[Opinion]: Will artificial intelligence kill online poker?

[karl coakley]

I'm sorry if this sounds argumentative but if you truly believe what you wrote here then you don't know the first thing about chess. To give you an example, there are 400 positions to analyze after both players made their first move. In the middle game there are up to 10^6 (1000000) positions a computer needs to analyze in order to calculate the correct move.

So no, writing a winning poker bot would not be impossible or require a lot of computing power. We already have HUD's and various other tools available for online poker that tracks player stats like aggression, how often they go to showdown, how often they call a re-raise and various other stats. We also have tools that can calculate the profitability of playing a certain hand in a certain position a certain way. Combine the two and you have a pretty strong poker bot.

No, the reason why you don't run into poker bots on major poker sites is simply because these sites have measures to detect bots. For example they track mouse movement and use it determine whether a human or program is playing.

I'm not an expert chess player, hopefully (maybe) average... but it doesn't take a math expert to know that there are a limited amounts of pieces and moves. I don't know how many moves that need to be analyzed, but your numbers show the disparity. If there are 1 million moves that need to be analyzed in the middle of a game, that would really be nothing. Deep blue could do 200 million a second. https://www.research.ibm.com/deepblue/meet/html/d.3.3a.html

Even way back when a dual core could do 8 million a second. http://www.spiegel.de/international...e-last-man-versus-machine-match-a-450566.html

In chess there is a definitive amount of pieces and moves. If programmed to play perfect, the computer CAN analyze all the pieces and moves to make the best (correct) move.

This is where poker is different.

You NEVER have a complete board and pieces to analyze. There also is no "perfect" way to play poker because it varies from person to person. I could teach you my style (which is a positive ROI) and you could play exactly the way I do and still not win. Its because it changes every day, session, and hand. Every decision is different.

I'm not taking away anything from chess players. I'm just stating that because the computer has the full information to analyze, it can, and make a perfect decision. It will never have the full information in poker and has to fill in the blanks. A computer will never do this better than a human. You can't program something to be creative.

I see stuff like this alot because of Libratus beating a few pros heads up. This actually is pretty laughable. While Libratus did beat the pros, it wasn't real poker. If you read a little more into the story, behind the headlines, there were important rules that certainly hindered the pros.

In order to weed out some of the luck involved with heads-up No-Limit Hold’em, a few special rules were implemented. First, the players and the AI were given 20,000 chips at the start of each hand. Blinds were 50/100. By resetting the chip stacks every hand, players – living or otherwise – had plenty of room to make plays and good runs by a player couldn’t snowball into a big stack versus small stack scenario.

Additionally, hands were mirrored, meaning that pairs of players received reversed hands. For example, if Chou was dealt 2-7 offsuit and Libratus got pocket Queens in one hand, McAulay would be dealt Queens and the AI would get 2-7 in a mirrored hand. This way, hands were distributed evenly (though deal was not predetermined), so neither the humans nor Libratus could benefit from getting a sick run of cards.

Finally, once players were all-in with a call before the river, no more cards were dealt. In these situations, the winning probabilities were calculated and players received a percentage of the pot corresponding to their equity in the hand. One would think that this would mean that there would be fewer chances taken on all-in calls, but at the same time it also meant that nobody could get lucky and suck-out on an all-in.

https://www.pokernewsdaily.com/poker-ai-libratus-soul-crushes-humans-29342/

Sorry, but that is not real poker.

We are so far away from a program that can beat a full ring game (or tournament) that it isn't funny.

 

[twizzybop]

Not to quash such an idea and bring Polk into the conversation
But

In 2015, Polk was selected to play heads-up no limit hold'em against A.I. poker bot Claudico, along with professional poker players Dong Kim, Jason Les and Bjorn Li. Each player was set to play 20,000 hands against Claudico for a team total of 80,000 hands. The human players ended up defeating Claudico for 732,713 chips, with Polk beating the bot for 213,000. The team received a total of $100,000 for the victory.[7]

 

[HennieP]

@karl oakley

We are not as much on opposite sides of this debate as it might appear. I'm not saying a computer can play with the same human intuition and imagination as a human. On this I agree, we're far away. What I am saying is that there are only a few hands at any given time that will be EV+ to play. So if you made a bot that didn't account for human emotion but only ever played EV+ hands it could be done.

Let's say we have such bot. The bot plays an EV+ hand like QQ and raises pre-flop. His human opponent re-raises. The bot re-evaluates based on the human's position and expected opening range and calls. They see a flop and the flop is AKA. The human makes a pot sized bet. The bot again evaluates based on pot odds, EV and the human's likely range and folds on the assumption that the human has something like AA/KK/AK.

The human could have been bluffing of course but our bot doesn't care about that. All he cares about is the profitability of a play and the odds of being ahead or behind.

Now you put this bot against the pro's and tell them its a bot and they will clean it out for sure. The bot can't bluff and the bot can't interpret bluffs. You put this same bot in a online ring game and tell nobody it's a bot? It's gonna be a very profitable little bot.

Edit - I just thought of this as well. How sure are we that the bot that lost to Doug Polk and friends were actually programmed to win? It would be very bad PR for online poker if a bot could defeat the best players in the world don't you think?

 

[lourivalveras]

One of the requires of online Cassinos os the comitment of the players. A bot is off the question and should easily get caught by the online system. Be carefull is Never good enough.

 

[John Turnner]

I dream that poker rooms are banned from using programs. Without them to play interesting and honest.

 

[Jared rich]

Artificial intelligence will for sure change the dynamic but organic understanding will prevail.

 

[cwatt3131]

I think it's important to note that all machines are merely an output of what humans put into them. Yes, they are extremely powerful and mathematically superior in terms of calculations, but they will never be able to beat more than 1 person or win MTT's.

 

[Poker Orifice]

Seems like most in this thread have not followed the news with poker & AI.
Back in 2015 there was a challenge with Claudico taking on 3 of the best players in the world. The humans came out on top (pretty sure most in this thread wouldn't have though).

Another challenge takes place in 2017 and this time the '4' humans all end up as losers & AI (Libratus) comes out ahead by more than $1.7 million.

There are many articles online about this... if you're too lazy to search here is just one https://www.wired.com/2017/01/mystery-ai-just-crushed-best-human-players-poker/

 

[Poker Orifice]

Edit - I just thought of this as well. How sure are we that the bot that lost to Doug Polk and friends were actually programmed to win? It would be very bad PR for online poker if a bot could defeat the best players in the world don't you think?

AI DID beat the best players in the world (Libratus). (you are referring to the challenge that took place in 2015).


... maybe the players were 'programmed' to lose?? LOL

Hint, do a minimal bit of research online and you won't have to guess as much or as often.

 

[HennieP]

AI DID beat the best players in the world (Libratus). (you are referring to the challenge that took place in 2015).


... maybe the players were 'programmed' to lose?? LOL

Hint, do a minimal bit of research online and you won't have to guess as much or as often.

Hint, don't attack people's comments if you have no clue what's going on.

What research was I suppose to do? I did not bring up the argument in the first place, I merely commented on what previous posters stated which was that a bot lost to humans. Following this conversation I merely asked how sure they were the bot was meant to win. So again, please tell me what research this question required on my part?

So, I suggest you do a minimal bit of reading so you won't make an ass of yourself as often.

 

[No1eJoker]

I agree with you, the scary fact is: Bots don't even have to play a perfect strategy. And they don't have to beat the best players.
To make an impact they just have to beat the average player. And there's bad news on that front: We're there already.
For virtually any poker game there already is a bot that plays better than the average, decent human player. So while poker in general might not yet be solved in a theoretical sense, it's solved enough for a decent bot to beat a decent player.

 

[cranberry]

The future of mankind in the future will depend on artificial intelligence. Anyway - it is inevitable. Artificial intelligence is rapidly developing and has unlimited potential.
I think, that hardly profitable for someone to destroy on-line poker by introducing artificial intelligence - because the profit from this will be 0.

P.S. The cooperation of people and artificial intelligence can achieve more, than people and artificial intelligence separately. The best poker player today is a person armed with a computer

 

[Jmckay]

A.i. will kill alot more than just poker.