Fun little blackjack code (low content)

would be neat to see a simulation where the person with the 100k bankroll had simulations where it had a set number of hands dealt, at $x per bet, playing according to the standard black jack strategy chart.

There are professional blackjack players, aren't there? it seems list the vast majority of participants will lose over any significant number of hands, just like any other casino gamble, but there are some long term winners, aren't there? (For the record, I am asking - I am not sure...haven't googled or anything).

Pro blackjack players work by varying there bet sizing when the deck is "hot". I'm actually building programs that can be used as the building blocks for various card games, and i'm definitely going to build a blackjack simulator (it would actually simulate hands instead of generating random numbers) with a lots of different bet sequences to prove to my degen friends you cant beat BJ without card counting

Your bankroll is 1000000 you are betting 1 a hand
The highest your bankroll reached was: 1000054

ran pretty good for a little while anyway

Now write it so that you double your previous bet if you lose

Your bankroll is 100000 you are betting 1 a hand
The highest your bankroll reached was: 100123
The number of hands it took for you to go broke was 195765
assuming 109 hands an hour it would have taken 1796.0091743119267 hours for you to go busto
The casino made 55.67900288611345 dollars an hour
The largest betsizing you made was 65536.0

65536 is a cool power of two (^_^)

revised code
+ Show Spoiler +

import random

bankRoll = 100000
constBankRoll = bankRoll
bigRoll = bankRoll
handCounter = 0
betSize = 1
constBetSize = betSize
bigBet = betSize
handsAnHour = 109
winning = False


print("Your bankroll is", constBankRoll," you are betting", betSize ," a hand"

while bankRoll > 0:
winning = False
if random.random() > .49 :
bankRoll -= betSize
else:
bankRoll += betSize
winning = True
if bankRoll > bigRoll:
bigRoll = bankRoll
if winning == False:
betSize = betSize * 2
if bankRoll >= constBankRoll:
betSize = constBetSize
if betSize > bigBet:
bigBet = betSize
#print(bankRoll)
handCounter += 1
hoursToBust = handCounter / handsAnHour
print("The highest your bankroll reached was:", bigRoll)
print("The number of hands it took for you to go broke was", handCounter)
print("assuming", handsAnHour," hands an hour it would have taken", hoursToBust ," hours for you to go busto"
print("The casino made", constBankRoll / hoursToBust, " dollars an hour"
print("The largest betsizing you made was", bigBet/2,)

cool initiave so far. if you want to refine it i'd start with improving this part to more accurately reflect the full range of possible real 'results' from 1 hand of blackjack and how often each can happen.

On April 03 2013 18:49 Rinny wrote: here is the code- python 3.2.3 but it shouldn't matter, notes its not a full black jack simulator, it just generates a random number from 0 to 1 and if the numbers less than or equal to .49 the player wins.

what would happen irl in this game you model if the payout worked exactly as it does in your model is that the players would quit very fast as it's too low variance and there isnt much chance for short term upswings. real blackjack is more likely to have a higher number for highest your br reached. another interesting thing to get into if youre gonna do more work on it is to read up a little bit about gambling psychology and try to make a stochastic model for how often a player will keep playing and when he will quit. From a documentary I once saw, casinos do this kind of work when deciding how to program slot machines, so it has built into it the optimum number of payouts and near-misses to keep peoples dopamine spiking at an optimum frequency for them to keep playing while still netting massive ev for the house. it's all about keeping u playing as long as possible for them.

being pro blackjack is actually pretty boss. watching the discards go into the shuffle and tracking the cards through the shuffle; watching which cards get tossed in front of an ace (and seeing that zone wont be split during the shuffle) so when that card sequence comes out you know you have a high probability of getting dealt an ace - Ace sequencing. The shit you have to do these days to get an edge goes way beyond simple card counting.

-