def calculate_hand_strength(current_hole_cards, board_cards, full_deck):
"""
Implemented as described in the page 21 of the thesis in: http://poker.cs.ualberta.ca/publications/davidson.msc.pdf
"""
our_cards = current_hole_cards + board_cards
our_cards_set = frozenset(our_cards)
if len(our_cards_set) == 2:
return STRENGTH_TABLE_FOR_2_CARDS[our_cards_set]
else:
pokereval = PokerEval()
ahead = 0.0
tied = 0.0
behind = 0.0
our_rank = pokereval.evaln(our_cards)
# considers all two card combinations of the remaining cards
deck = list(full_deck)
for card in our_cards:
if card not in deck:
print "Warning! Card not in deck: "+str(card)+", current_hole_cards: "+str(current_hole_cards)+", board_cards: "+str(board_cards)
deck.remove(card)
opponent_cards_combinations = itertools.combinations(deck, 2)
for opponent_card1, opponent_card2 in opponent_cards_combinations:
opponent_rank = pokereval.evaln([opponent_card1] + [opponent_card2] + board_cards)
if our_rank > opponent_rank:
ahead += 1.0
elif our_rank == opponent_rank:
tied += 1.0
else:
behind += 1.0
hand_strength = (ahead + tied/2.0) / (ahead + tied + behind)
return hand_strength
def calculate_hand_strength(current_hole_cards, board_cards, full_deck):
"""
Implemented as described in the page 21 of the thesis in: http://poker.cs.ualberta.ca/publications/davidson.msc.pdf
"""
our_cards = current_hole_cards + board_cards
our_cards_set = frozenset(our_cards)
if len(our_cards_set) == 2:
return STRENGTH_TABLE_FOR_2_CARDS[our_cards_set]
else:
pokereval = PokerEval()
ahead = 0.0
tied = 0.0
behind = 0.0
our_rank = pokereval.evaln(our_cards)
# considers all two card combinations of the remaining cards
deck = list(full_deck)
for card in our_cards:
if card not in deck:
print "Warning! Card not in deck: " + str(
card) + ", current_hole_cards: " + str(
current_hole_cards) + ", board_cards: " + str(
board_cards)
deck.remove(card)
opponent_cards_combinations = itertools.combinations(deck, 2)
for opponent_card1, opponent_card2 in opponent_cards_combinations:
opponent_rank = pokereval.evaln([opponent_card1] +
[opponent_card2] + board_cards)
if our_rank > opponent_rank:
ahead += 1.0
elif our_rank == opponent_rank:
tied += 1.0
else:
behind += 1.0
hand_strength = (ahead + tied / 2.0) / (ahead + tied + behind)
return hand_strength
print ""
print "f6 result = %s\n" % pokereval.poker_eval(
game="holdem",
fill_pockets=1,
iterations=iterations_high,
pockets=[["Js", "Jc"], ["__", "__"], ["__", "__"]],
dead=[],
board=["__", "__", "__", "__", "__"],
)
print ""
print "f7 result = %s\n" % pokereval.poker_eval(
game="omaha",
fill_pockets=1,
iterations=iterations_high,
pockets=[["Js", "Jc", "7s", "8c"], ["__", "__", "__", "__"], ["__", "__", "__", "__"]],
dead=[],
board=["__", "__", "__", "__", "__"],
)
print ""
hand = ["As", "Ad"]
print "handval %s = %d " % (hand, pokereval.evaln(hand))
print ""
hand = ["Qc", "7d"]
print "handval %s = %d " % (hand, pokereval.evaln(hand))
pokereval = None
board=["__", "__", "__", "__", "__"])
print ""
print "f6 result = %s\n" % pokereval.poker_eval(
game="holdem",
fill_pockets=1,
iterations=iterations_high,
pockets=[["Js", "Jc"], ["__", "__"], ["__", "__"]],
dead=[],
board=["__", "__", "__", "__", "__"])
print ""
print "f7 result = %s\n" % pokereval.poker_eval(
game="omaha",
fill_pockets=1,
iterations=iterations_high,
pockets=[["Js", "Jc", "7s", "8c"], ["__", "__", "__", "__"],
["__", "__", "__", "__"]],
dead=[],
board=["__", "__", "__", "__", "__"])
print ""
hand = ['As', 'Ad']
print "handval %s = %d " % (hand, pokereval.evaln(hand))
print ""
hand = ['Qc', '7d']
print "handval %s = %d " % (hand, pokereval.evaln(hand))
pokereval = None
def calculate_hand_potential_without_heuristics(current_hole_cards, board_cards, round_id, full_deck):
"""
Implemented as described in the page 23 of the thesis in: http://poker.cs.ualberta.ca/publications/davidson.msc.pdf
"""
our_cards = current_hole_cards + board_cards
out_cards_set = frozenset(our_cards)
# hand potential array, each index represents ahead, tied, and behind
ahead = 0
tied = 1
behind = 2
hp = [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]]
hp_total = [0.0, 0.0, 0.0]
total = 0.0
pokereval = PokerEval()
our_rank = pokereval.evaln(our_cards)
# considers all two card combinations of the remaining cards for the opponent
opponent_cards_combinations = list(itertools.combinations(full_deck, 2))
indices = []
for index, cards in enumerate(opponent_cards_combinations):
card1, card2 = cards
if card1 in our_cards or card2 in our_cards:
indices.append(index)
for index in reversed(indices):
opponent_cards_combinations.pop(index)
for opponent_card1, opponent_card2 in opponent_cards_combinations:
opponent_rank = pokereval.evaln([opponent_card1] + [opponent_card2] + board_cards)
if our_rank > opponent_rank:
index = ahead
elif our_rank == opponent_rank:
index = tied
else:
index = behind
# hp_total[index] += 1.0 # original version
# all possible board cards to come
deck = list(full_deck)
dealt_card = current_hole_cards + board_cards + [opponent_card1] + [opponent_card2]
deck_without_dealt_cards = [card for card in deck if card not in dealt_card]
if round_id == 1: # flop
cards_combinations = list(itertools.combinations(deck_without_dealt_cards, 2))
# cards_combinations = random.sample(cards_combinations, int(len(cards_combinations)*0.2))
for turn, river in cards_combinations:
# final 5-card board
board = board_cards + [turn] + [river]
our_future_rank = pokereval.evaln(current_hole_cards + board)
opponent_future_rank = pokereval.evaln([opponent_card1] + [opponent_card2] + board)
if our_future_rank > opponent_future_rank:
hp[index][ahead] += 1.0
elif our_future_rank == opponent_future_rank:
hp[index][tied] += 1.0
else:
hp[index][behind] += 1.0
total += 1.0
hp_total[index] += 1.0 # new version
else: # turn
# cards = random.sample(deck_without_dealt_cards, int(len(deck_without_dealt_cards)*0.75))
cards = deck_without_dealt_cards
for river in cards:
# final 5-card board
board = board_cards + [river]
our_future_rank = pokereval.evaln(current_hole_cards + board)
opponent_future_rank = pokereval.evaln([opponent_card1] + [opponent_card2] + board)
if our_future_rank > opponent_future_rank:
hp[index][ahead] += 1.0
elif our_future_rank == opponent_future_rank:
hp[index][tied] += 1.0
else:
hp[index][behind] += 1.0
total += 1.0
hp_total[index] += 1.0 # new version
# the original formula:
# ppot = (hp[behind][ahead] + hp[behind][tied]/2.0 + hp[tied][ahead]/2.0) / (hp_total[behind] + hp_total[tied]/2.0)
# npot = (hp[ahead][behind] + hp[tied][behind]/2.0 + hp[ahead][tied]/2.0) / (hp_total[ahead] + hp_total[tied]/2.0)
# ppot: were behind but moved ahead: cant use the original hp_total, because the result isnt normalzied and because it dont work for the heuristics
# added hp[ahead][ahead] so already good hands wouldnt be given a below average potential
ppot = (hp[ahead][ahead] + hp[behind][ahead] + hp[behind][tied]/2.0 + hp[tied][ahead]) / (hp_total[behind]*1.5 + hp_total[tied]*1.0 + hp_total[ahead]*1.0)
# npot: were ahead but fell behind
# npot = ((hp[ahead][behind]/total)*2.0 + (hp[ahead][tied]/total)*1.0 + (hp[tied][behind]/total)*1.0)/4.0
return ppot
#!/usr/bin/python
from pokereval import PokerEval
p = PokerEval()
contador = 0
with open('./p054_poker.txt') as archivo:
for juego in archivo:
cartas = juego.rstrip().lower().split(' ')
mano_a, mano_b = cartas[:5], cartas[5:]
if p.evaln(mano_a) > p.evaln(mano_b):
contador += 1
print contador
def calculate_hand_potential_without_heuristics(current_hole_cards,
board_cards, round_id,
full_deck):
"""
Implemented as described in the page 23 of the thesis in: http://poker.cs.ualberta.ca/publications/davidson.msc.pdf
"""
our_cards = current_hole_cards + board_cards
out_cards_set = frozenset(our_cards)
# hand potential array, each index represents ahead, tied, and behind
ahead = 0
tied = 1
behind = 2
hp = [[0.0, 0.0, 0.0], [0.0, 0.0, 0.0], [0.0, 0.0, 0.0]]
hp_total = [0.0, 0.0, 0.0]
total = 0.0
pokereval = PokerEval()
our_rank = pokereval.evaln(our_cards)
# considers all two card combinations of the remaining cards for the opponent
opponent_cards_combinations = list(itertools.combinations(
full_deck, 2))
indices = []
for index, cards in enumerate(opponent_cards_combinations):
card1, card2 = cards
if card1 in our_cards or card2 in our_cards:
indices.append(index)
for index in reversed(indices):
opponent_cards_combinations.pop(index)
for opponent_card1, opponent_card2 in opponent_cards_combinations:
opponent_rank = pokereval.evaln([opponent_card1] +
[opponent_card2] + board_cards)
if our_rank > opponent_rank:
index = ahead
elif our_rank == opponent_rank:
index = tied
else:
index = behind
# hp_total[index] += 1.0 # original version
# all possible board cards to come
deck = list(full_deck)
dealt_card = current_hole_cards + board_cards + [
opponent_card1
] + [opponent_card2]
deck_without_dealt_cards = [
card for card in deck if card not in dealt_card
]
if round_id == 1: # flop
cards_combinations = list(
itertools.combinations(deck_without_dealt_cards, 2))
# cards_combinations = random.sample(cards_combinations, int(len(cards_combinations)*0.2))
for turn, river in cards_combinations:
# final 5-card board
board = board_cards + [turn] + [river]
our_future_rank = pokereval.evaln(current_hole_cards +
board)
opponent_future_rank = pokereval.evaln([opponent_card1] +
[opponent_card2] +
board)
if our_future_rank > opponent_future_rank:
hp[index][ahead] += 1.0
elif our_future_rank == opponent_future_rank:
hp[index][tied] += 1.0
else:
hp[index][behind] += 1.0
total += 1.0
hp_total[index] += 1.0 # new version
else: # turn
# cards = random.sample(deck_without_dealt_cards, int(len(deck_without_dealt_cards)*0.75))
cards = deck_without_dealt_cards
for river in cards:
# final 5-card board
board = board_cards + [river]
our_future_rank = pokereval.evaln(current_hole_cards +
board)
opponent_future_rank = pokereval.evaln([opponent_card1] +
[opponent_card2] +
board)
if our_future_rank > opponent_future_rank:
hp[index][ahead] += 1.0
elif our_future_rank == opponent_future_rank:
hp[index][tied] += 1.0
else:
hp[index][behind] += 1.0
total += 1.0
hp_total[index] += 1.0 # new version
# the original formula:
# ppot = (hp[behind][ahead] + hp[behind][tied]/2.0 + hp[tied][ahead]/2.0) / (hp_total[behind] + hp_total[tied]/2.0)
# npot = (hp[ahead][behind] + hp[tied][behind]/2.0 + hp[ahead][tied]/2.0) / (hp_total[ahead] + hp_total[tied]/2.0)
# ppot: were behind but moved ahead: cant use the original hp_total, because the result isnt normalzied and because it dont work for the heuristics
# added hp[ahead][ahead] so already good hands wouldnt be given a below average potential
ppot = (hp[ahead][ahead] + hp[behind][ahead] + hp[behind][tied] / 2.0 +
hp[tied][ahead]) / (hp_total[behind] * 1.5 + hp_total[tied] *
1.0 + hp_total[ahead] * 1.0)
# npot: were ahead but fell behind
# npot = ((hp[ahead][behind]/total)*2.0 + (hp[ahead][tied]/total)*1.0 + (hp[tied][behind]/total)*1.0)/4.0
return ppot