The ethics of poker bots - what are your thoughts?

[wayniac]

There have been occasions where I have played online poker and honestly felt that I was up against "bots". Maybe it's just delirium, but there are certain players that always bet in such a predictable manner and always play the odds. These players coincidentally never have chat enabled, which is almost criminal at the end of a big tournament that goes for 4 hours or so. Whether I am correct or not about their current status, the possibility that artificial bots could be unleashed on poker sites is not out of the realms of possibility. With serious development, these bots could conceivably hold the edge over most amateur players and "get in the money" enough to justify their operation. They could also play many games at the same time. Most sites do explicitly ban the use of such programmes but I doubt that the makers could not circumvent their protection measures. Anyway, before I start sounding like a paranoid lunatic, I will just end with a simple question.... Would it bother you if others were using poker bots in a tournament you are in?

[mrmuzeman]

Re: The ethics of poker bots - what are your thoughts? It wouldnt bother me cos im also a poker machine baby! In all seriousness though bots are a limit poker thing. Theyd be shite at no limit and since I dont play limit I dont really care.

[GaF]

Re: The ethics of poker bots - what are your thoughts? I haven't got a problem with bots in themselves (cos I'm not aware of any that exist that can take on no limit with any decency - shocking when you think computers can beat World Chess Masters, but can't beat average muppets at poker!!!) - where I'm worried is if you have bots working in tandem - if you've got the bots working well, and in cooperation on the same table, they'll be unstoppable!!!!! Fortunately I don't think they exist (yet!!!!) cos when they do, we'll know cos we'll all be losing consistently - it can (unfortunately) only be a matter of time.... I know one exists (can't remember it's name, but PartyPoker searches for it, which is overcome by using an equivalent of PC Anywhere and 2 machines) - but understand that it's AI isn't very good.......

[Sir Puntalot]

Re: The ethics of poker bots - what are your thoughts? No bot could match a top class poker player, it would take programming beyond belief.

[GaF]

Re: The ethics of poker bots - what are your thoughts?

No bot could match a top class poker player' date=' it would take programming beyond belief.[/quote'] But it only has to beat your average muppet, not world class players, to be profitable ...... surely not that hard? And maybe not today.......but one day bots will beat allcomers (just like in Chess).........

[mrmuzeman]

Re: The ethics of poker bots - what are your thoughts? Pokers a game of incomplete information and in no limit hands can be played in several ways none necessarily right or wrong sometimes. In Chess everythings there to see, its very different. I dont believe a bot could be any better than just another player tbh. Plus theres races and all ins and stuff where luck can play part etc, in chess you know what will happen and there is zero luck.

[guesswest]

Re: The ethics of poker bots - what are your thoughts? There are bots out there that are doing well at no limit. Vexbot played Phil Laak a little while ago, it did lose, but was a good match and people put them about even going into it. Pokibot, which is one of the commercially available bots, claims to have reverse engineered Vexbot, but based on reports of it's performance, is a doubtful claim. Nevertheless, it's already happening. Winholdem and Pokibot are the two commercially available bots, as GaF pointed out the AI is pretty crap, they about break even at micro limits , but people writing their own codes into these bots are making money already. There's an active forum over at Winholdem where people exchange tips for tweaking the software. Who knows what bots people are operating on sites tho - if you do have a truly good bot I can't see much incentive to sell the source code. The leap from fixed to NL comes when the bots can learn about their opponents as well as playing the odds, that's already happened with projects like Vexbot and the AI is getting better all the time - people will be playing online with bots like this soon, if they aren't already.

[wayniac]

Re: The ethics of poker bots - what are your thoughts? wow! Interesting stuff guesswest... I guess I'm not going crazy after all. I'm gonna go check out that thread and see if I can fish up some more info for people here. Thanks mate.

[wayniac]

Re: The ethics of poker bots - what are your thoughts? U of A holds winning hand in video poker Edmonton research in high demand for electronic games

Shawn Ohler

The Edmonton Journal

Monday, November 07, 2005 EDMONTON -- Poker pro Phil Laak had the great nickname (The Unabomber). He had the hot girlfriend (actress Jennifer Tilly) and the reputation at the table (crazy). But nothing could help him against the stranger from Edmonton. The stranger from Edmonton was killing him. Laak and the stranger were playing heads-up hold-'em in a Las Vegas poker room, and the stranger was a maniac, hyper-aggressive, raising relentlessly, forcing Laak to fold hand after hand. Sixty hands in, to no one in particular, Laak said: "I have a feeling I'm getting outplayed." Darse Billings heard Laak. And Billings, who knew more about the Edmonton stranger than perhaps anyone, nodded in agreement. The stranger had a great nickname, too -- Vexbot, a computer powered by artificial intelligence and programmed by Billings and the University of Alberta Computer Poker Research Group, a world-leading team of Edmonton computer science professors and students. Vexbot's high-profile match in July against Laak, famous in poker circles for his hooded-sweatshirt-and-shades nod to the original Unabomber, began a rush of good luck for Billings' group. Televised tournaments and big-money tours in the U.S., Canada and Europe mean Texas Hold-'Em poker is among the fastest-growing games in the world, and the Edmonton research is in increasingly high demand. BioTools, a business offshoot of the U of A group, recently licensed some of its artificial intelligence to the makers of Stacked with Daniel Negreanu, an Xbox, PlayStation 2 and PC hold-'em poker game that comes out in February. BioTools's own game, Poker Academy, is sold at Wal-Mart and will soon hit shelves at Best Buy and Future Shop. And Poker Academy Pro 2.0, an expert players' learning tool which the company has stacked with its most cutting-edge A.I., has drawn rave reviews in U.S. gaming mags. From a commercial standpoint, it's a successful culmination of years of work that began in 1997 with a program called Loki. Loki won regularly on play-money Internet servers but proved too predictable and slow to adapt against stronger human players. In 1999, Loki was rewritten as Poki, which boasted more sophisticated adaptive abilities and does well to this day in 10-handed Internet "ring games" against good players. Stacked employs the Poki A.I., and Negreanu, the slight, fair southern Ontarian who is to poker what Mike Weir was to golf before his swing abandoned him, insists it's wily enough to give even hold-'em vets a rough ride. "In every other computer poker game on the market, once you figure out how the computer plays, there's no more game. It's game over," Negreanu said. "What Stacked does, especially the more sophisticated bots, is adapt to your play. It's the only A.I. that does that, and it's crucial to a realistic experience, because that's what good players do: adapt." But Poki has a fatal flaw: it's very vulnerable in a heads-up, or two-player, game. So the team created a more defensive heads-up specialist dubbed Sparbot, which applied theories created by Nobel laureate mathematician John Nash, subject of the Oscar-winning movie A Beautiful Mind. Then came Vexbot, an A.I. system that constantly tries to exploit the weaknesses it finds in its poker foe. Laak was unnerved by Vexbot's incessant pressure in Vegas, and Billings said that showed when Vexbot made a big, unexpected raise on a hand's final card. "Phil thought for a long time about what to do, and finally said: 'If that's a bluff, it's over for humanity.' And he folds. And I'm thinking, 'Oh my God, that could have easily been a bluff,' because I knew the mental state the program was in. It was playing extremely aggressively." Sure enough, when Billings later scanned the computer's records, Vexbot had bluffed the poker pro out of a very large pot. "The computer's post-game analysis, which factors out most of the luck, shows Phil legitimately won, but it was quite a close match," says U of A professor Jonathan Schaeffer. "We didn't get thumped." It's proof of the academic rigour of Schaeffer and Billings's group that, despite Vexbot's showing in Vegas, the program has already been all but abandoned as a research tool. "Now we're pretty negative on Vexbot. It's technology that we're putting to bed," Schaeffer said. "Our initial enthusiasm for it was because Vexbot was this maverick kamikaze. But once people understood how aggressive it was, people were quickly able to zero in and exploit it, because it was too slow to adapt.... "It's a very challenging opponent against an average player. But it's not the program that's going to beat the world champion, and that's our standard." The team is already working on A.I. that will immediately adapt to an opponent's tendencies. And that future, that promise and purity of science, is clearly what lures Billings more than the university's connection to Daniel Negreanu and Stacked's attendant hype. "I don't have a TV. I don't follow pop culture. I don't know if Stacked is supposed to be the next greatest thing. I know it's a game coming out, but that's about it," Billings shrugged. A former professional poker player lured back to academia by the complicated elegance his game's "imperfect information" poses to A.I. scientists, Billings figures the next great hold-'em bot will come after he finishes his PhD and writes it himself. "The bot I'll write will be considerably better than anything we've produced so far. I don't know if it'll be a world beater, but it'll definitely be a big improvement." Billings says this with the matter-of-factness one uses to tell a child that water freezes at zero. It is a claim with no arrogance. It would be unwise to bet against him.

[brael]

Re: The ethics of poker bots - what are your thoughts?

But it only has to beat your average muppet, not world class players, to be profitable ...... surely not that hard? And maybe not today.......but one day bots will beat allcomers (just like in Chess).........

And that's where the fun ends... poker is supposed to be about fun... bots could kill this off. Poker, as with many sports, needs human error to let your opponent in. If we've got machines playing poker beyond a level that we can't match, and if we don't know we're playing against them, we'll get thumped everytime and eventually give up. How much can a bot win when it's up against 9 other bots? Surely they'd be there for days, if not weeks or months? IMO it could turn poker into nothing but a futile game of tic-tac-toe! Taxi...! :loon

[Sir Puntalot]

Re: The ethics of poker bots - what are your thoughts?

And that's where the fun ends... poker is supposed to be about fun... bots could kill this off.

I doubt it Brael, how could you possibly know what/who you were playing?

[PAULM03]

Re: The ethics of poker bots - what are your thoughts? Been thinking about some of the comments in this thread and I think some of you are wrong in saying that no limit would be the most difficult game to create a bot for. Why? Well as Mr.M says poker is a game of incomplete info, however the amount of information you have decreases with each card shown. If you're dealt AA you know pre-flop that you have the best hand. However once the flop is dealt, in the majority of situations you don't know this anymore and the amount of uncertainty will usually increase with each card dealt. With this in mind you want your bot to see as little flop/post-flop action as possible, and the game that would fit this bill is no-limit. I would bet that a no-limit bot programmed with a solid pre-flop strategy would make more money than a bot playing limit.

[mrmuzeman]

Re: The ethics of poker bots - what are your thoughts? Why? By going all in all the time b4 the flop? Everyone would just fold. In limit there us a more 'correct' way to play and thus a bot could be able to fold, call, raise etc a lot lot easier. There is no doubt about this as bots are already commonplace in limit poker.

[PAULM03]

Re: The ethics of poker bots - what are your thoughts?

Why? By going all in all the time b4 the flop? Everyone would just fold. In limit there us a more 'correct' way to play and thus a bot could be able to fold, call, raise etc a lot lot easier. There is no doubt about this as bots are already commonplace in limit poker.

Who said anything about going all in all the time? I just said a solid pre-flop strategy. Every card that comes is potentially help to your opponent isn't it? And it must stand to reason that in limit more people with mediocre drawing hands/poor pairs/nothing will see the flop, because you can't control the bet size. In fact thinking about it pot-limit must in some ways be more of a skill game than no-limit because you need a better all-round game, whereas in no-limit there's more onus on your pre-flop strategy. I might be wrong, I don't make any great claims about my own game or my knowledge of the game, but it just seems logical.