I recommend everybody not to try to be too clever and refer just to mathematical formulas -- although it might sound very professional. I think the ability to handhold D800 with various lenses (VR or not) and focal lengths is easy to test personally. Then you know for sure (or average).
I just had a quick, single set of a test images of a brick wall (I know, not the best possible target) holding a 24-70 f/2.8 at 70 mm using 1/30, 1/60, 1/125 and 1/500 second shutter speeds taken from a stand-up position. It seems that when looking at the images using 100% zoom, there is a clear difference between 1/60 and 1/125 but I could not see any difference between 1/125 and 1/250 (or 1/500). Of course, it is possible that the only shot at 1/125 just happened to be better than average.
When the images are shown at full size in a 24 inch display, I could not see blur in 1/30 second image. So, unless you are cropping or zooming a lot, a slightly blurred image could well be usable.