diff --git a/network.py b/network.py
index 657b924..86e0fe2 100644
--- a/network.py
+++ b/network.py
@@ -323,42 +323,9 @@ class Network:
         scores_means = [tf.reduce_mean(score) for score in all_scores]
         transformed_means = [tf.reduce_mean(score) for score in transformed_scores]
 
-
-        return ([scores_means, transformed_means])
-
-
         print(time.time() - start)
 
-        # count = 0
-        # # loop over boards
-        # for a_board in boards:
-        #     a_board_scores = []
-        #
-        #     # loop over all rolls, for each board
-        #     for roll in all_rolls:
-        #
-        #         # find all states we can get to, given the board and roll and the opposite player
-        #         all_rolls_boards = Board.calculate_legal_states(a_board, player*-1, roll)
-        #         count += len(all_rolls_boards)
-        #         # find scores for each board found above
-        #         spec_roll_scores = [self.eval_state(sess, self.board_trans_func(new_board, player*-1))
-        #                             for new_board in all_rolls_boards]
-        #
-        #         # if the original player is the -1 player, then we need to find (1-value)
-        #         spec_roll_scores = [x if player == 1 else (1-x) for x in spec_roll_scores]
-        #
-        #         # find the best score
-        #         best_score = max(spec_roll_scores)
-        #
-        #         # append the best score to a_board_scores, where we keep track of the best score for each board
-        #         a_board_scores.append(best_score)
-        #
-        #     # save the expected average of board scores
-        #     all_rolls_scores.append(sum(a_board_scores)/len(a_board_scores))
-        #
-        # # return all the average scores
-        # print(count)
-        # return all_rolls_scores
+        return ([scores_means, transformed_means])
 
 
     def calc_n_ply(self, n_init, sess, board, player, roll):