class testNNPV implements NeuralNetworkPolicyAndValue<int?, TicTacToePlayer> {

class TestNNPV implements NeuralNetworkPolicyAndValue<int?, TicTacToePlayer> {

  }
  @override
  Map<int?, double> getQValues(GameState<int?, TicTacToePlayer?> game) {
    return <int?, double>{
      0: 0,
      1: 0,
      2: 0,
      3: 0,
      4: 0,
      5: 0,
      6: 0,
      7: 0,
      8: 0,
    };

    var x = TicTacToePlayer.X;

        .getSimulationResult(iterations: 100, nnpv: testNNPV());

        .getSimulationResult(iterations: 100, nnpv: TestNNPV());

  Map<MoveType, double> getQValues(GameState<MoveType?, PlayerType?> game);

  Map<MoveType?, double> initialQs = {};
  double q = 0;

    this.q = 0,

    if (initialQs.isEmpty && nnpv != null) {
      initialQs = nnpv!.getQValues(gameState!) as Map<MoveType?, double>;
    }

          gameState: gameState,
          backpropObserver: backpropObserver,
          move: move,
          parent: this,
          root: root,
          c: c,
          depth: depth + 1,
          nnpv: nnpv,
          useValueAfterDepth: useValueAfterDepth,
          valueThreshold: valueThreshold);

        gameState: gameState,
        backpropObserver: backpropObserver,
        move: move,
        parent: this,
        root: root,
        c: c,
        depth: depth + 1,
        nnpv: nnpv,
        useValueAfterDepth: useValueAfterDepth,
        valueThreshold: valueThreshold,
      ); //q: initialQs[move] ?? 0);

      return priorScore;

      return 0;

    var ucb = ((winsByPlayer[player] ?? 0 + (draws * 0)) / visits) +
        (c * priorScore * sqrt(log(parent!.visits.toDouble() / visits)));
    return ucb;

    if (priorScore == 1.0) {
      return ((winsByPlayer[player] ?? 0 + (draws * 0.5)) / visits) +
          (c * sqrt(log(parent!.visits.toDouble()) / visits));
    } else {
      // Q[s][a] + c_puct*P[s][a]*sqrt(sum(N[s]))/(1+N[s][a])
      return q +
          c *
              priorScore *
              sqrt(parent!.visits.toDouble()) /
              (1.0 + visits.toDouble());
    }

          (aVisits == 0 || bVisits == 0)) {

          (aVisits == 0 && bVisits == 0)) {

      currentNode.visits += 1;

      double reward = 0.0;

        reward = 0.5;

        reward = currentNode.parent?.currentPlayer() == winner ? 1 : 0;

      // Q[s][a] = (N[s][a]*Q[s][a] + v)/(N[s][a]+1)
      currentNode.q = (currentNode.visits * currentNode.q + reward) /
          (currentNode.visits + 1.0);
      currentNode.visits += 1;

  double c = 1.41421356237; // square root of 2

    double? c,

        c: c,

        c: c ?? 1.41421356237, // square root of 2