I am absolutely new to decision theory . I came across this following payoff matrix in the book.(Math. Stats : John E Freund).
Player A
I II
Player B 1 7 -4
2 8 10
The value of the game is given as 8 units.However i . have a question . I agree the optimal strategy for Play B is "2" . But i want to argue that the optimal choice for player A is 2 . This is because in decision theory the assumption is made that "each player must choose a strategy without knowing what the opponent is going to do and that once a player has made a choice it can not be changed" ..Going by this logic, the optimal strategy for player A should be II , because i see in that case , the loss would be less overall for the moves of player B . This , logic if true gives the value of the game to be 10 , which is not correct. I want to know , where i am wrong.
The statement that "each player must choose a strategy without knowing what the opponent is going to do and that once a player has made a choice it can not be changed" just means the choice in each time the game is played are made without information of what the opponent played this time. The players are allowed to examine the whole payoff matrix and think about what the opponent might do. In this game Player B finds that move $2$ dominates move $1$. No matter what A does, B is better off playing $2$ than playing $1$, so he should play $2$ all the time. For A, neither strategy dominates the other, but A "knows" B will play $2$, so by playing $1$ A can lose $8$ instead of $10$. If B were so foolish to play $1$ it is true that A could do better by playing $2$, but we don't count on foolishness.