Bandit based Monte-Carlo Planning https://sites.ualberta.ca/~szepesva/papers/ecml06.pdf