There is no conpericy theory, if these cards can clock up to 1600mhz with no problem, why would AMD not put it into the performance segment with a 300 buck MSRP for a reference card. It would be simple to do right? Something has stopped them from doing so, pretty simple right? If we are to believe in the rumors of the overclocks... You have been touting those overclocks as of late, but when it comes to this you can't. So why is that? So why aren't they doing this? Business wise it would be a smart move. It would lock in 90% of the market with their cards, and only leave nV with 10% or so to play with the current Pascal till midrange Pascal arrives. And no GF stuff you just stated, its the same damn chip as the one they are releasing in a few days. No need for a 40 or 44 CU chip. So yields should be the same. You asked me before why AMD aren't binning chips for voltage and power usage. You know why, this is the reason why, some chips can do it some chips can't, and the yields of those chips that can are extremely limited. So the need to do it will actually back fire. Now if you remember in chip design clocks, are on of the base factors in it. So if they are getting chips with 1600 mhz on a decent basis (even if yields are low), they probably were going for something higher than what they ended up getting in their design schema. Things like that just don't happen by accident. just like nV, they didn't plan on the clocks they got, it was better than what they expected, but they were planning for high clocks though.