fn negamax(self: *@This(), alpha_orig: isize, beta_orig: isize, depth: usize) !isize {
        if (depth == 0) return self.evaluate();

    fn negamax(self: *@This(), alpha_orig: isize, beta: isize, depth: usize) !isize {
        if (depth == 0) return try self.quiescenceSearch(alpha_orig, beta);

        var beta = beta_orig;

        }
        // node (move) fails low
        return alpha;
    }
    // quiescenceSearch helps the engine not thow away its pieces because of
    // a temporary seemingly good move.
    fn quiescenceSearch(self: *@This(), alpha_orig: isize, beta: isize) !isize {
        var alpha = alpha_orig;
        const evaluation = self.evaluate();
        // fail-hard beta cutoff
        if (evaluation >= beta) {
            // node (move) fails high
            return beta;
        }
        // found a better move
        if (evaluation > alpha) {
            // PV (principal validation) node
            alpha = evaluation;
        }
        var ml = try MoveList.init(0);
        try self.generateMoves(&ml);
        for (ml.slice()) |move| {
            var bck: GameState = undefined;
            self.backup(&bck);
            self.ply += 1;
            if (!self.makeMove(move, .captures)) {
                self.ply -= 1;
                self.restore(&bck);
                continue;
            }
            const score = -try self.quiescenceSearch(-beta, -alpha);
            self.ply -= 1;
            self.restore(&bck);
            // fail-hard beta cutoff
            if (score >= beta) {
                // node (move) fails high
                return beta;
            }
            // found a better move
            if (score > alpha) {
                // PV (principal validation) node
                alpha = score;
            }