| Article ID: | iaor20051889 |
| Country: | Germany |
| Volume: | 32 |
| Issue: | 1 |
| Start Page Number: | 3 |
| End Page Number: | 16 |
| Publication Date: | Jan 2003 |
| Journal: | International Journal of Game Theory |
| Authors: | Benam M., Arous G.B. |
In a game against nature, a player has two real valued functions, f(0,x) and f(1,x), defined for x=0 and x=1. He chooses a succession, X1, X2, . . . , Xn of realisations of x, and calculates the mean value Xn. His reward at time n+1 is f(X[n+1],Xn) and he seeks a policy to maximise the long run average reward. This paper discusses the optimal strategy for the player.