- added reward sum and reward mean

tic_tac_toe.py

@@ -21,36 +21,42 @@ def play(player_provider: PlayerProvider, k_max=10000, with_debug=False):
 		run = True
 		other_player = players[-1]
 		while run:
-			for player in players:
+			do_stop = False
+			this_reward = 0
+			other_reward = 0
+			for this_player in players:
 				if with_debug:
 					print(f"---------------------------------------------------")
 					print(f"- Game {k:06d}, Move {move} -----------------------------")
 					print(f"---------------------------------------------------")
 				last_state = state
-				state, has_moved = player.move(state)
+				state, has_moved = this_player.move(state)
 				if with_debug:
 					print(to_state_string(last_state, state))
 				if not has_moved:
-					player.end_game(state, 0.0)
-					other_player.end_game(state, 0.0)
-					run = False
+					this_reward = 0
+					other_reward = 0
+					do_stop = True
 					if with_debug:
-						print(f"{player.mark}: No more moves")
-				if player.has_won(state):
-					player.end_game(state, 1.0)
-					other_player.end_game(state, 0.0)
-					run = False
+						print(f"{this_player.mark}: No more moves")
+				if this_player.has_won(state):
+					this_reward = 1
+					other_reward = 0
+					do_stop = True
 					if with_debug:
-						print(f"{player.mark}: Has won the game")
+						print(f"{this_player.mark}: Has won the game")
 
-				other_player = player
-				if not run:
+				if do_stop:
+					this_player.end_game(state, this_reward)
+					other_player.end_game(state, other_reward)
+					run = False
 					break
 
+				other_player = this_player
 				move += 1
 
 
-do_training = 1
+do_training = 0
 do_human_player = 0
 do_load_values = 1
 
@@ -72,7 +78,7 @@ p_hx = HumanPlayer(mark='X', name="Jens")
 p_mx = MachinePlayer(mark='X', params=MachinePlayer.Params(p_explore=0.1, alpha=0.1), values=x_values)
 p_mo = MachinePlayer(mark='O', params=MachinePlayer.Params(p_explore=0.1, alpha=0.1), values=o_values)
 
-K = 50
+K = 5
 N = 1000
 if do_training:
 	for k in range(0, K):