In New AI Benchmark, Computer Takes On Four Top Professional Poker Players 89
HughPickens.com writes: Stephen Jordan reports at the National Monitor that four of the world's greatest poker players are going into battle against a computer program that researchers are calling Claudico in the "Brains Vs. Artificial Intelligence" competition at Rivers Casino in Pittsburgh. Claudico, the first machine program to play heads-up no-limit Texas Hold'em against top human players, will play nearly 20,000 hands with each human poker player over the next two weeks. "Poker is now a benchmark for artificial intelligence research, just as chess once was. It's a game of exceeding complexity that requires a machine to make decisions based on incomplete and often misleading information, thanks to bluffing, slow play and other decoys," says Tuomas Sandholm, developer of the program. "And to win, the machine has to out-smart its human opponents." In total, that will be 1,500 hands played per day until May 8, with just one day off to allow the real-life players to rest.
An earlier version of the software called Tartanian 7 (PDF) was successful in winning the heads-up, no-limit Texas Hold'em category against other computers in July, but Sandholm says that does not necessarily mean it will be able to defeat a human in the complex game. "I think it's a 50-50 proposition," says Sandholm. "My strategy will change more so than when playing against human players," says competitor Doug Polk, widely considered the world's best player, with total live tournament earnings of more than $3.6 million. "I think there will be less hand reading so to speak, and less mind games. In some ways I think it will be nice as I can focus on playing a more pure game, and not have to worry about if he thinks that I think, etc."
An earlier version of the software called Tartanian 7 (PDF) was successful in winning the heads-up, no-limit Texas Hold'em category against other computers in July, but Sandholm says that does not necessarily mean it will be able to defeat a human in the complex game. "I think it's a 50-50 proposition," says Sandholm. "My strategy will change more so than when playing against human players," says competitor Doug Polk, widely considered the world's best player, with total live tournament earnings of more than $3.6 million. "I think there will be less hand reading so to speak, and less mind games. In some ways I think it will be nice as I can focus on playing a more pure game, and not have to worry about if he thinks that I think, etc."
Count cards (Score:1)
Re: (Score:3, Informative)
Will the computer be prohibited from counting cards? Humans may bluff, but they cannot fake statistics.
If you are not counting cards, you are playing poker wrong. This is not blackjack with multiple decks.
Re: (Score:3, Informative)
In Texas Hold 'em, you see your own two cards and the up-to-5 community cards each hand, and the deck is shuffled between hands. Everyone knows what cards have been seen and what have not all the time without any card-counting skill.
Re: Count cards (Score:1)
Counting cards will not help
Re: (Score:1)
Re: Count cards (Score:4, Interesting)
The guy that you linked to thinks that knowing how many outs you have is "card counting" -- no. you also apparently think so, which means that you cannot possibly have anything to add on this subject (and your ignorance on this subject is not a secret to you, so why are you pretending?)
Re: (Score:3)
That site doesn't understand what card counting is. When you count cards, it changes how much you bet on a hand you haven't seen yet. You increase your bet (or enter the game altogether) only when the player odds are higher than normal.
Counting "outs", the number of cards in the desk that will improve your hand, is not what's called card counting in casino games.
Re: (Score:3, Informative)
Your replies aren't quite making this clear for some reason, but counting cards is a strategy for blackjack, not poker. In poker, the deck is shuffled after each hand, so there is nothing to count. Of course, using whatever information is available to estimate the value of the hands' of your opponents is part of the basic strategy of poker that every sane player, including this computer, uses.
Re: (Score:3)
I was also confused how this could work in holdem. After some googling it appears that "counting" in poker only refers to certain stud games where hands ahead of you are exposed.
In holdem the only (?) additional info you could have over a competitor is if someone in early position folded and revealed their cards (obviously uncommon). That's info that the small blind wouldn't have had, and could be very useful if you're in late position, especially on the button. But imo that's not really "counting", just
Re: (Score:1)
With a computer, its more than "counting cards" - a computer can remember every single card played and compute probabilities far more precisely than a human ever could. So the question is - how much real information can people *really* get from other players betting behavior? Is it really enough to offset the advantage from simply being able to fully compute the probabilities? Furthermore, if the computer model includes factoring in information revealed by bidding and uses that to obfuscate its betting b
Re: (Score:2)
Maybe you should work out that there not all card games are blackjack. There are no revealed cards that are then hidden before a reshuffle in a texas hold em game, so there is nothing to count. Seven card stud, as one example, would be a poker game in which there are such cards - but such a small number that all the player's (well or the ones who can actually play) remember them.
That's Great and All (Score:3)
But can it make me a sandwich?
I think it's important to note that while we're good at making AI good at one thing, we're still a very long way from making it good at general skills.
Re: (Score:3)
But can it make me a sandwich?
I doubt that an AI program will be able to do that anytime soon. However, it's well within the capability of today's genie:
Human: Genie, make me a sandwich.
(Flash! smoke dissipates to reveal sandwich in place of human)
Genie: OK, you're a sandwich.
Re: (Score:2)
Re: (Score:3)
The article claims a machine is more sanitary. But think about it, food gets everywhere, is a robot going to clean the machine? or a human? It's only more sanitary if it is cleaned thoroughly and regularly otherwise bacteria and mold could run rife.
Re: (Score:1)
It's retarded to even think there is some general skill in animals or anything.
Either you program sandwich making directly or you program its artificial desire to eat and the ability to taste, rank, and make food out of raw materials, and give it some time to learn to make an eatable sandwich.
It's supposed to learn and do things the way people find best. So it's just a matter of time before we deprecate most humans since we cannot improve our thought process or upgrade our brain speed.
Re: (Score:2)
Yes,
http://money.cnn.com/2015/04/2... [cnn.com]
But would you trust a chef who won't eat their own food?
Re: (Score:1)
Given that a robot doesn't go to the bathroom, in some respects the robot is to be trusted more.
Bert
Re: (Score:2, Redundant)
This.
Why would I want an AI that plays games? I want an AI that takes the boring tasks off me so I can play games!
Re: (Score:2)
You want an AI that can make you a lot of money playing games like poker so that you can play the games you want to. Just as long as not a lot of other people have the AI.
Re: (Score:2)
So ... you enter it into a competition where exactly that is being tested? I'd rather let that AI play in some online casinos...
Wait... who says that's not already happening?
Re: (Score:2)
Poker is a game of chance, straight probability calculations nothing more nothing less. Beyond that it is about reading the players and that has nothing to do with poker. So I am 'ass'uming they are inserting butt plug stress monitoring probes and using that to associate stress with betting styles and to be able to assess when a players arse puckers with relation to the quality of their cards and the bets they make.
Re: (Score:2)
We'll have to see if sudo is installed.
Mind games (Score:5, Insightful)
not have to worry about if he thinks that I think, etc.
I think he's wrong on this. A computer would still need to consider what his opponent thinks he holds and raise accordingly.
Re: (Score:2)
Re:Mind games (Score:4, Interesting)
I think he's wrong on this. A computer would still need to consider what his opponent thinks he holds and raise accordingly.
Isn't necessary for chess... the top competitive chess programs (like the foss stockfish...) are not the best suited to beating humans... they still beat humans repeatedly, without mercy, game after game after game. Even the world (human) chess champion (Magnus Carlsen) admits that playing one of these engines is like repeatedly ramming your head into a wall.
Re: Mind games (Score:2)
That's because chess and poker are different. In chess all players know the complete state and the best move is to make the beat move possible, regardless of what your opponent may be thinking.
This is why poker is harder. Chess isn't an AI problem,because it does not need to learn about the nature of your play. The engine need merely take the board and solve for the best move. But a good poker AI must find a way to infer your thinking patterns else it is just playing an odds game and will dump excess inf
Re: (Score:2)
In chess all players know the complete state and the best move is to make the beat move possible, regardless of what your opponent may be thinking.
For a highly restricted definition of "best", sure...
Chess isn't an AI problem,because it does not need to learn about the nature of your play.
It does if it wants a maximal score in a tournament...
You have defined "best" to mean "best against this opponent" in poker, but have arbitrarily defined "best" to not mean the same thing in chess.... there is no justification for using separate definitions here.
Re: (Score:2)
No -- There is an absolute best way to win in chess, from every possible position, and it can be calculated. The human opponent can only screw it up for themselves. (yes, I am aware that the 100% absolutely perfect chess program has not yet been written, that will win when given a time limit, but given unlimited time, a computer will always win at chess. The only reason computers have not already done this is because of time limits. It's the same reason all casinos have betting limits -- in a game wit
Re: (Score:3)
This doesnt define what the "best move" is if at least one player is not one of these "optimal players."
Here then is the problem: The grandparant learned about minimax and now think that he is an expert on chess engines, but hasn't actually put enough thought into it to even pretend to be an expert. He injects long series of assumptions into his arguments in order to reinforce his reliance on knowing what minimax is as being the fo
Re: (Score:2)
No -- There is an absolute best way to win in chess, from every possible position, and it can be calculated.
Wrong. If every move leads to a draw in minimax, that does not mean that ever move is equal. Your thinking is extremely shallow on this subject.
Your sort of thinking is the same reason that the top chess engines sometimes suicide against its opponent because while the top engine sees that it is lost (and thus immediately commences to delay the loss as long as possible by sacrificing every piece that it can) its opponent does not see it.
Its straight up suicide, and now you please explain how straight up
Re: (Score:2)
not have to worry about if he thinks that I think, etc.
I think he's wrong on this. A computer would still need to consider what his opponent thinks he holds and raise accordingly.
Actually, in heads-up play using GTO strategy, this is not necessarily true.
AI has great chances (Score:1)
Poker games take time (hours), people grow tired, computers don't.
People struggle at memorizing chances, taking shortcuts, computers have exact picture talking into account every single bit.
All one needs is behavior that is random enough, for human players not to guess if computer is bluffing.
Then, of course, there is luck factor, so results will fluctuate quite a bit.
Re: (Score:1)
This is a good point. Computers have no emotion, either. Even the best human players are affected to some small degree by their emotions, especially when they are tired.
Not much of an issue in Hold 'em. Good players can handle those odds with little effort.
Re: (Score:2)
People getting emotional isn't always an advantage for the computer; it's part of the complexity of playing against them. Some people will rampage where their aggressiveness ramps up after losing a hand due to luck (a bad beat). Some will bet less aggressively because they're stinging from the loss and worried about their stake. You can't just model the human opponent's patterns and expect them to be consistent.
Re: (Score:3)
People struggle at memorizing chances, taking shortcuts, computers have exact picture talking into account every single bit.
Memorizing chances isn't very important in no-limit. A rough estimate is all you need because other factors will completely dominate whatever error exists in your estimate. When the implied odds can vary between ~1:1 and 100:1, the second or third digit of your estimate of the chances of making a winning hand (for instance, ~2.5:1 against making a flush) is drowned out.
In car analogy terms, its like worrying about if insurance will cover the broken taillight after your car has been t-boned at an intersec
Re: (Score:3)
It's my impression that pro players often get amateur players through bet sizing, if your call/fold response doesn't match the equity of your hand they'll pretty easily see that they can milk you for value or push you into folding. Or that the amateurs are bad at getting the maximum value out of their good hands because they give the pros easy call/fold odds. Of course there's a lot more to bet sizing than your own two cards, but you can't bluff properly without having a pretty good clue about what you repr
Less Mind Games? (Score:4, Insightful)
I think there will be less hand reading so to speak, and less mind games.
Isn't that what makes the game so interesting? Any good card game involves messing with your opponent's head.
Re: (Score:2)
Re:A Strange Game (Score:5, Interesting)
In fact poker has rules (antes, blinds) where the whole point is to make not playing a losing move.
In a global thermonuclear war, it would be like making unused bombs self detonate in one's country after some time.
Texas Hold'em is not complex (Score:3, Informative)
Re: (Score:1)
Jesus Christ you are a nitwit. It's not the game that's the challenge, its beating your opponent, in this case, some of the best players in the world.
God damn this place has gone downhill. Go away, 3,000,000+ ID boy.
Re: (Score:2)
Can I play you for money. Whatever stakes you want (please be a billionaire, please be a billionaire)
Re: (Score:2)
Re: (Score:1)
Re: (Score:1)
The concern isn't about making good decisions, but about the behavior and use of the system's capabilities and emergent properties.
Re: (Score:2)
Re: (Score:2)
You want AI that's capable of making good decisions even when the information is incomplete.
World's best? (Score:2, Informative)
What, are you kidding me? Bluff has him ranked as number 36 [bluff.com]. That's pretty good, but a far cry from "widely considered the world's best."
Re: (Score:2)
I imagine Polk is some very average player who got roped into this project and so, to give the project credibility, they trump him up into the stratosphere.
Re: (Score:2)
Oh surely.
Doug Polk's Global Poker Index ranking:
Ranking Position National rank Total score Highest
GPI rank #164 for 1 weeks #85 1,942.19 pts #77 [globalpokerindex.com]
PoY 2015 rank #1228 for 1 weeks #471 254.16 pts #353 [globalpokerindex.com]
Re: (Score:2)
I paid attention pretty thoroughly to offline poker, from '02 to about '10... and I've only ever heard of him as a guy that writes articles for Bluff. He's certainly not a name anyone I know would say is #1, when there's players out there who churn through (lose-win) more than this guy has made in his entire life, on a weekly basis.
Heads-up No Limit cash game, not tournaments (Score:3)
He really is among the best heads up no limit players though.
Tell (Score:2)
Unfortunately the AI has a tell, his hard-disk lights flash when he's bluffing.
Re: (Score:2)
If someone cheats ... (Score:2)
Crappy, sensationalist reporting. (Score:5, Informative)
TFS refers to TFA which refers to another TFA, and all of them are pathetically written. Here's a link at CMU discussing the competition [cmu.edu]. This is the second link in TFS, but it's not clear that all of the other links in the first paragraph are just trash.
In any case, a couple of points and/or musings:
As a final note: may I please encourage submitters and/or our illustrious editors to not fluff up submissions with links to crappy articles that miss most of the important points? Just the source link would have been enough - it's a good article with real information written in actual English.
Re: (Score:2)
Heads-up or even 4-handed, you're going to clear 100 hands an hour on a single table.
My guess. (Score:2)
Online poker .com would pay BIG (Score:2)
This would be very profitable. If I were an evil asshole with no ethics I would love to buy or rent such software out and put in a fake virtual player for each session. No way I would ever lose money unless through very odd anomalies. Oh hell who am I kidding this is how half of Las Vegas works where they hire statisticians to favor the house as much as possible.
You are dumb to ever think you can win and are just smarter than the other guys
Re: (Score:2)
... this shows how little you understand of poker. Primarily, that poker is a game played between people, and it does not involve the house. The house makes some money for providing the service of dealing the cards, but it is not involved in the game of poker.
Just playing the odds is an extremely exploitable strategy.
Doesn' t the computer have a huge advantage? (Score:2)
Re: (Score:2)
I'm sure you could run a side-channel attack on the computer for tells, and I'm also sure the computer could be fitted with a camera and appropriate algorithms to read your heart rate, blood perfusion rate, respiration rate, rate of sweat production, etc, for information about your general level of anxiety, surprise, etc.
Re: (Score:2)
And even then, all you can do is throw models at it and strategies to attempt to use on those models.