(Selfish mining in the sense that people are working on chains in private.)
Taking deposits in advance wouldn't address the "issue" either because they
could still work on the chains in private. The deposits have nothing
to do with that, they just prevent the NaS issue.
To me its not really a "problem" per se, but the implication is that
there's no free ride with proof of stake, at least as far as I can see.
People will still expend energy to come up with the
best solutions and the energy expenditure
won't go down.
Perhaps we are talking past each other but how do TaPoS/deposit miners expend the same or more energy than PoW miners under the same attack?
With Vitalik's proposal their isn't an exponentially increasing dynamic difficulty using more physical resources but only a increasing difficulty either penalizing the miners or taking deposit, all of which use practically no energy.
I don't really know to be honest, and I suppose it would depend on the specific PoS implementation...
but essentially, its back to the NaS problem of people preparing multiple chains. The only difference
is that the deposits prevent you from broadcasting them all immediately...but miners could still
spend extra resources and go "outside" the intended system (in a similar way that selfish mining
can happen in Bitcoin outside the intended methodology)..and when that happens, people start spending
extra resources (computing power).
I don't know under what circumstances this would/could happen or not, just putting it
out there as a possible reason why there might ultimately be no efficiency win -- and this
is what the truthcoin article is saying as well.