I wasn't talking about the initial post. I was talking about the fact that in _very stable_ states the same numerical simulation with a supposedly well-established set of interactions produces garbage. VVC> The point is that any metastable state in numerical simulations VVC> requires fine-tuned Hamiltonians.