I recently came across this paper Complex Dynamics in Learning Complicated Games which looks at why some games are easy to master whereas some harder skill based games such as Go or chess are much harder to master.
In simple games with a limited number of moves the most effective strategy is fairly easy to deduce, as a result the game becomes boring. However, in this paper it is argued that in more complex games where each move has a large number of possible permutations then the players action become less rational and it becomes more difficult to find an optimum approach. In effect the action of players can become random.
The authors of the study posit that this may have implications for financial markets (no shit Sherlock) It is always intrigues me that game theorists and academics fail to understand that the real world is dirty, our knowledge is imperfect and the people making the decisions are even less perfect.
One of the things that has always interested me about game theory is its somewhat sterile nature and I have wondered if this sterile nature does not simply consign it to abstract games that centre around geopolitcal conflict. The background for these games is that all players have a a near perfect understanding of the game being played. This understanding is termed the equilibrium point and is perhaps the largest flaw in the applicability of game theory to real world problems.
Our knowledge is constrained by our cognitive ability, the time we have to reach a decision and the quality of information we have. In some ways game theory does recognise this but it fails to understand that whilst we have bounded decision making capability we have an almost unbounded capacity for being irrational. It is this irrationality that undermines the usefulness of fields such as game theory and its red headed cousin economics.