Slashdot Mirror

← Back to Stories (view on slashdot.org)

Artificial Intelligence in Poker

Posted by michael on from the deep-blue-calls-your-bluff dept.

Markian Hlynka writes "The University of Alberta's research into Poker AI is featured in this New York Times article. There is also detailed discussion of the game of Poker, and the 'new breed' of players who have honed their abilities online. See the U of A's poker project for more information."

17 of 279 comments (clear)

  1. Poker AI? riight... by KDan · · Score: 5, Interesting

    Poker is not a card game, it's a people game (aka don't play the cards, play the people). It's all about bluffing and reading other people's bluffs. I'm baffled that people even bother playing poker on the internet. Even with webcams the game wouldn't be the same at all.

    Daniel

    --
    Carpe Diem
    1. Re:Poker AI? riight... by AssFace · · Score: 5, Interesting

      reading the article is too much I see.

      it references that and points out how much of "reading the player" is overhyped and easily faked out.

      whereas the real information is in the trade at hand - the exchange of money. watching the bets and the amounts in them at varying spots in the game.

      I have a few friends that have won online tournaments and they approach it from a very mathematical point of view. They do very well in person or over the net.
      Using the "read" approach, unless the read is of the play on the table, is only going to work with people that aren't aware of the read and therefore not faking the attributes.

      I personally prefer to look for the security holes in the online software :)
      (There was a famous one in '96 or so where the system was using the random function built in - I think in Turbo Pascal IIRC - they had it exposed by posting their random code on the net to prove that they were being fair. A consulting firm then exploited that to show that they only needed to see one or two cards beyond what was in their card to then show what everyone else in the game was holding... there is much higher security in it all these days, and better/smarter programming).

      Another firend in college found a site that had a hole, not in the security, but in the method at which they gave out tokens - as long as you kept playing, there was a reward of some number of tokens as an incentive to keep you playing.
      He then ran some numbers and proved that with that, they were open to an exploit of the Martingale system. He ran it on them for a good amount of time and it failed - he basically proved that their code was cheating on the inside.
      He called them on it and after a few heated e-mails, got all his money back and was banned from the site.
      I could go on and on - but that is going off topic.

      --

      There are some odd things afoot now, in the Villa Straylight.
    2. Re:Poker AI? riight... by Alaric42 · · Score: 2, Interesting

      I would assume that the free tokens given out over time ended up being an infinite supply of money as long as you stretched out the hand length or somesuch.

    3. Re:Poker AI? riight... by wmajik · · Score: 3, Interesting

      Not a personal attack, but I'm assuming you are not a poker player.

      At heart, poker is a human game where your ability in the mental martial arts dictate your ability to succeed. That being said however, this purity usually only shows at the highest levels of poker, heads-up (1v1) no limit poker; as showcased in the World Series of Poker (WSOP).

      In the average poker game at the local casino or with your buddies on Thursday night, the 'mental game' usually falls along the wayside along with the beer bottles. A main caveat of playing perfect poker is to play as if you could see what everyone else at the table is holding. The better you are able to put a man to his hand (figure out what his/her cards are) the better your play becomes, theoretically speaking. I say theoretically because like it says in the article: "Game-theory models usually assume that every player uses the best possible strategy, something that rarely if ever happens with humans."

      Many times, you can have a table filled with players who have no idea what they are doing and betting blindly with an ace high. In this scenario, against a good player holding ace high, you would know to re-raise him even if you were holding garbage, as ace high is not a strong enough hand to call with. Against a weak player, they may even re-raise you back not realizing that they are making a terrible play according to game theory, BUT, a play that happens to actually work in reality. Only in poker can you tell someone that the only reason they won was because they played terrible :)

      Anyhow, my example was meant to illustrate that in poker, sometimes you play the person and other times you play for the best hand. Being able to do both and knowing when to mix it up is what separates a true professional from the average player.

      Then again, talking is very theoretical too for that matter. But then you have to realize there are many people who make a profit playing online (myself included), so don't take my word for it, look at my stack of chips :grin:.

      If anyone is interested, I play at Party Poker.

      Good place for beginners and they'll add $25 to your account if you sign up with the code MAGIC.

    4. Re:Poker AI? riight... by gehrehmee · · Score: 3, Interesting

      Think that's frightening?

      Check out this article on the Second International Roshambo (Rock Paper Scissors) Programming Contest. It's actually quite interesting to understand some of the justifications and rationalles that go into attemping to win at a normally un-winnable game:

      Game theorists have analysed rock-paper-scissors and come to the conclusion that the optimal way to win a game of rock-paper-scissors is to play completely randomly; random play will win as many throws as it loses and hence draw every match. However, consider trying to win a tournament by drawing every match!

      Therefore, when trying to win a rock-paper-scissors tournament, you should assume that players will be trying to win the whole event and hence will not be playing optimally. Therefore you shouldn't play optimally - instead, you should figure out how to play in order to beat your opponent.

      --
      "You know, Hobbes, some days even my lucky rocketship underpants don't help" -- Calvin
    5. Re:Poker AI? riight... by gehrehmee · · Score: 2, Interesting

      Oh, and before I forget, the Roshambo Programming Contest is hosted by the University of Alberta, the same university responsible for the above poker AI article.

      They've also created the world's best checkers player, human or machine. Chinook utilized a distributed computing solution for mapping and optimizing its checkers stratagy back in 1989. IIRC from the talk Dr. Jonathan Schaeffer gave on it, this distributed network accounted for 80-90% of the Internet traffic between the United States and Canada in its day :)

      --
      "You know, Hobbes, some days even my lucky rocketship underpants don't help" -- Calvin
    6. Re:Poker AI? riight... by gehrehmee · · Score: 2, Interesting

      Let's say there's 100 bots. Just to make it ridulously clear, let's say 98 of them are using random choice. Our bot, which will use the "suboptimal" solution, will still come up against all of the random bots in a draw, since they're playing randomly. The only thing left to affect our total score is that one other "suboptimal" bot. If our sub-optimal bot can beat it on average, we get a slight edge, and win the tournement. The same applies to the 50-50 case. Basically, you can't win or lose against the random bots. So you might as well play your own stratagy. The other problem to look out for is that going into a match, you don't know anything about your opponent. It could be a random bot, or any kind of strategic algorithm. Once again, if it's random, it doesn't matter what you play, you'll still come out even. But on the slightest chance that it's non-random, you can try your strategy. You can also take it a step further, and try to deduce what the other guy's stratagy is. If you know exactly what his Roshambo algorithm is going to pick, then you can always beat him. Then again, you run the risk of misintepreting a strategy (or a series of random plays *as* a strategy), and playing yourself into the ground. Or you could end up playing your adaptive algorithm against someone else's adaptive algorithm, and getting into some really interesting competition :)

      --
      "You know, Hobbes, some days even my lucky rocketship underpants don't help" -- Calvin
  2. Gem by mao+che+minh · · Score: 4, Interesting
    I was following various links about the topic (artificial intelligence and poker) when I found this little gem. Wicked awesome site design, so I can only assume that his software building skills are as magnificent: POKER WITH AI-LEARNING

    I say we help him beta test not only his program, but also help him stress-test his web server.

  3. Re:Tell me... by Gorm+the+DBA · · Score: 3, Interesting
    RTFA...One of the things covered is that bluffing and reading tells, although effective at taking the occasional hand and a marginally winning strategy, is not as effective as it is given credit in the Poker Community. A better long term strategy is just playing strong hole card positions (especially in Hold-em, where 5/7 of the ending hand is the same between all players.)

    Now, they also say the machine has to be able to bluff, but the trick was to get it to do it the right amount, and at the right time postionally. Reading the opponent isn't as important as seeing the right situation in the cards.

  4. Re:Tell me... by rkent · · Score: 5, Interesting

    You RTFA:

    Peter Muller, a friend of Mr. Rao's who has played against the same bot, said the approximations in the game-theory model left a weakness and limited the bot's chances to do more than break even. Game-theory models usually assume that every player uses the best possible strategy, something that rarely if ever happens with humans.

    "An optimal game theoretic strategy might ensure that you don't lose, but it won't be effective at exploiting an opponent's weaknesses," Mr. Muller said. "The best players learn how to exploit predictability, but don't do it often enough so that the opponents catch on."

    In other words, it's easy to bluff a computer; you just play strongly and it'll assume you have a good hand and probably fold to you. Unless it's got a good hand, in which case you're screwed. Or if it has adoptive modelling that remembers how often you bluff, then you're REALLY screwed. Generally, though, it sounds like the Alberta AI just plays tightly, using "classes" of hands to avoid getting confused by the billions of possible hands, which does limit losses, but doesn't generally win big.

  5. Re:Why poker? by xTown · · Score: 2, Interesting

    Limit Hold 'Em is a game of patterns. There are a lot of variables to take into account--position, pot odds, expected value, number of bets/raises already in, your cards, and many many more. You make playing decisions based on all of these variables. For example, in late position with few callers you can consider playing cards that you would fold in early position. Then there's bluffing and semi-bluffing.

    Look at "Positively Fifth Street" by James McManus. He talks a little bit about the Alberta project and the rationale behind it.

  6. For non-robots, a simplified poker method by RobertB-DC · · Score: 5, Interesting

    I have never been any good at poker... in high school, playing nickel-ante poker, I lost about $25 to just one of my friends. Typically, after about 15 minutes of play, everyone was playing with "my" money.

    But recently, I spent some quality time with a hand-held poker game, and played the "hundreds or thousands" of games as described in the article. Not enough to become an expert, but I did come up with a technique to make my 100 credits last longer.

    I hacked away as much complexity as I could. The heart of my method is to forget about the effect of getting two cards you need. The chances of getting two specific cards is something like 1/52 * 1/52 = 1/2704 -- too small to care about. So the entire method is about the next card.

    Of course, I put it online: How To Lose Less At Video Poker. At the risk of slashdotting my own server, I'm curious if anyone can find any obvious flaws in the method.

    I found this Java-based tutorial that purports to generate the "optimum payout" -- it often disagrees with me, presumably because it's trying for big payouts. My method doesn't promise profit, only smaller losses.

    An important disclaimer: I've never used my method with any non-trivial amount of actual cash. Here in Texas, there are video poker machines in every Quickie Mart, but I just don't see the appeal. Now, if they would put in a Pac-Man machine...

    --
    Stressed? Me? Of course not. Stress is what a rubber band feels before it breaks, silly.
  7. Unix friendly online rooms? by unperson · · Score: 2, Interesting

    Okay, I know this is a lil' off topic, but as one of the (presumably) many slashdot readers who play poker...has anyone found an online poker room that you can play for real $$$ where I don't need to download some windowsware to play?

    I've toyed around with the java-based yahoo rooms (which last I checked, didn't have a real $$$ option). All of the big name poker rooms that I've seen through friends require a windows based client. I've been dying to give it a try.

    I'd also be interested in anything anyone wants to post under this thread about poker room security. Are there many malicous online poker room opponents out there? I've seen a few cases where someone was about to sweep in a hold'em hand, holding the nuts(*), only to be booted off the game at the last second...any thoughts?

    (*) Attn Trolls: This is an actual poker term.

  8. Good, let's forget about chess... by feidaykin · · Score: 2, Interesting
    Chess "AI" is really pointless. All that is needed for playing chess is a fast computer and lots of stored moves, mainly because the number of possible moves is quite finite.

    While poker is an interest game to tackle, I think I'd have to agree with others here that it is more of a people game and hard to a machine to understand the nuances of "bluffing" and other things that we silly humans do.

    What I'd really like to see is computer AI able of playing the Japanese/Chinese board game called Go at advanced levels.

    As it is right now, the best Go AI is only at intermediate levels it terms of skill. Here is a URL comparing Chess to Go...

    http://www.villagenet.com/users/bradleym/Compare.h tml

    --

    "To confine our attention to terrestrial matters would be to limit the human spirit." -Stephen Hawking

  9. Does anyone play checkers anymore? by swordgeek · · Score: 2, Interesting

    No, it's not a troll. :-)

    But a few years ago, checkers was solved as a mathematical problem. There is a computer program that can play a perfect game of checkers, all the time. That project was headed by Jonathan Schaeffer, one of the people involved in this Poker AI project.

    Just a footnote, to let you all know that this group has some serious history in gaming AI.

    --

    "People who do stupid things with hazardous materials often die." -- Jim Davidson on alt.folklore.urban
  10. Re:NYT WANTS TO STEAL YOUR THOUGHTS! by Anonymous Coward · · Score: 1, Interesting

    Make up bull$#!t details. They have no right to the information you provide, so you have no obligation to tell them the truth. (*)

    Women living alone are supposed never to reveal their sex to anybody who doesn't need to know it. Suppose some cyberstalker was tapping internet connections at random till he found one that belonged to a woman - and then began stalking her and eventually raped her?

    It is actually illegal anyway to target advertisements based on a person's answers. For instance, let's suppose the advertisers want to sell shoes to women and video games to men, so they show you an advert for a pair of shoes or a video game depending on what sex you said you were. But they've just violated the Sex Discrimination Act 1976. A woman has a right to see video game adverts and a man has a right to see shoe adverts.

    Anyway, nobody has any right to know what is between my legs unless they are proposing marriage.


    (*) This of course does not address the real issues ..... A compromise would be to make up one bogus account and share it widely.

  11. Wow, that's kinda neat... by StressGuy · · Score: 2, Interesting

    The first ever complex program I wrote was, in fact, an attempt to make an intelligent poker player. It was written in BASIC for an IBM with an 8086 processor and about 7-feet long when printed out. It made ASCII representations of the cards and I had my own random number generator that used the time the program started as the "seed" value. It had an independent routine for "bluffing" and made it's more rational decisions based upon what I was to eventually learn was a pretty decent implementation of "fuzzy logic". ...problem is, it played lousy poker...and I could never figure out why. At least it followed the rules though.

    --
    A goal is a dream with a deadline