I believe that the stock VID is set at the chip fab and written into the chip itself, which means it's done before the chip reaches Asus, GB, MSI, etc. So with Fermi cards I suppose there's a simple binning process they could deploy wherein they assume lower VID = better chip. But I think stock VID is still really just an estimation of how 'good' each chip is really going to turn out to be once it's actually put onto a card and overclocked ...
Maybe someone could ask MSIAlex on the the "Ask Me Anythign Nvidia-Related" thread he started a ways back
In reality, the VID shows absolutely nothing relevant to overclocking. A low VID means it doesn't require as much voltage for stock as others so in theory, you have more voltage headroom when overclocking.
As I have said already in this thread, binning is nothing more than sorting them based on data. Whether it be VID, clocks, etc.