Path design

Traditionally, a game is well defined and each player has his own information set and risk attitude. Based on this, various forms of game can be discussed.

However, the game itself is part of the evolution. How it evolves also depends on each player's decision and the historical path. It's similar to Andrew Lo's adaptive market hypothesis (AMH). And it's different from the learning theory on that the complete form of the game is unknown. There is not a definite goal to put all backward reasoning to an end.

So rather than mechanism design, we should focus on path design. It's still kind of mechanism design, just in a dynamic way. And most important, feedback is needed in order to see how far the game has gone and how each player's information has evolved. Based on these information, we are able to design the next feasible path.

I am not quite crystal about the theory and its possible applications right now. I will review some literature and come back with more rigorous and applicable ideas later. So stay tuned on.