Chapter 1 · The dawn of AI (1943–1990)1950s
Search and game playing
Game AI looks ahead through possible moves and picks the one with the best outcome (minimax).
Chapter 1 · The dawn of AI (1943–1990)1950s
Game AI looks ahead through possible moves and picks the one with the best outcome (minimax).
The computer must choose a move.
The computer imagines each of its moves, then each reply, down to the end of the game. It assumes you will reply as well as possible, scores each move by its worst case, and takes the best of those.
From any position there are some legal moves. Each leads to a new position with its own moves. Drawn out, the possibilities form a tree.
You are X. After each of your moves the computer shows every reply it considered. Tap one of its options to see how you could answer it.
Your move (X).
Make your first move and the computer’s search will appear here.
Noughts and crosses is small enough to search to the very end, so this computer never loses.
What does minimax assume about the opponent?
Why can't a computer search every chess position?
What does pruning do?