Single Comment

Yes -- to play a game like this well the computer would have to use what's called 'classical' game theory rather than, I suppose, 'combinatorial' game theory. In classical game theory, which is used for games of simultaneous movement, the possible choices for each player form the rows / columns of a matrix, and the entries of the matrix describe the value of the result to one of the players. The optimal strategy for each player is a vector giving the probability that the player should give to each possible choice. If the matrix is known then the calculation of the optimal strategies is straightforward. But the conventional ways of evaluating the value of a game position for standard chess would not apply here, so figuring out the entries to the matrix would be difficult. It might be a good research project for some grad student studying game theory, though.