Beej's Bit Bucket

a day ago

The minimax algorithm is used for turn-based games where each player assumes the opponent will make optimal moves to minimize their loss and maximize their gain.
It builds a look-ahead tree of possible board positions, with leaf nodes scored as win (+∞), loss (-∞), or draw (0) for the current player.
Values are propagated upward: maximizing for the current player's turn, minimizing for the opponent's turn, until the root selects the best move.
A recursive pseudocode demonstrates the algorithm, including tracking the best move and handling depth limits with heuristic evaluation when the tree cannot be fully explored.
For complex games like chess, heuristic functions evaluate board positions (e.g., control of center) because full exploration is impossible.
Techniques like alpha-beta pruning reduce the tree size, and the negamax variant simplifies code for zero-sum games.

Hasty Briefsbeta