Unless $ is not a consideration in your hardware purchases, then 460 SLI was and still remains better. It is
very competitive with 5870 in CF for hundreds of $$$ less. Not only is a single
GTX460 @ 900mhz keeping up with a 5870, but with better SLI scaling, 460 SLI will be very close while costing $300 less. 5870 has little overclocking headroom and scales worse when overclocked compared to Fermi architecture.
Sorry
a lot of people could care less if 5870 is 10% faster than a GTX470, they would rather take $100 savings towards their next graphics card upgrade where that $100 will buy them 30-50%+ performance increase after selling their old card. In addition, I am sure many opted for the $400 GTX460 SLI setup and sacrificed 20% or so performance increase that 5870 CF would offer,
pocketing $300 in the process.
Well, I had crafted a well thought out, long-winded response to this but thanks to a chat window popping up from Digsby my back button crushed my dreams of an elegant, intellectual response.
Here's the casual version, quickly put out. Price/performance works only when all other things are equal (enough). For example, sli scaling means squat to those who want single card solutions, and I'd bet a very large sum of money that single card solutions dominate the computer landscape. Heat and noise matter to some people, and of course that was fixed in the GTX460 but as you well know, the GTX460 came out a long time after the 5850 and 5870; the damage had been done.
My point is that whether or not x is better price/performance than y actually doesn't matter to a lot of people, so you're correct (even though that was a typo on your part, so no need for you to be 'sorry'
🙂): many people
could care less about x's merely being faster than y for a non-proportional increase in price, but they do not! Many people could care less about the early adoption of technology at a high price premium, but they don't. Many people couldn't care less about price, as long as the performance is
good enough for them, and unless you have some hard data to show lots of AMD jumping ship to Nvidia (from a 58XX series card to a 460), I've got a tough time believing that so many people are obsessed with price/performance (without taking into account heat, noise, brand preference, whatever) as you claim.
Edit: you totally misunderstood what I meant in my note to OCGuy.
a) he claimed that cherry picking results is what lead to the decently positive conclusion for AMD in this article, and
b) I was curious whether he made that claim (post history doesn't go back that far) when the previous articles were released because they were based on cherry picked results as well (by OCGuy's metric) and if I recall correctly, I was one of the only people crying foul then.
That's it.