INTERACTIVE EXPLORERMONTE CARLO TREE SEARCH
Think ahead.
Then think again.
Watch a search algorithm turn uncertainty into a move. Play Tic-Tac-Toe as X, then open the tree and see where O spent its simulations.
01 — PLAY THE GAME
Tic-Tac-Toe
X YOUO MCTS AGENT
Your turn
Choose any open square to begin.
505,000
Explore lessExplore more
THE IDEA
Search by
sampling futures.
MCTS does not need a handcrafted strategy or a trained model. It grows a decision tree by repeatedly exploring possible continuations, simulating outcomes, and using those outcomes to focus the next search.
UCB1 SELECTION SCORE
win rate + C × √(ln(parent visits) / child visits)Balance reward against uncertainty.