On analysis of partition games

The paper examines disjoint subsets of strategies for the Lotto games in order to provide a criteria of their “relative strength”, i.e. to define which strategies are more likely to win. The proposed methods are based both on the results of simulation and analytical techniques from combinatorial search theory. We showed that analysis of disjoint subsets of the set of (n,m)-partitions allows one to choose strategies with high “winning ability”. Our focus is on the simulation of tournaments performed to check these assumptions. Actors in these elimination tournaments are the partitions and the payoffs are identical to that for the Lotto game. The simulation demonstrated that a contestant using the strategies from specially designed disjoint subsets wins with frequency near 0.9 in the elimination tournament if other contestants play the Nash equilibrium for the given (n,m)partitions.