This project is read-only.

Monte-Carlo Search


A Monte-Carlo search works by generating a list of potential moves, and for each move playing out thousands of games at random on the resulting board. The move which leads to the best set of random games for the current player is chosen as the best move.


TobiasV wrote Mar 15, 2016 at 3:29 PM

How do you determine the best set of random games?

If you play all moves in a seperate game in memory this exponentially grows at each turn.

Turn 1: 300 potential moves -> Starts 300 games in memory.
For all 300 games next move has 298 potential moves -> Starts 298 games for each of the 300 games => 89700 games in memory. How many moves before you can determine if a given game is winning or losing?

Without some smart decision making you'll end up with too many games in memory to handle?

Or am I thinking wrong?