Cryptographic Implications for Artificially Mediated Games [article]

Thomas Kellam Meyer
2009 arXiv   pre-print
There is currently an intersection in the research of game theory and cryptography. Generally speaking, there are two aspects to this partnership. First there is the application of game theory to cryptography. Yet, the purpose of this paper is to focus on the second aspect, the converse of the first, the application of cryptography to game theory. Chiefly, there exist a branch of non-cooperative games which have a correlated equilibrium as their solution. These equilibria tend to be superior to
more » ... the conventional Nash equilibria. The primary condition for a correlated equilibrium is the presence of a mediator within the game. This is simply a neutral and mutually trusted entity. It is the role of the mediator to make recommendations in terms of strategy profiles to all players, who then act (supposedly) on this advice. Each party privately provides the mediator with the necessary information, and the referee responds privately with their optimized strategy set. However, there seem to be a multitude of situations in which no mediator could exist. Thus, games modeling these sorts of cases could not use these entities as tools for analysis. Yet, if these equilibria are in the best interest of players, it would be rational to construct a machine, or protocol, to calculate them. Of course, this machine would need to satisfy some standard for secure transmission between a player and itself. The requirement that no third party could detect either the input or strategy profile would need to be satisfied by this scheme. Here is the synthesis of cryptography into game theory; analyzing the ability of the players to construct a protocol which can be used successfully in the place of a mediator.
arXiv:1001.0054v1 fatcat:kzmm6pgnljdbnkmsbxvi7ixuai