3 search tree T (with branching factor b and depth d)
4 set of handles H of size as number of leaves (b^d)
5 search algorithm: chooses a move until arrives at a leave,
6 returns the handle in the leave
7 playout algorithm: selects a move until arrives at a leave,
8 returns outcome of the handle in the leave
10 passes to next level a next-selector(selector, level,
13 the mean of the best handle less the mean of the selected handle
16 playout algorithms to compare:
18 VCT: UVB once, then UCB