219 total views
Addendum: See Vitalik in the comments on #3:
I read the article; very unconvinced. The key part is the bottom of page 11 where the article reveals its model for how costly attacks on proof of stake systems are. The paper seems to think that in a proof of stake system, you win by having a longer chain than the other chains, and slashing is only there to prevent literal double-signing. This completely ignores the entire set of recent developments in PoS literature around Tendermint, Casper CBC, Casper FFG, etc, which are the entire basis for claims about PoS’s greater security. These newer protocols use a form of slashing where it is provably impossible to revert a finalized block without slashing 1/3 of the validator set, so an attacker loses not just interest but also principal.
The section on permissioned chains completely fails to model the reputational losses (and possible legal consequences) that would be incurred by nodes on the chain if they misbehave.