Minimax

 

Moving on with adversarial games, my AI lecture presented an Interesting explication 

of the Minimax, here thebasis for the pseudocode describing how a computer might play 

( and win at) tic tac toe. The idea is to play another game in which X is trying to maximize 

the score, and 0 is trying to minimize it. A terminal win for X is +1, one for O is -1,

and a draw is 0.

I can use alpha beta pruning to diminish the number of iterations: some paths 

are discourging from the get-go.

There are 255, 168 possible tic tac toe games. No computer is going to calculate

all that. An algorithm that shows more probabilities of winning might be 

good enough!! (Depth-limited minimax).

                                            *     *     *

(Don't want to misrepresent things here. Tic tac toe is always a draw if 

both sides play optimally 👧. And because the board can be turned and/or flipped

 - think see-through board - there are only three games, depending on

where the first move is) .

                                            *     *     *

GOT IT!!

Fall is here...