Computer chess programs typically use some form of heuristics to assess how good a board position is, or other aspects of the game play. These chess heuristics can then be used with game tree search techniques such as minimax or Monti Carlo tree search. In older chess programs the heurstic might be hand codes, but now the heurustic may be a value network or policy network; forms of deep neuarl network that have been trained on real or simulated games.
Used on page 228