Last active
July 25, 2018 17:09
-
-
Save ShlomoAbraham/b698cf5d9de09b3546eb886187ac5af7 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| import random | |
| import itertools | |
| # solution is a dictionary: | |
| # - Key is a tuple of (score: int, cards_remaining: int) | |
| # - Value is a tuple of (continue_drawing: bool, expected_value: float) | |
| solution = { | |
| (-1, 1): (True, 0), | |
| (1, 1): (False, 1) | |
| } | |
| # build up solution (dynamic programming) | |
| def calculate(current, total_remain, b_remain, r_remain): | |
| if (current, total_remain) not in solution: | |
| if b_remain == 0 or total_remain == 0: | |
| solution[(current, total_remain)] = (False, current) | |
| else: | |
| b_ev = calculate(current + 1, total_remain - 1, b_remain - 1, r_remain) | |
| r_ev = calculate(current - 1, total_remain - 1, b_remain, r_remain - 1) | |
| b_pct = b_remain / total_remain | |
| r_pct = r_remain / total_remain | |
| ev_draw = (b_ev * b_pct) + (r_ev * r_pct) | |
| do_draw = ev_draw > current | |
| ev = ev_draw if do_draw else current | |
| solution[(current, total_remain)] = (do_draw, ev) | |
| return solution[(current, total_remain)][1] | |
| def do_strategy(deck, score, cards_remaining, verbose): | |
| if cards_remaining <= 0: | |
| return 0 | |
| next_step = solution[(score, cards_remaining)] | |
| if verbose: | |
| print('{0}: {1}'.format((score, cards_remaining), next_step)) | |
| if next_step[0]: | |
| card = deck[-1 * cards_remaining] | |
| new_score = score + 1 if card[1] == 'black' else score - 1 | |
| return do_strategy(deck, new_score, cards_remaining - 1, verbose) | |
| else: | |
| return score | |
| COLORS = ['red', 'red', 'black', 'black'] | |
| RANKS = '23456789TJQKA' | |
| DECK = tuple(card for card in itertools.product(RANKS, COLORS)) | |
| my_deck = random.sample(DECK, 52) | |
| calculate(0, 52, 26, 26) | |
| result = do_strategy(my_deck, 0, 52, True) | |
| print(result) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| import random | |
| import itertools | |
| import functools | |
| import numpy as np | |
| solution = { | |
| (-1, 1): (True, 0), | |
| (1, 1): (False, 1) | |
| } | |
| def calculate(current, total_remain, b_remain, r_remain): | |
| if (current, total_remain) not in solution: | |
| if b_remain == 0 or total_remain == 0: | |
| solution[(current, total_remain)] = (False, current) | |
| else: | |
| b_ev = calculate(current + 1, total_remain - 1, b_remain - 1, r_remain) | |
| r_ev = calculate(current - 1, total_remain - 1, b_remain, r_remain - 1) | |
| b_pct = b_remain / total_remain | |
| r_pct = r_remain / total_remain | |
| ev_draw = (b_ev * b_pct) + (r_ev * r_pct) | |
| do_draw = ev_draw > current | |
| ev = ev_draw if do_draw else current | |
| solution[(current, total_remain)] = (do_draw, ev) | |
| return solution[(current, total_remain)][1] | |
| def do_strategy(deck, score, cards_remaining, verbose): | |
| if cards_remaining <= 0: | |
| return 0 | |
| next_step = solution[(score, cards_remaining)] | |
| if verbose: | |
| print('{0}: {1}'.format((score, cards_remaining), next_step)) | |
| if next_step[0]: | |
| card = deck[-1 * cards_remaining] | |
| new_score = score + 1 if card[1] == 'black' else score - 1 | |
| return do_strategy(deck, new_score, cards_remaining - 1, verbose) | |
| else: | |
| return score | |
| SIM_LENGTH = 1000 | |
| def calc_eval(deck: np.ndarray): | |
| total = 0 | |
| vector = np.array(deck.copy()) | |
| for i in range(SIM_LENGTH): | |
| np.random.shuffle(vector) | |
| total += vector.cumsum().max() | |
| return total / SIM_LENGTH | |
| def choice(deck): | |
| return calc_eval(deck) > 0 | |
| def run_mo(deck): | |
| total = 0 | |
| for i in range(52): | |
| c = choice(deck[i:]) | |
| if c: | |
| total += deck[i] | |
| else: | |
| return total | |
| return total | |
| COLORS = ['red', 'red', 'black', 'black'] | |
| RANKS = '23456789TJQKA' | |
| DECK = tuple(card for card in itertools.product(RANKS, COLORS)) | |
| calculate(0, 52, 26, 26) | |
| results = [] | |
| for i in range(100): | |
| my_deck = random.sample(DECK, 52) | |
| mod_deck = [1 if t[1] == 'black' else -1 for t in my_deck] | |
| result = do_strategy(my_deck, 0, 52, False) | |
| mo_result = run_mo(mod_deck) | |
| results.append((result, mo_result)) | |
| import statistics | |
| mean = statistics.mean([r[0] for r in results]) | |
| mo_mean = statistics.mean([r[1] for r in results]) | |
| outperforms = len([r for r in results if r[0] > r[1]]) | |
| ties = len([r for r in results if r[0] == r[1]]) | |
| mo_outperforms = len([r for r in results if r[0] < r[1]]) | |
| # print(result) | |
| # print('mo_result {0}'.format(mo_result)) | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment