4 ms·

> This trust emerges from the mathematical and game-theoretic properties of the system, without depending on the trustworthiness of individual network participa

> This trust emerges from the mathematical and game-theoretic properties of the system, without depending on the trustworthiness of individual network participants. I'm very uncomfortable with the idea of game theory as the base of trust. Mathematics as used in cryptography is a pretty solid base, but game theory feels like something entirely different. How do I know that there isn't a participant in the game that is willing to just smash the board? The existing blockchains like Bitcoin do have a certain amount of centralization, so this is not about a very large number of individuals where you could find some comfort in statistics. What if a state actor decides to smash the board, they could probably exert enough pressure on the large players? What if there are more subtle ways to extract short term gains that game theory didn't anticipate?

woah 8y ago
Do you think that the “smash the board” case isn’t considered? Not that any given game theoretic analysis isn’t flawed, but the “madman attack” is one of the main scenarios that’s been considered in blockchain from the original bitcoin paper onwards.
DennisP 8y ago
Protocols don't necessary assume that players are profit-seeking. Ethereum's Casper for example allows that players may wish to smash the board; however, it makes attacks very expensive, limits the damage they can do, and makes it easy for the network to recover. Subtle ways to get extra gains can be a problem; e.g. people didn't figure out selfish mining attacks until Bitcoin was several years old. Cryptoeconomics is a new field, and it advances as people discover attacks and invent solutions to them.