Monte Carlo Tree Search 3.1

Defensive strategy improved to trust the evaluation function more during the first few moves.
Defensive evaluation function with larger reward for blocking than making pairs made