AI gambling

>>8289781
Yeah I used my AI to play the roulette.
I won over $100k before they realized I was getting instructions from an earpiece and they backroomed me.

In terms of poker, computers are only unbeatable in heads up limit Texas Hold'em. Poker isn't even gambling if you aren't shit.

>>8289797
Yeah, but they don't have to be unbeatable. The smallest advantage will make profit in the long run, especially when running hundreds of instances on different poker sites.

because gambling websites aren't exactly legal, and furthermore don't allow users to plug in external programs. And if you kept winning then they'd treat you like a card counter and kick you out

never ever forget: the house always wins. If you're winning too much they don't have to let you play

>>8289797

A computer would starch full ring LHE and HUNLHE at most SPRs. Would probably beat almost any other games as well.

>>8289781

It is banned. Of course.

>>8289912

It is legal outside of the US and in 2 states iirc. They don't care how much you win, you're not playing the house.

>>8289785
>To play the roulette

There is no strategy to roulette, the odds are always against you..

>>8289939
>retards actually believe this

>>8289928

>They don't care how much you win, you're not playing the house.

The house would rather have four people breaking even then three losers and one big winner. They're a business, not a charity. Also payouts get taxed so smaller payouts means a smaller loss.

>>8289939
Actually no You can always win roullete if you have enough money.

>bet 1 buck on black
>Lose
>Bet 3 buck on black
>Win
>You have now made a profit

You just keep betting larger amounts until you win. Statistically, it's foolproof.

File: 1458026806253.jpg (84KB, 850x400px) Image search: [Google]

84KB, 850x400px

A computer can never win poker by just looking at cards because computers can't bluff or read people. Statistically, with 5 people playing, you should only win every one out of 5 hands. The reason people can win more than they lose is though bluffing their way out of hands they would lose with if they got to the River.

I've played poker against machines and one thing they always do is fold if you reraise them. That makes them look at their cards and calculate they would probably lose if they called, because they always assume you aren't bluffing.

>>8291149
>Statistically, it's foolproof.

You just need an infinite amount of money to start with and a casino that will allow infinitely high bets.

But yeah then you're all set.

>>8289928
This. There are probably tons of bots that haven't been caught yet, and quite a few that have on lots of sites. Also online poker is in somewhat of a legal grey area as far as playing it goes...
>>8291157
Poker is a lot more complicated than you think, friendo. I am shit at poker and programming, but could probably program a bot to beat you, just because I know a little bit about the math behind it and you obviously know jack shit.

>>8289781
Yes, in a course about Combinatorics we played around with a fake poker AI. Assuming each hand is equally likely, the probability that a type of hand occurs is simply the number of possible hands with that type divided by the total number of hands. For example, the probability to obtain a straight flush is 40/2598960 which is less than one in ten thousand. The probability to obtain a straight (or a straight flush) is 10240/2598960 ~ 0.004, or 4 in a thousand ect.

Worked great with a regular deck of cards, or 5 decks, but online poker you have no idea what they're using as everything is proprietary and almost certainly rigged to be more random than a regular meatspace casino game.

>>8291149
You think maybe that is why roulette wheels usually have such low limits?

>>8291157
>because computers can't bluff or read people
Nobody can read each other over the internet and software can imitate a bluff very precisely using alternating betting strategies and taking occasional calculated risks.

>>8291157
>That makes them look at their cards and calculate they would probably lose if they called, because they always assume you aren't bluffing.
No, they always assume there are given statistical probabilities of winning a hand, even a relatively simple NN-based AI could easily track opponents and determine their betting strategy among a given list of strategies, one of which could be the profile of a bluffer.

>>8291149

If you're banning winners then people have no reason to play and will move to a different site. The tax on payouts is paid by the players (if they decide to file their winnings). Also it wouldn't even matter as the total amount being withdrawn is unaffected by the differences in the wins and losses of the players.

>>8291149

A martingale strategy is incredibly bad. Although you usually win a small amount, the probability that you bust is high enough to keep the odds in the house's favor unless you have infinite credits, in which case playing for profit is pointless.

>>8291157

I hereby grant you the dumbest poster award.

>>8291168

Most popular sites are incredibly likely to have fair programming due to the massive amount of data available. Also, more "random" is better if your strategy assumes a fair game.

>>8291175

This has to do with the bankrolls of each casino and there are limits in every game (against the house). Even if you gave them a big edge, they wouldn't coinflip you for billions because they don't want that kind of variance (neither does anyone else who gambles for a living).

>>8291179

No one always bluffs or always bets for value unless they are incredibly bad. Software can easily imitate a reasonably balanced range.

>>8291196

A computer probably wouldn't need to adjust its default strategy against a "bluffer", as such a player will already lose to any reasonably balanced range. Although if you wanted it would be trivial to maximally exploit such an obvious mistake.

>>8289781
When I played poker years ago there were already winning poker bots online, in cash games as well as tournaments. Occasionally some were outed because people got suspicious and found that a couple of players played pretty much identically over a large number of hands.

Like another anon already said, limit holdem is a solved game for a bot, no way you're going to win. Honestly I could see a bot do well in HU PLO but I wouldn't know if individuals could manage to create one.

>>8291956
Also to add, in general the less players and the smaller the stacks, the easier it will become for a bot. It would be pretty easy for a bot to win any game as long as it's heads up and with 20 or so big blinds or less.