My minimax algorithm should, in theory, work fine. I cannot find any obvious problems. But... whenever I run it as 'O', my algorithm says that the player is going to win. However, tic-tac-toe is a drawn game. I am very confused!
import time
positionsChecked = 0
maxDepthChecked = 0
drawn = 0
wonForAI = 0
wonForPlayer = 0
def areMovesLeft(board):
for i in board:
for j in i:
if j == '-':
return True
return False
def evaluate(board, humanSprite, AISprite):
movesLeft = areMovesLeft(board)
if not movesLeft:
return 0
for i in range(3):
if board[i][0] == board[i][1] and board[i][1] == board[i][2]:
if board[i][0] == humanSprite:
return -10
elif board[i][0] == AISprite:
return 10
if board[0][i] == board[1][i] and board[1][i] == board[2][i]:
if board[0][i] == humanSprite:
return -10
elif board[0][i] == AISprite:
return 10
if board[0][0] == board[1][1] and board[1][1] == board[2][2]:
if board[0][0] == humanSprite:
return -10
elif board[0][0] == AISprite:
return 10
if board[0][2] == board[1][1] and board[1][1] == board[2][0]:
if board[0][2] == humanSprite:
return -10
elif board[0][2] == AISprite:
return 10
return 0
def minimax(board, isMaxPlayer, depth, humanSprite, AISprite, alpha, beta):
global positionsChecked, maxDepthChecked, drawn, wonForAI, wonForPlayer
best = 0
if depth > maxDepthChecked:
maxDepthChecked = depth
movesLeft = areMovesLeft(board)
if not movesLeft:
positionsChecked += 1
return 0
score = evaluate(board, humanSprite, AISprite)
if score == 10:
return score - depth
positionsChecked += 1
if score == -10:
positionsChecked +=1
return score + depth
for i in range(3):
for j in range(3):
if board[i][j] == '-':
movesLeft = True
if isMaxPlayer:
best = -1000
for i in range(3):
for j in range(3):
if board[i][j] == '-':
board[i][j] = AISprite
var2 = minimax(board, not isMaxPlayer, depth + 1, humanSprite, AISprite, alpha, beta)
best = max(best, var2)
board[i][j] = '-'
alpha = max(alpha, best)
if alpha <= beta:
break
break
return best
else:
best = 1000
for i in range(3):
for j in range(3):
if board[i][j] == '-':
board[i][j] = humanSprite
var2 = minimax(board, not isMaxPlayer, depth + 1, humanSprite, AISprite, alpha, beta)
best = min(best, var2)
board[i][j] = '-'
beta = min(beta, best)
if beta <= alpha:
break
break
print(best)
if best > 0:
wonForAI += 1
elif best == 0:
drawn +=1
elif best < 0:
wonForPlayer += 1
return best
def calcBestMove(board, humanSprite, AISprite):
global positionsChecked
global maxDepthChecked
maxDepthChecked = 0
positionsChecked = 0
print("Best move being calculated...")
timeBeforeMove = time.time() # I am using this to calculate how long the program takes to find the best move
bestValue = -1000
for i in range(3):
for j in range(3):
if board[i][j] == '-':
board[i][j] = AISprite
moveValue = minimax(board, False, 1, humanSprite, AISprite, -1000, 1000)
board[i][j] = '-'
if moveValue > bestValue:
bestValue = moveValue
bestRow = i
bestCol = j
print(f"\nThe value of the best move is {bestValue}. The best move is ({bestRow + 1}, {bestCol + 1})")
timeAfterMove = time.time()
timeTaken = timeAfterMove - timeBeforeMove
print(f"The time it took the AI to find the best move is {timeTaken} seconds.")
print(f"The AI searched {positionsChecked} positions.")
print(f"Deepest search is {maxDepthChecked}")
print(f"Drawn is {drawn} and won for player is {wonForPlayer} and wonfor ai is {wonForAI}")
board[bestRow][bestCol] = AISprite
def printBoard(board): # this prints the board
print(" | | ")
print(f" {board[0][0]} | {board[0][1]} | {board[0][2]} ")
print(" | | ")
print("---------|---------|---------")
print(" | | ")
print(f" {board[1][0]} | {board[1][1]} | {board[1][2]} ")
print(" | | ")
print("---------|---------|---------")
print(" | | ")
print(f" {board[2][0]} | {board[2][1]} | {board[2][2]} ")
print(" | | ")
def switchTurns(humanSprite, AISprite, isHumanTurn, board): # controls the flow of the game
printBoard(board)
if isHumanTurn:
while True:
row = input("Enter row: ")
column = input("Enter column: ")
if (row == '1' or row == '2' or row == '3') and (column == '1' or column == '2' or column == '3'):
row = int(row)-1
column = int(column)-1
if 0 <= row <= 2 and 0 <= column <= 2 and board[row][column] == '-':
board[row][column] = humanSprite
break
else:
if not (0 <= row <= 2 and 0 <= column <= 2):
print("Out of range!")
else:
print("Box already filled")
else:
print("Not valid values!")
else:
calcBestMove(board, humanSprite, AISprite)
gameStatus = evaluate(board, humanSprite, AISprite)
if gameStatus == 0:
movesLeft = False
for i in range(3):
for j in range(3):
if board[i][j] == '-':
movesLeft = True
if not movesLeft:
print("It's a tie!")
else:
switchTurns(humanSprite, AISprite, not isHumanTurn, board)
else:
printBoard(board)
if gameStatus == -10:
print("Human wins!")
elif gameStatus == 10:
print("AI wins!")
return
def chooseFirstOrSecond():
playerGoesFirst = input("DO you want to be first or second? (f/s) ")
if playerGoesFirst == 'f' or playerGoesFirst == 's':
if playerGoesFirst == 'f':
switchTurns('X', 'O', True, [['-', '-', '-'], ['-', '-', '-'], ['-', '-', '-']])
if playerGoesFirst == 's':
switchTurns('O', 'X', False, [['-', '-', '-'], ['-', '-', '-'], ['-', '-', '-']])
else:
print("Invalid!")
chooseFirstOrSecond()
chooseFirstOrSecond()
Let me give you an example of the game I played:
DO you want to be first or second? (f/s) s
| |
- | - | -
| |
---------|---------|---------
| |
- | - | -
| |
---------|---------|---------
| |
- | - | -
| |
Best move being calculated...
The value of the best move is -4. The best move is (1, 1)
The time it took the AI to find the best move is 1.8505730628967285 seconds.
The AI searched 59492 positions.
Deepest search is 9
Drawn is 17158 and won for player is 20451 and wonfor ai is 2820
| |
X | - | -
| |
---------|---------|---------
| |
- | - | -
| |
---------|---------|---------
| |
- | - | -
| |
Enter row: 1
Enter column: 3
| |
X | - | O
| |
---------|---------|---------
| |
- | - | -
| |
---------|---------|---------
| |
- | - | -
| |
Best move being calculated...
The value of the best move is -4. The best move is (2, 1)
The time it took the AI to find the best move is 0.04171943664550781 seconds.
The AI searched 1322 positions.
Deepest search is 7
Drawn is 17529 and won for player is 20903 and wonfor ai is 2862
| |
X | - | O
| |
---------|---------|---------
| |
X | - | -
| |
---------|---------|---------
| |
- | - | -
| |
Enter row: 3
Enter column: 1
| |
X | - | O
| |
---------|---------|---------
| |
X | - | -
| |
---------|---------|---------
| |
O | - | -
| |
Best move being calculated...
The value of the best move is -6. The best move is (2, 2)
The time it took the AI to find the best move is 0.0019884109497070312 seconds.
The AI searched 66 positions.
Deepest search is 5
Drawn is 17531 and won for player is 20939 and wonfor ai is 2862
| |
X | - | O
| |
---------|---------|---------
| |
X | X | -
| |
---------|---------|---------
| |
O | - | -
| |
Enter row: 2
Enter column: 3
| |
X | - | O
| |
---------|---------|---------
| |
X | X | O
| |
---------|---------|---------
| |
O | - | -
| |
Best move being calculated...
The value of the best move is 9. The best move is (3, 3)
The time it took the AI to find the best move is 0.0 seconds.
The AI searched 4 positions.
Deepest search is 3
Drawn is 17531 and won for player is 20941 and wonfor ai is 2862
| |
X | - | O
| |
---------|---------|---------
| |
X | X | O
| |
---------|---------|---------
| |
O | - | X
| |
AI wins!
(Above 0 is an AI victory, below zero is a player victory.) As you can clearly see, it says that I am winning all the way up to the last turn... I have no idea why...
I think you flipped a comparison operator in your alpha-beta pruning. Specifically, in the if isMaxPlayer:
branch,
if alpha <= beta:
break
should be
if alpha >= beta:
break