Slashdot Mirror
Poker Program Battles Humans In Vegas
Bridger writes "Poker software called Polaris will play a rematch against human players during the 2008 World Series of Poker in Las Vegas.
Developed by an artificial intelligence group at the University of Alberta in Canada, Polaris will be pitted against several professionals at the Rio Hotel between July 3rd and 6th. 'It's possible, given enough computing power, for computers to play "perfectly," where over a long enough match, the program cannot lose money,"' said associate professor Michael Bowling.'"
When it's bluffing, it blinks twice.
I'd love to see one of the guys slick at handleing cards, slip a couple extra aces into the deck, or the like. Would the program adapt? Draw a laser and call him a no good sack of mostly water?
I Need someone to rebuild a Digitech Digital Delay pedal for me....for me...for me...for me.
If they want to correctly display the advanced AI "poker face": :|
I'm at least as good as this software! I also won't loose any money playing Poker. In fact, I'm better, because I don't loose any money in the short, near, or long term! I just don't gamble!
"It's possible, given enough computing power, for computers to play "perfectly," where over a long enough match, the program cannot lose money,"
Unless the match is infinitely long, that is not true.
Only to find out it's Wii tennis, a very small subset of "tennis"
The statements made regarding this subject apply only to the subset of poker being played, seven-card limit Texas Hold'em.
the program cannot lose money," said associate professor Michael Bowling
Riiight. So why are you an associate professor instead of cleaning up with online poker?
Poker is a zero sum game. Pit two of these 'perfect' players against each other, and one of them will lose money.
Give me Classic Slashdot or give me death!
professional poker is a psychological game. Unless the computer has the feeling of anxiety it will have an edge.
What I find impressive is the fact it lost in the past. It would also be interesting to see what it can do with some sort of lie detector software.
One of our competitors trademarked the term "hypothesis". From now on, we will call them "boneheaded ideas".
"It's possible, given enough computing power, for computers to play "perfectly," where over a long enough match, the program cannot lose money,"
So what happens when you pit two of these against each other?
professional poker is a psychological game. Unless the computer has the feeling of anxiety it will have an edge.
Poke is almost entirely a game of skill, not chance, at professional levels. The average dufus at his weekend poker game will play for luck. Professionals play the other players. A computer has no tells, and can't read them in a human player. The computer therefore has a distince edge against the amateur, and a distinct disadvantage against the pro.
What I find impressive is the fact it lost in the past. It would also be interesting to see what it can do with some sort of lie detector software.
The only lie detector that has any hope of working - as you should know, if you read /. - is a professional poker player.
Best application of AI ever
i don't believe it. he's bluffing
intellectual property law is philosophically incoherent. it is your moral duty to ignore it or sabotage it
I for one welcome our new poker playing robot overlords.
This is akin to the Human Chess Player vs The Computer. Why is this even news?
Need a Russian bride? I have a large supply in a warehouse waiting for you. I offer a great trade-in plan too! Robert H
when I google your phrase about 100%+ slots in Vegas, I get no results whatsoever -- because why the FUCK would a casino keep a machine on the floor that lost money every year (or every five years, or every decade). They know the odds. You put money. You pull. You put money. You pull. Nothing changes whether one person continues for forty years or a hundred thousand people continue through forty years, it looks the same from the casino's perspective for that particular slot machine.
We're not so sure about that, though. Poker is a very complicated game incorporating not only mathematical betting and statistical odds but also the important skills of expectation, observation and learning, psychology and deception, intimidation and subterfuge.
So to break it down:
Expectation - This is a "statistical odds" issue.
Observation and Learning - Yes, and this is a program designed by the Machine Learning group at UoA.
Psychology and Deception - Deception is simply varying your play by making it probabilistic rather than static. Psychology is only necessary if you wish to take advantage of your opponents' in order to maximize your own play, it has nothing to do with a game theoretical approach. Also, the computer has no such psychological weaknesses.
Intimidation and Subterfuge - Subterfuge basically goes back to deception again. Intimidation in the world of online poker is essentially proper bet sizing, timing, and position.
An AI player has very little downside here, except for the fact that hold'em is a very computationally intensive game. The UoA team has spent the last decade developing better algorithms and game theoretic approximations to combat this.
All that said, they are still a long way from being able to sign their laptop up for the main event at the WSOP, or sit it down at the Bellagio's high stakes table.
Also, the competition is in July (i.e. right now), not June like the article says.
Polaris will be pitted against several professionals at the Rio Hotel between June 3rd and 6th
Uhm is this date correct?
If it is give me the results already if it isn't the article was wrong and the submitter copied the error.
http://poker.cs.ualberta.ca/man-machine/
First match was a draw.
It is possible for a casino to have 100%+ machines, but only if the jackpot is "progressive" and only at times when it has progressed enough.
Put two of these in a room and see who wins.
I dont know about over 100%, but I know some of the low denomination machines pay off really well. I guess the idea is you think winning is so easy, you might as well go to a larger denomination machine.
:x
Most 100%+ slots I've seen, in Vegas, stipulate that you only get 100% of your money back, "with perfect play". Which would mean the majority of people would still loose plenty of money. Besides, even if you did double your money on a 106% slot you'd probably blow it all on craps five seconds later anyway.
If the machine "loses" (assuming 100% utilization) less than $4/hour on average, they almost certainly come out ahead on amenities/drinks; family members and friends playing other games; people getting bored of the low-payoff slots and losing money on other games; etc. Slots are there partly to keep "non-gamblers" busy pulling a lever, while their acquaintances piss away larger sums.
Once the machine gives away around minimum wage or higher, you might start getting crazies and obsessives working it.
"They were pure niggers." – Noam Chomsky
and thus, if the human is good enough, they might be able to pick out patterns in the decisions the computer makes and exploit them. The computer can't learn how to adapt to each opponenent nearly as well as the human players can. I play hold 'em all the time and I can tell you if you can't keep up with someone's changes in strategy you're going to lose. Also, if you play exactly the same every hand, everyone else on the table will pick up on it, and own you.
i have a roll of electrical tape.
Playing perfectly = keeping 100% of your money in your wallet
Computers play 'perfect' poker, i.e. the computer looks at the cards on the table, the cards in it's hand, and the bet's it opponent has made, and figures out if calling the bet, or raising the bet is appropriate statistically, based on the chance of it's payout vs. it's chance of getting 'the winning hand'. In the end, when two perfect players play against each other, they both get the same amount of information, and the difference between their scores will be determined by the random-chance of the deck. The critical flaw of the 'perfect play' system is too rigid adherance to the system. By betting on a predictable system, they grant their opponent an insight into their hands, and therefore, more information. A bet from a perfect player communicates more than just the new odd's on the pot for his opponent, but also allows his opponent to have a better read on what hand the perfect player /has/. More information=better play.
Of course, this is seven card limit texas hold'em, a very formulaic and restricted betting structure will severly limit players abillity to exploit their informational advantage.
Professional poker requires an adaptive system which compensates for the meta-game of people's betting. It's possible their AI does this, but simply playing 'perfect poker' is no acheivment, any grade 12 math student with a bit of practice and a head for odds can play /near/ perfect (i.e. perfect enough that in the limited set of hands they see they have a good chance of making no mistakes). If beating professionals was as easy as not having any tells, and being able to perform reasonable amounts of math, the final table at poker tournaments would be alot more diverse.
100%+ "pure" slot games aren't exactly common (because as you say, they will invariably lose you money). What is very common is 100%+ return poker machines or games with similar levels of user input, where the machines pay out more than 100% if you play 'perfectly', forever. Of course perfect play is often unintuitive and involves things like taking the safe bet rather than higher payout options - not something most people in Vegas are renowned for.
I'm not sure why anyone thinks this is such a feat. In LIMIT hold'em, bluffing, psychological aspects, and implied odds are diminished to the point of meaning next to nothing. It is almost a purely computational game. So, yes, a computer can play technically "perfect".
There are already poker "bots" out there that will play pretty much perfectly when it comes to Limit Hold'em. I'm not sure why this is so different.
I want to see this team of academics write some code that will beat a human at *No-Limit* Hold'em. Or maybe *Pot-Limit* Omaha. NEVER going to happen.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing. The machine would turn into a professional folding station that only plays AA, KK, or AK. Guess what? That strategy isn't winning any games or any period of time in a no-limit or pot-limit world.
AK? That's Called an Anna Kournikova...She looks good, but never wins ;-)
'It's possible, given enough computing power, for computers to play "perfectly," where over a long enough match, the program cannot lose money," said associate professor Michael Bowling.'"
Impossible.
Here's why. Put four of these at the same table with no humans.
Someone will *have* to be the loser.
Weaselmancer
rediculous.
We're not so sure about that, though. Poker is a very complicated game incorporating not only mathematical betting and statistical odds but also the important skills of expectation, observation and learning, psychology and deception, intimidation and subterfuge.
Lol @ poker players who don't understand math
In my head, this computer's monitor is sitting at the table using a likeness of the dealer from that old Intellivision poker game.
Actually, video poker machines with +100% payout are harder to come by than they used to be. There are actually websites out there where people track which machines have the best payouts. These are professionals who make their living (or a good portion of it) gaming the system.
Some of you are probably asking, "Why the FSCK would they make machines that payout more than they take in???"
It's mostly what the parent said. Most people don't play perfectly and therefore lose money playing. They make the payout system complicated enough so that very few people can actually play well enough to consistently win. Then they pull in the people who think they are good enough, so it works in their favor!
>Java has, and will always be slower then assembly.
It all depends on who writes the code.
You are fundamentally mis-understanding the importance of variance. Try to think of it in the REALLY big picture of the long term. If I offer you a game you spin a wheel where the bet is $1000, and 999 out of 1000 times you will lose everything, and 1 time out of 1000 you will get paid $100,001 dollars, my game is now paying out OVER 100%. Now this is a simplified example, but almost NOBODY is willing to take those swings of variance for such a ridiculously small edge. The casino has almost unlimited amounts of money, and can ALWAYS bear the swings. This is also besides the point that there are MAYBE only a handful of them on the casino floor and they are always filled in the WAY back of the casino, ensuring that for giving up a tiny edge, hundreds of gamblers will come in and try, and give up.
Over the short term, one wins by chance.
Over the long term, they tie.
I think it's more likely that casinos have 0% slots. The slot machine never ever pays out. This of course doesn't dissuade your average gambler.
bite my glorious golden ass.
You Bluff.
You have been eaten by a Grue.
You die.
Moved to http://soylentnews.org/. You are invited to join us too!
+1 informative if i had points
I assumed that it was no-limit hold'em because that's what most people play, and all the poker shows on TV are predominantly no limit based, it's just more exciting.
Orbis terrarum est non altus satis
'It's possible, given enough computing power, for computers to play "perfectly," where over a long enough match, the program cannot lose money," said associate professor Michael Bowling.'" I don't think this statment is accurate. In poker there is no such thing as playing perfectly, you are in the end always left to chance to some extent. How could the above statement be true if it were pitted against several copies of itself? There is no way they could all play perfectly, if one is in the lead it means all the others are losers. If one of the machines were to win a few lucky hands at the start... it already has an advantage of having a bigger pot, at this point it can then maximize its odds using some strategy, however this is not guaranteed and again to get to this point it still up to chance.
The Pros will still win, because the Pros have more information than the computer, if the computer is always playing optimally.
That's what makes the Pros pros. People who play mathematically optimum poker lose, because they are ignoring the information that is important: What cards does the opponent have?
In fact, I'd wager (hah!) that a computer playing mathematically optimum poker is at a disadvantage, as it makes it much easier for a Pro to determine what cards the computer has.
paintball
I play good enough Poker that I can take down money off of regulars at online poker sites(I've trained for thousands of hours), and I've toyed with programming poker bots. Even with a program that plays "perfect" Poker, it isn't guaranteed to even finish in the money. Look at this obvious situation: Poker bot gets "AA" in the beginning of the tournament, Joe Blow gets "KK", they both raise and re-raise each other until they're both all in. The pokerbot made the right play, and Joe Blow can't be faulted much either. The board is 2h 5d 7d Kc Qc. The kings beat out the AA. The perfect Poker bot lost. The only way to know that the Poker Bot made the right play was that a Poker veteran confirmed it.
When you play advanced Poker, you can get into situations where you're playing the odds that your opponent does not have a hand. In this style of Poker, you're sometimes playing with weaker hands than Poker books suggest you start with. You can make the right play in advanced Poker, but other Poker vets wouldn't agree with you because they're not advanced enough to understand your play.
In advanced poker it is sometimes hard to figure out who made the right play because there is no mathematical basis for it. Are you going to compute someone's algorithm in their brain to determine what strength of hand their betting pattern is on? People can do this, but there is no hard and fast rules for getting reads. A computer can get reads better than a human because it can memorize everything perfectly, but to code something that handled all those reads would take a lot of time... And who is gonna code that? A professional poker player? The programmer is probably going to be a random dude.
So I'd say it isn't trivial to program a "perfect" Poker player. How are you going to judge which poker bot is the best? Are you going to pit bots against each other over millions of games? Are you going to pit bots vs humans? Anyone can claim they have a perfect poker bot, but I sincerely doubt it. And if you leak your source code, anyone who knows they're playing against your bot will have a huge advantage.
God spoke to me.
If I was playing a computer on the first hand I would go ALL in (and do it blind). ...program would hopefully calculate my SIZE_OF_BALLS() variable as an out of bounds condition and give up. If that didn't work at least I'd be done and could go back to drinkin' in the casino bar.
While it is true that limit holdem is mostly computational, the way pros extract value (read: money) from the lesser pros is by adapting. Most reasonable pros essentially play perfectly when it comes to pot odds, game theory, card value, position, stack sizes, etc. However, what separates them from the top pros is their ability to factor in additional information from the previous hands with that person (or machine). Will the player (or machine) play more aggressively if it has been steamrolling the opposition? If a pro believes that to be true, she will respect aggressive players bets less.
Also from TFA, "It's possible, given enough computing power, for computers to play 'perfectly,' where over a long enough match, the program cannot lose money,"
While this is true, a "long enough match" would be VERY VERY long - hundreds of thousands if not millions of hands. For pros over a couple thousand hands of limit holdem against good opposition, winning or losing is primarily dictated by variance.
By well titted cocktail waitresses handing out free drinks.
Engineering is the art of compromise.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing. The machine would turn into a professional folding station that only plays AA, KK, or AK.
why can't a computer bluff?
... the house needed any further advantages to ensure that it always won.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing. The machine would turn into a professional folding station that only plays AA, KK, or AK. Guess what? That strategy isn't winning any games or any period of time in a no-limit or pot-limit world.
This is where the machine learning comes into play. Each time the human bets half her stack right off the bat, the machine lowers its calling (and raising) standards.
I want to see this team of academics write some code that will beat a human at *No-Limit* Hold'em. Or maybe *Pot-Limit* Omaha. NEVER going to happen.
It's certainly harder, but I wouldn't say never. In some scenarios it might not be too hard to beat an average human, like having the AI buy in short and try to get all in preflop or on the flop.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing.
Maniacs who do that "constantly" are easy to beat, just call with hands that are on average better than theirs. An AI with halfway decent opponent modeling should be able to adjust to that pretty quickly.
How to solve most of our problems: 1.Lots of nuclear plants. 2.Cure aging.
If the machine wins, the maker will get a military contract, then get killed by some hot terminator chick and a crazy woman named Sarah Connor.
Wait... I've seen this before.
You moved your mouse. Please restart Windows for changes to take effect.
I want to see this team of academics write some code that will beat a human at *No-Limit* Hold'em. Or maybe *Pot-Limit* Omaha. NEVER going to happen.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing. The machine would turn into a professional folding station that only plays AA, KK, or AK. Guess what? That strategy isn't winning any games or any period of time in a no-limit or pot-limit world.
This team of academics has written poker bots that play no limit and pot limit poker, and the bots are fairly decent at it. Doing a half stack bet frequently will most likely end up with the professional being busted out.
The same techniques would work for pot limit omaha. The only difference being that pot limit omaha is a bit more combinatorially complex.
LetterRip
What's more, people don't realize that "playing perfectly" is a misnomer. It deals with making statistically correct decisions given the available information, which in poker, is not all of the information.
People familiar with the game understand that perfect play doesn't mean "you always get it right". It means you are playing based on the idea of expected value. This is discussed in the theory of poker.
When you talk about multiway play, you also are overlooking that decision making changes significantly in multiway pots, which is discussed in Morton's theorem.
I'm also at reasonably sure that given time, a pattern of "perfect play" would be found to be exploitable. Since the computer is using incomplete information, it must make assumptions. These assumptions, if anticipated, could be used to influence the computer's decisions.
Since this story is about limit, it's no wonder the computer has so much success. The structure of limit makes it a very straight mathematical game, where you're frequently calling river bets when you are certain you are beaten, simply because statistically it is correct to do so.
This structure is not present on all poker games, and indeed, many of the most popular games are not limit, but pot limit or no limit instead. I'd be interested to see what kind of progress is being made on those games, which strike me as vastly harder problems.
I'm not sure why anyone thinks this is such a feat. In LIMIT hold'em, bluffing, psychological aspects, and implied odds are diminished to the point of meaning next to nothing. It is almost a purely computational game. So, yes, a computer can play technically "perfect".
There still exists a psychological aspect to a limit hold'em game, particularly at higher stakes.
But even at a 3-6 or a 4-8 table, I find a chance for a bluff oh, about once every four or five hours. It's not a big aspect of the game, but it's there. You have to know the players well enough to know who's actually capable of folding, AND get into a situation where they can believe they're beat, but it's possible, and even more important, valuable. Not required to make money at the low limits, mind, but can score a couple extra bucks every now and again.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing. The machine would turn into a professional folding station that only plays AA, KK, or AK. Guess what? That strategy isn't winning any games or any period of time in a no-limit or pot-limit world.
Bah. Even I can write code that can handle that: "move in over the top, with nothing".
That said, a computer's got to walk before it can run. If it can sit down at a 20-40 game somewhere, it's not doing too bad. Once they've figured out how to write a limit bot that can handle that, then they can start working on a no-limit bot.
Given the play we see from supposed "pros" on TV, I fear what the researchers will determine is random play backed by a large bankroll is indeed the optimal strategy in modern no-limit hold-em. :-/
Comment removed based on user account deletion
...a malicious adversity can be introduced through study and analysis. The players will observe the behavior of a program with finite possibilities.
Oblig /. reply
1)Observe
2)Divide
3)???
4)Conquer...Profit?
Of course perfect play is often unintuitive and involves things like taking the safe bet rather than higher payout options - not something most people in Vegas are renowned for.
its actually completely the opposite for most video poker games, such as throwing away a made flush (already a winner) that is almost a straight/royal flush --an example would be like KQJT2 all of clubs, the correct play is to give up the guaranteed win of a flush, and draw for the jackpot hands (royal and straight flush)
I want to see this team of academics write some code that will beat a human at *No-Limit* Hold'em. Or maybe *Pot-Limit* Omaha. NEVER going to happen.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing. The machine would turn into a professional folding station that only plays AA, KK, or AK. Guess what? That strategy isn't winning any games or any period of time in a no-limit or pot-limit world.
What is it about poker that everyone that can play thinks they're great at it.. Please give us more of your poker insight, o' master of poker AI..
Are you really suggesting an AI can't tell when someone is overbetting? That's not even AI, it's just basic probability and statistics.
If you even go to somewhere like Full Tilt Poker you'll likely encounter bots with the ability to pick off people which overbet constantly, and you wouldn't even call that software "AI", just basic rules and guidelines.
// MD_Update(&m,buf,j);
I think it's more likely that casinos have 0% slots.
Casino machines are heavily regulated. Instead, the gambling industry is searching for other ways to screw their clients: http://www.salon.com/news/feature/2008/06/16/gambling_science/
The machine would turn into a professional folding station that only plays AA, KK, or AK. Guess what? That strategy isn't winning any games or any period of time in a no-limit or pot-limit world.
Against a loose player, this strategy is ideal actually.
I want to see this team of academics write some code that will beat a human at *No-Limit* Hold'em. Or maybe *Pot-Limit* Omaha. NEVER going to happen.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing. The machine would turn into a professional folding station that only plays AA, KK, or AK. Guess what? That strategy isn't winning any games or any period of time in a no-limit or pot-limit world.
I love when I read idiotic statements like this. I mean how many times do we have to see lines like "x amout of memory/storage/power will always be enough" before we learn how stupid they are. Pretty much if someone utters phrases containing never or always and doesn't have a mathematical proof revolving around the total amount of information in the universe you can be pretty sure they're full of it or just uninformed.
Now to your point and why it's overstated....you are partially correct, currently they can not do this but it's largely a problem of lack of input and computational power. We all know the computational power will get solved eventually thanks to Mr. Moore's "law". What's left is enough input for a computer to make a read. This likely means facial and voice readers (assuming a table of full players where there is some chatter). At this point the computer would have an advantage because it would give away much less information than it would take in. Additionally it would have perfect recall of behavioral vs. play information over the history of the game. This would allow the computer to play correctly based on various decision distributions using advanced game theory far beyond basic pot odds and expectation customizing these distributions to the opponents as the game progressed and more historical behavioral information became available. Also you won't see the all too common bet size mistakes from the computer.
It's a massive computational problem and currently the computer doesn't have enough inputs to solve it, but to say it will never be able to play No-Limit effectively against a pro is silly. No-limit poker is purely luck in the short term and purely skill in the long term and anything that is purely skill can eventually be solved by a computer with enough input information and computational power. We just don't CURRENTLY have these two prerequisites in place.
I don't care how well such a program is coded... it will absolutely buckle under the pressure of a professional who constantly bets half his stack on nothing. The machine would turn into a professional folding station that only plays AA, KK, or AK. Guess what? That strategy isn't winning any games or any period of time in a no-limit or pot-limit world.
I'm not sure why you think this. The computer is incapable of "buckling under pressure". It either makes a good play or a bad play.
People assume that the computer is always going to be playing based on pot odds, but a well-designed system will take much more than that into account.
So if you deploy a half dozen around a virtual table, the net loses and gains of each would be equal? Otherwise, the net money supply would have to grow?
Lurking at the bottom of the gravity well, getting old
I have a brilliant algorithm for a poker program that can never lose!
User: "Would you like to play a game of poker?"
Computer: "No, thanks."
Given enough computing power, a machine will always be able to calculate the exact probability of every possible hand that its opponent might have, now & into the near future, without being fooled by stuff like bluffs or feints. (It's also possible that it might be able to detecting cheating like slipping an extra card into play when a result occurs which it has calculated to be "impossible".)
The entire point of a bluff/feint is that the cards are random, and calculating the exact probability of every possible hand has nothing to do with the actual hands that your opponents have. "My hand is better than 92% of the possible hands my opponent could have" is just enough to lose all of your chips 8% of the time. As the variance in possible betting increases, the usefulness of probability in long-term play decreases. No-limit poker isn't about the cards, it's about the behavior of your opponents.
"We have to go forth and crush every world view that doesn't believe in tolerance and free speech." - David Brin
NEVER going to happen
I've learned a few things in life and at the top of the list is that people who say this are always wrong.
I'm not saying I'm great at it. Not at all. I'm not great at baseball, either, but I'm a fairly good student of its strategy. Same with poker.
I'm not suggesting an AI can't tell when someone is overbetting. "Overbetting" is a simple computation. Detecting an overbet is simply looking at the bet compared to the pot-odds and the potential hands of the bettor.
I *am* suggesting that an AI is not going to be able to determine if a professional player is overbetting in order to bluff... or in order to incite a call.
Let's be clear. Computer programs do not "learn". All they do is examine history and then make some computational decision based on historical probability.
Any professional poker player will absolutely tear that strategy apart. Because poker pros are all about playing the weaknesses of their opponent.
If you honestly believe that a group of Doctors in this field would spend their time making a simplistic little bot that would degenerate into only playing AA, KK or AK, then it's obvious that the real problem they are working on here is completely over your head.
"Fairly decent"? You mean they lose and are close to break-even, right? They are still losers. There isn't a no-limit/pot-limit bot in existence that wins over the long run. Let's be clear about that point.
And most of these bots are run at micro-limits. You won't see these no-limit/pot-limits bots running at $5-$10, $20-40, or $50-$100. Why? Because these bots suck.
If you know of a no-limit/pot-limit bot that consistently wins over the long haul, please post a link.
Oh... a computer can bluff. But let's qualify that. A computer can bluff BADLY.
A computer bluff would be the equivalent of some donkey randomly throwing a massive bet out there hoping for a bunch of folds.
But a computer isn't going to know that "John Doe" across the table just took a huge bad beat, is tired, is close to the end of his bankroll... and then bluff based on that information.
A computer simply isn't going to know all of the information and nuances necessary in order to execute a truly good bluff.
Hmm... Lots of misspellings of the word "lose" in the same location. This must be a thread about gambling!
So! Some guys have got or with additional investments can get a bot whose purpose is to not lose money at a card game over time? I'm going to go out on a limb and assume it would also accept gaining money over time as acceptable.
One should not too quickly dispense with important concepts such as "aggression" and "extracting maximum value from a hand" using outside data including "physical tells" and "the metagame".
I'm no bigshot, but I will asert with confidence that this new dealiebob has a standing invitation to visit Bobby's Room and risk some actual treasure in the form of green bricks w/ rubber bands.
Even lesser men than the likes of Doyle Brunson, Barry Greenstein, and Phil Ivey would line up to fly y'all down and take your developers out to Prime before or after the big game.
If you think no-limit poker is simply a massive computational problem, then you don't know anything about poker.
Anything that is purely skill can eventually be solved by a computer? Are you nuts?
Anything that is purely MATHEMATICAL skill can eventually be solved by a computer. But "skill" is not necessarily computational. Especially in the no-limit/pot-limit case.
Seriously, go read some books around no-limit and pot-limit gambling. You'll quickly find that the "edge" in these games isn't about the numbers.
This group of doctors is solving a problem that might yield a small amount of additional expected value over commercially available limit poker bots.
This group of doctors IS NOT taking on the problem of sitting a computer down at a table of 6 or 7 no-limit/pot-limit professionals -- and having the computer win. Why? Because they can't solve it.
A computer sitting down at a no-limit table, surrounded by pros, would be pounded into playing only the highest value hands. Because it would "learn" that the only way it could predictably win is to play those hands.
Hell, the best no-limit computer/bot out there would be one that doesn't even consider its own cards. The best no-limit bot would simply pick a hand it wanted to represent at the beginning of each hand... and then play each hand as if it actually had those cards.
At least then it would be closer to a professional level of play.
Given a long enough game, there will be a plausible hand which seems like a winner which is beaten by another hand and results in a huge loss.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
Betting $1000, 1000 times will cost you 1 million. Paying out 100,001 dollars every 1000 games actually only a shade over 10%.
Well, that's the end of online poker, everyone could be using a computer doing it for them instead. Only face to face poker is reliable.
i happen to live in Edmonton Alberta where this program was created and talked with one of the programmers, originally, the program had a limited number of actions, that is to say it treated similar hands the same way, ACE and TWO were treated the same as ACE and THREE, i was told by this go around the program would treat each hand individually, and just so you know the computer does bluff sometimes,
I mean, I know there tends to be some lag on Slashdot, but seriously!
Oblig. Wargames misquote: "The only way to win is not to play".
I don't think I'd want an opponent called Polaris, I'd be afraid of a missile up my ass if I beat it.
To have a right to do a thing is not at all the same as to be right in doing it
The grandparent obviously meant a payout of $1,000,001.
I can see computers doing well at this fairly soon, as the options are much more limited and (comparatively) easily enumerated.
The pros they have lined up didn't ring any bells with me, but last year Phil Laak opened up a can of whup-ass on the machine. It will be interesting to see how it goes this time around.
Incidentally, I think pot-limit is a more skillful game in tournament format; it takes away one of the few obvious moves fish have against pros. In a ring game, though, NLTH is definitely 'The Cadillac of Poker'.
[ ]Half Empty [ ]Half Full [x]Twice as big as it needs to be
The argument isn't that it isn't harder it's that the computer can eventually notice the same things people do.
I don't see why a computer can't randomly keep a few hands here when statistically it should fold. That way, whoever the computer is playing has no clue when it is a bluff and when it is not. The computer would have no tell. It's brilliant!
Having worked in a casino running numbers (Crystal Reports makes baby jesus cry) and working in marketing (who needs a soul anyway?) I can tell you that the slots themselves make money, and when you do have a big payout, it serves as advertising. When you have a gigantic payout, it's partly covered by the promotion (the biggest payouts tend to be part of some kind of network game) and people come from other markets to visit your casino. In California slots sell booze (it's actually illegal to give away alcohol in California) but in Vega$ the slots make money and some establishments give you some of it back in booze. If you can stand well booze, though, you probably get the best value for your money at the penny slots... and people at dollar slots aren't there for the free drinks. They're there because they have too much money and want to give it away.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
go T go
That's only because you're an imbecile who doesn't know how to play, as yet, there is no computer that consistently beat a skilled human player. RTFA to reiterate that if you're too stupid to realize it by now. I've put m money up, and have no takers. That says more than any argument you have, and frankly, I'll put my money up to you too, get whatever program you like and put up a decent stake, and we'll get to it.
Don't assume everyone sucks at poker just because you suck at poker.
I genuinely believe you're stupid enough to think a solvable game is equivalent to poker. I really do think you're that stupid.
I also notice that you failed to address every single false accusation that I refuted, and are pretending you weren't lying in your previous post.
You clearly have no idea how poker works, and are willing to say whatever you need to in order to avoid admitting that. Unfortunately, your statements are a shibboleth that you can't undo, and they clearly paint you as nothing more that a novice player with no skill or knowledge of the game at all.
it's actually illegal to give away alcohol in California
How does Napa work into that?
A robot chess match is one thing . . . but a poker match is something else . . . I don't think it really works simply because it takes away the whole bluffing element of the game. The machine can only play using probabilities so it can only play its own hand. However, poker is about more than playing your own hand.