Any anonymous poster claiming 6 defective cameras in a row is either a troll or not knowing how to evaluate their camera's focusing. Depending on the forum, troll is more likely.
A new buyer really should go to a local pro oriented store and deal with them instead of on-line. An experienced store owner would be able to short circuit many of the reported problems. As it is, people upgrading have only forums for advice but there is no hands on access to the testing setup or usage for effective diagnosis. Nikon learned from the problem with so many perfectly good cameras returned when the D7000 was released so they supplied small locally owned stores as a higher priority than mega-on-line stores when introducing the D800. Yes, still, most of the returned D800's, a scan of the forums reveals, were from sight-unseen purchases from on-line. Those who had hands-on experience with it and a knowledgeable dealer seemed much less likely to have a problem. The odds of having the exact same problem with a random sampling of cameras is pretty low so other causes make more sense to investigate. There is still no indication that tests in this thread were done in a a way to determine if there is a problem or not. It is hard to come up with an appropriate diagnosis without some trustworthy data. A diagnosis, to be correct, has to account for all evidence, if any of it does not fit, the diagnosis is wrong. All we know now is that a self diagnosis of 3 cameras being defective was arrived at for reasons we do not know, with evidence we do not know. If all three cameras perform so close together, it is more logical that they are all performing correctly since everything has to be within tolerance to get such consistent reproducible results from random sampled units. If one has a defect, it would be expected that results would not be consistent with the others. Too much time has been wasted on speculation and assumptions in this case since none of us are there to see the testing methods. Stan St Petersburg Russia