Open Side Menu Go to the Top
Register
Poker Bot: 'A Nuclear Weapon For Poker' Poker Bot: 'A Nuclear Weapon For Poker'

09-29-2014 , 09:53 AM
A lot of people here have the wrong idea about the limitations of software. (I have started several software companies, have written chess and poker programs, and have developed "learning" hedge fund algorithms that outperform human experts). A few points:

Stars may be detecting many bots, but I am confident that the best ones go undetected. Bot-detection depends on the software playing in a machine-like way, which means unnatural levels of repetitiveness (of speed of action, strategic lines, lack of mistakes etc). A smart bot developer can program in all of these kinds of fallibility and randomness, putting in enough variation to fool the detection algorithms, but not so much that it falls out of profitability.

Some people are assuming that bots will always be GTO-based. That is a solid starting point, but just as the most successful players understand GTO but move away from it to exploit specific player leaks when they spot them, so a well-constructed bot can incorporate this capability.

Multi-player games are an increased challenge, but much more so for humans than for computers. The successful bots in these complex situations will not be ones based on clever multi-player GTO theory, but much dumber ones that learn through genetic approaches. With the ability to try random strategies and let the best ones "evolve", that sort of software will easily crush anything that tries to implement a human-designed algorithm. They will use evolution to derive base strategies, and then on-the-fly learning approaches to detect and exploit player habits and table dynamics.

I have no doubt at all that on-line poker will eventually be killed by these programs, and I am suspicious that some of the steady increase in toughness of games in the last couple of years is already coming from this factor.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-29-2014 , 10:25 AM
Quote:
Originally Posted by Hig Hurtenflurst
Databases -> Statistical analysis -> Algorithm -> Nuclear Bum-Hunting Bots
Uranium -> Metal container -> Drop from plane -> Nuclear Bomb

Hmmm, something seems to be missing...

Juk
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-29-2014 , 01:02 PM
Quote:
Originally Posted by LukeSilver
I am sure we have all played naked at some point.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-29-2014 , 04:04 PM
Quote:
Originally Posted by Paragon
Someone was kind enough to let me know there were import errors in the latest version of HM2. I uploaded new logs that shouldn't have any problems (I run the site).
Cool. HEM would only import hands that went to showdown, so I'll have another look at your site and see if you've fixed that issue. If you're the guy behind Prelude, then I'm very impressed.

Quote:
Originally Posted by ESKiMO-SiCKNE5S
Wasn't this Terminator? Or Jurassic Park? Or something?
It wasn't a quote, but my ideas may have been influenced by those films (and stuff like 'I Robot' and 'Westworld'). I'm a great believer that technological "progress" is unstoppable, but I'm also more pessimistic about it than most people, without actually being a luddite.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-29-2014 , 06:20 PM
Good post, thanks.

Quote:
Originally Posted by Rimlog
A lot of people here have the wrong idea about the limitations of software. (I have started several software companies, have written chess and poker programs, and have developed "learning" hedge fund algorithms that outperform human experts). A few points:

Stars may be detecting many bots, but I am confident that the best ones go undetected. Bot-detection depends on the software playing in a machine-like way, which means unnatural levels of repetitiveness (of speed of action, strategic lines, lack of mistakes etc). A smart bot developer can program in all of these kinds of fallibility and randomness, putting in enough variation to fool the detection algorithms, but not so much that it falls out of profitability.

Some people are assuming that bots will always be GTO-based. That is a solid starting point, but just as the most successful players understand GTO but move away from it to exploit specific player leaks when they spot them, so a well-constructed bot can incorporate this capability.

Multi-player games are an increased challenge, but much more so for humans than for computers. The successful bots in these complex situations will not be ones based on clever multi-player GTO theory, but much dumber ones that learn through genetic approaches. With the ability to try random strategies and let the best ones "evolve", that sort of software will easily crush anything that tries to implement a human-designed algorithm. They will use evolution to derive base strategies, and then on-the-fly learning approaches to detect and exploit player habits and table dynamics.

I have no doubt at all that on-line poker will eventually be killed by these programs, and I am suspicious that some of the steady increase in toughness of games in the last couple of years is already coming from this factor.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-29-2014 , 07:00 PM
Quote:
Originally Posted by David Sklansky
I said "adjusting" not analyzing. They won't ever beat a ring game with bad players for more than human experts will. They also will never do as well in heads up tournaments as human experts will as long as there are many bad players in the tournament.

And if you reply that they can be programmed to adjust, that, besides being hard to do, opens them up to being double crossed as long as smart opponents realize they are up against a computer that is trying to adjust.
Sadly the computers of the world will soon be well equipped to handle all of your cunning, deceitful, double crossing triple range merging tricks and chicanery David, despite your very human protestations and claims that such programming is "hard to do."

Its like, the future, man.

EDIT: didnt click link but Im guessing this is in some 30bb donkament format. Wake me up when bots crush 500bb deep, and I will start paying attention.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-29-2014 , 08:52 PM
"Over the past few decades computer scientists have tested the capacity of their game playing against top human players, with famous matches like Garry Kasparov vs. IBM’s super computer Deep Blue in chess and poker pros Phil Laak and Ali Eslami taking on the University of Alberta’s limit-hold’em heads-up bot Polaris. At first humans were able to compete, but eventually the programs gained the advantage."


/
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 02:03 AM
Quote:
Originally Posted by Burdell
Especially multiway non-fixed-limit poker, which probably does not have a Nash equilibrium, seems many years away from the computers being all that good at it. Even HU hold-em has far more complexity than computing power for finding a solution. The researchers behind this bot simplified the game by looking at three streets at a time and then tries to approximate the solution.
What does it mean when people say computers doesn't have enough computer power to find a solution? I read this multiple times, but why do they say this. Is it about some hardware limitation, or that the adequate software isn't develop yet?
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 04:15 AM
Quote:
Originally Posted by monTiel.
What does it mean when people say computers doesn't have enough computer power to find a solution? I read this multiple times, but why do they say this. Is it about some hardware limitation, or that the adequate software isn't develop yet?
I think they mean a hardware limitation. The game tree is too big for current computers to compute all the possible variations. But it might still be possible for a clever algorithm to prune the game tree to a computable size.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 07:26 AM
Quote:
Originally Posted by LukeSilver
i was concerned enough to do some research on the matter and found a few things out firstly people have been worried about this since 2005 and in 9 years no bot has managed to kill online poker yet.
Did that research include the Svenska Spel case where some $500k was repaid to some 25,300 players?

http://no.donkr.com/Articles/svenska...mpensated-5457

Still to come to trial BTW.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 09:51 AM
Quote:
Originally Posted by David Sklansky
If the game is not holdem, not heads up, and has a structure that encourages multiway pots there will never be a non adjusting bot that plays better than the best humans.
Even if that's true, that doesn't matter. Online poker is dead as soon as the bots can defeat a certain percentage of players. At that point the poker economy is crushed.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 09:54 AM
Quote:
Originally Posted by Rimlog
A lot of people here have the wrong idea about the limitations of software. (I have started several software companies, have written chess and poker programs, and have developed "learning" hedge fund algorithms that outperform human experts). A few points:

Stars may be detecting many bots, but I am confident that the best ones go undetected. Bot-detection depends on the software playing in a machine-like way, which means unnatural levels of repetitiveness (of speed of action, strategic lines, lack of mistakes etc). A smart bot developer can program in all of these kinds of fallibility and randomness, putting in enough variation to fool the detection algorithms, but not so much that it falls out of profitability.

Some people are assuming that bots will always be GTO-based. That is a solid starting point, but just as the most successful players understand GTO but move away from it to exploit specific player leaks when they spot them, so a well-constructed bot can incorporate this capability.

Multi-player games are an increased challenge, but much more so for humans than for computers. The successful bots in these complex situations will not be ones based on clever multi-player GTO theory, but much dumber ones that learn through genetic approaches. With the ability to try random strategies and let the best ones "evolve", that sort of software will easily crush anything that tries to implement a human-designed algorithm. They will use evolution to derive base strategies, and then on-the-fly learning approaches to detect and exploit player habits and table dynamics.

I have no doubt at all that on-line poker will eventually be killed by these programs, and I am suspicious that some of the steady increase in toughness of games in the last couple of years is already coming from this factor.
Very good post. People seem to have an idea that a bot is just a series of if/then/else steps. There are a bunch of very powerful machine learning algorithms out there, and the field is only in its infancy.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 06:23 PM
Quote:
Originally Posted by cdog
Wake me up when bots crush 500bb deep, and I will start paying attention.
I can't speak about 500bb deep, but how do you feel about a couple of 6-max bots winning more than 10bb/100 at 1000NL?

This one has apparently just been banned on Party Poker, after taking a quarter of a million dollars from the poker economy:


(Weird) thread about the Party Poker bots here: http://forumserver.twoplustwo.com/28...-view-1474198/
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 08:27 PM
Quote:
Originally Posted by Dima2000123
Even if that's true, that doesn't matter. Online poker is dead as soon as the bots can defeat a certain percentage of players. At that point the poker economy is crushed.
Not if the game is such that its still fun for amateurs and gives them a good chance for winning sessions.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 09:06 PM
Quote:
Originally Posted by Rimlog
A lot of people here have the wrong idea about the limitations of software. (I have started several software companies, have written chess and poker programs, and have developed "learning" hedge fund algorithms that outperform human experts). A few points:

Stars may be detecting many bots, but I am confident that the best ones go undetected. Bot-detection depends on the software playing in a machine-like way, which means unnatural levels of repetitiveness (of speed of action, strategic lines, lack of mistakes etc). A smart bot developer can program in all of these kinds of fallibility and randomness, putting in enough variation to fool the detection algorithms, but not so much that it falls out of profitability.

Some people are assuming that bots will always be GTO-based. That is a solid starting point, but just as the most successful players understand GTO but move away from it to exploit specific player leaks when they spot them, so a well-constructed bot can incorporate this capability.

Multi-player games are an increased challenge, but much more so for humans than for computers. The successful bots in these complex situations will not be ones based on clever multi-player GTO theory, but much dumber ones that learn through genetic approaches. With the ability to try random strategies and let the best ones "evolve", that sort of software will easily crush anything that tries to implement a human-designed algorithm. They will use evolution to derive base strategies, and then on-the-fly learning approaches to detect and exploit player habits and table dynamics.

I have no doubt at all that on-line poker will eventually be killed by these programs, and I am suspicious that some of the steady increase in toughness of games in the last couple of years is already coming from this factor.
Although I agree with most of this I again claim three things: (Although only number 1 really matters.)

1. There are poker games that are more fun, allow looser play and more multi way pots and enough volatility such that even "perfect" players or bots won't really ruin the game or intimidate the recreational player.

2. Bots will never be programmed to take full advantage of bad players because it is a needless risk given they are certain to have an edge without doing this. Thus great players will outperform them in situations where there are bad players in the mix.

3. Bots that adjust away from GTO based on how players play will probably increase their winnings UNLESS the opponent is an expert who KNOWS that this is what he is up against (while the computer doesn't know he knows). Do you see why?
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 09:23 PM
Quote:
Originally Posted by David Sklansky
Not if the game is such that its still fun for amateurs and gives them a good chance for winning sessions.
Bots are anti-fun. Amateurs will not enjoy the game if they know they are fighting bots instead of real people. Look at video games where people use bots and cheats, it really pisses people off. It's worse for poker because real money is on the line.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 09:26 PM
Quote:
Originally Posted by David Sklansky
Although I agree with most of this I again claim three things: (Although only number 1 really matters.)

1. There are poker games that are more fun, allow looser play and more multi way pots and enough volatility such that even "perfect" players or bots won't really ruin the game or intimidate the recreational player.

2. Bots will never be programmed to take full advantage of bad players because it is a needless risk given they are certain to have an edge without doing this. Thus great players will outperform them in situations where there are bad players in the mix.

3. Bots that adjust away from GTO based on how players play will probably increase their winnings UNLESS the opponent is an expert who KNOWS that this is what he is up against (while the computer doesn't know he knows). Do you see why?
I used to play limit games, all the HORSE games and some of the draw games. I couldn't stand the NLHE games bec I'd play 2 tables, had to wait for the 20 tablers to act, not much action and all of that. But this is online being discussed and after all of these years this forum still largely scoffs that there is anything other to be played than NLHE. I think some of that is bec the players would have to give up their suite of software aids.

All of the 'poker is dying' and 'the games are so much tougher' threads are about NLHE but there's seemingly no way to move those ppl off their asses and into the other games. So gl trying to convince them.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 09:29 PM
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 09:32 PM
Quote:
Originally Posted by Howard Beale
I used to play limit games, all the HORSE games and some of the draw games. I couldn't stand the NLHE games bec I'd play 2 tables, had to wait for the 20 tablers to act, not much action and all of that. But this is online being discussed and after all of these years this forum still largely scoffs that there is anything other to be played than NLHE. I think some of that is bec the players would have to give up their suite of software aids.

All of the 'poker is dying' and 'the games are so much tougher' threads are about NLHE but there's seemingly no way to move those ppl off their asses and into the other games. So gl trying to convince them.
I think they would rather learn new games than get a job.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 09:34 PM
What a delightful thread. I am so looking forward to seeing all the internet players in live games.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
09-30-2014 , 10:26 PM
Is there a lot of bots playing turbos mtts on ps? Seems to me it would be easy to do since its push/fold poker most of the time
Poker Bot: 'A Nuclear Weapon For Poker' Quote
10-01-2014 , 12:22 AM
PS or any site could implement periodic visual recognition tests. The most advanced robots are not even close to matching the human eye, and would take completely different software to boot
Poker Bot: 'A Nuclear Weapon For Poker' Quote
10-01-2014 , 01:12 AM
Quote:
Originally Posted by David Sklansky
Although I agree with most of this I again claim three things: (Although only number 1 really matters.)

3. Bots that adjust away from GTO based on how players play will probably increase their winnings UNLESS the opponent is an expert who KNOWS that this is what he is up against (while the computer doesn't know he knows). Do you see why?
Could you elaborate David? I don't see why you think a human exploitative player would necessarily be able to beat an exploitative bot. Aren't they both simply using Bayesian analysis to intuit their opponents strategy? What makes you think a human will be better at that than a computer?
Poker Bot: 'A Nuclear Weapon For Poker' Quote
10-01-2014 , 01:35 AM
Quote:
Originally Posted by David Sklansky
I think they would rather learn new games than get a job.
That's about what it would take. There is still the problem that the recs want to play NLHE also. Maybe if the sites would promote the other games w/ reduced rake or promos they'd get decent traffic, IDK, it doesn't affect me anymore until online is legal in the U.S.
Poker Bot: 'A Nuclear Weapon For Poker' Quote
10-01-2014 , 02:21 AM
Quote:
Originally Posted by Frankie Fuzz
Could you elaborate David? I don't see why you think a human exploitative player would necessarily be able to beat an exploitative bot. Aren't they both simply using Bayesian analysis to intuit their opponents strategy? What makes you think a human will be better at that than a computer?
Because the computer is lacking a piece of information that the human has. It doesn't know that the human knows that it is a computer adjusting based on previous play. (Unless it is programmed to assume that)

(If they both knew that the other knew that the other knew.... the strategy coalesces to GTO)
Poker Bot: 'A Nuclear Weapon For Poker' Quote

      
m