly a few thousands pieces to verify
anyway the statistic could not take as a reference as the piece you got from retail shop is pure luck