Oct

27

Under Von Neuman's Minimax theorem, a randomized mixed strategy in a two party zero sum game with two equilibrium points will improve the odds over using the preferred of the two strategies. The prime example is the soccer penalty kick. Examining the game matrix, kicking to the left to the kickers preferred side returns 38%, kicking to the right (the non preferred side) yields 72%, but a randomized 60/40 mix of the two will yield a 79% success rate. The rational reason is that it is harder for the opponent to discover the strategy. The optimum weights can be solved algebraically. Markets can be modeled as two party/ bull/bear zero sum game. The parties presumably know each other's strategies. Would it not behoove one to use some sort of randomized method to avoid having your own strategy used against you? Is this a reason to use a trailing stop for example?


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search