Monte Carlo Tree Search - Tic-Tac-Toe Visualization

A visualization of the Monte Carlo Tree Search algorithm, applied to Tic-Tac-Toe.

[JustAGiraffe]

What a great representation of MCTS and an awesome tool to go with it. Thanks for your work — it helped me understand the intuition behind the algorithm.

[DennisRhodes-cq]

though you say this video is not about MCTS, this video has helped me understand the basics of MCTS more efficiently than any other resource I've seen!

[abrahamcastanedomusic]

This is pretty cool man, I was looking for videos like this one, a step by step simulation is the best way to learn.

[applejuice]

Thanks this made the MCTS concept really clear

[kunmy]

Finally, I understand how the algorithm works. Thanks!!!

[tienna-yuri]

It was a great visualization about MCTS!! This help me a lot, thank you!!!

[albertocaruso]

Grande, tanti saluti dall'Italia

[stefano]

What rollout strategy is used in this example? Can it be changed?

[dipankarpurecha]

Hi, thanks a lot for this. I wanted to ask, do you save the states of the random simulations in order to avoid repeating the same simulation?

[KayKay-obtz]

Can u pls make chessbot by mcts with progressive bias and progressive unpruning and exploration bonus

[AC-jtwq]

Hi. I aspire to be like you. Thank you for the video.

[高天林-rw]

BRAVO, How can your repo only have 12 stars? ( 13 stars now

[koenvleugel]

What is the exploration parameter (c) in this example? 2.14? And how do you determine what the exploration parameter should be?

[erikm]

I dont understand how the simulation part works, 'a random game is simulated...' what does it mean, that every move is just randomized from that point until the game finishes? what if the decision tree is so deep that it cant finish, how do you then estimate the values?