someone mentioned poker bots a while back...

podgeen

news story

NickyOD

"He lurks online at the tables for the chicken-hearted; even there, where the biggest ante is 4 cents, he can't win consistently.
But Gabriel has a potentially powerful alter ego. In his spare time, he's perfecting a computer program to go online and play the game for him.
His BlackShark software is still a work in progress, but Gabriel has no doubt that such programs eventually will be championship quality. "In the future," he said, "robots are going to take over."

lol! he can't win at the penny table but thinks he can program a bot with the algorithms to win for him. Hmmm, yeah that makes sense.

Kingp35

I think the programmer himself needs to learn how to play before he can program a bot to do it!

Good luck to him. Still not good news for online players if they manage to make a half decent one. The average player will be screwed

jimbling

its not quite true that the programmer needs to be a good poker player to write a good bot...
ask yourself this.... was the human who developed the Deep Blue supercomputer as good a chess player as Garry Kasparov... obviously not. The point is that these things can be programmed to learn and adapt.... they also take in a huge amount of information to make choices with.


For the next 5 years at least, a good human poker player will demolish a bot, but the majority of poker players beware....

NickyOD

The real difference between the best pro chess players is the speed at which they make their decisions and a lot of the time the game ends in a draw. It's much easier for a chess bot to win because he is up against only one oponent, can predict all his oponents possible moves and make a very precise mathematical decision which will almost always have positive expectation. There is no "unknown" element to chess. All of the pieces are in play.

At a poker table bots are up agiainst several oponents with unknown cards, and unknown possible community cards to play for. Players moods change. They may play the same hand completely differently on seperate occasions. Its too dificult for poker bots to adapt to this at the same speed of the human brain, nor can they pick up tells from the chat or the speed at which a player makes a call or a raise. The feel factor cannot be represented in a mathematical algorithm. Poker bots will never ever be successful at high stakes. The best they can do is win at shorthanded low limit holdem and even there they are not exactly rocking the online poker world. Lets face it there are enough supernerds out there and online poker is around long enough. If it was that easy we'd have heard about it by now, but all I read about from the nerds with this software is how they are stuggling at anything higher than the penny tables where simple ABC poker will reward you, and those that say they are making 3BBs/100 at 2/4 are totally full of **** IMHO.

It would take a super intelleigent programmer who is also a highly successful poker player to create a successful poker bot to win at above average stakes limit holdem. which begs the question why, when he is already a winning player?

With sites like Party already using pretty sophisticated technology to catch these people out all it takes is a daily update to the way the HH or the cards are displayed on the screen (I hear winholdem can't read the cards on PS) for the programmer to be forced to constantly change and adapt his bot. Bots are easy for a host to detect. They cant multitable, they can't respond in the chat and they constantly take a long time to make decisions because the bot has so many scenarios to contemplate. I saw this happen on Full Tilt already. Bots will not take over, not in 5 years, not any time. Never.

Kingp35

jimbling wrote:

its not quite true that the programmer needs to be a good poker player to write a good bot...
ask yourself this.... was the human who developed the Deep Blue supercomputer as good a chess player as Garry Kasparov... obviously not. The point is that these things can be programmed to learn and adapt.... they also take in a huge amount of information to make choices with.


For the next 5 years at least, a good human poker player will demolish a bot, but the majority of poker players beware....

I dont think thats really true either. I agree that the people who wrote the bots for chess and backgammon didnt need to be great players because those games are based on a series of moves which judging from a particular move the bot can decide what to do.

Poker is totally different in that it is a psychological game more so than anything else. You could write a bot to learn a players habits in that how often he folds etc. and this can be calculated into the bots game play but if a new player joins the game the bot is screwed as it will take at least 50 hands for it to really grasp how this player plays. Also everyone knows there are so many players out there who you just cant predict what they are going to do. Basically the bot will work out the probability of winning and combine it with a players habits and decide then what to do. This mat sound good but Poker as you all know isnt only about the cards. You could teach a bot to play the cards according to its data but you cant teach it the psychological aspect of the game unless you yourself know it. That is the real skilled part of poker imo that only a human can really grasp. A bot will do something according to its data, a human can do absolutely anything and unless the bot actually knows what cards he has there is no way of really knowing what it should do. Its a tricky one really and tbh i cant see a bot capable of beating good players being devised for a long long time until AI has developed alot.

I could be wrong though because im definitely not an expert in bots.
Might be a good idea to post this in one of the comp forums to see what the people in there think. Im interested to know if anyone actually thinks that this could really work. If it somehow could it would spell serious problems for online poker

Imposter

I read somewhere about a bot who was let loose in 10 or maybe 20$ stt's. It was using a slightly modified version of Sklansky's idiot proof tournament strategy. It was going grand until one of the modifications gave it away. I think it was on 2+2 I read it. I can beleive that such a bot is possible even for cash games, it could possibly work if the opponents were stupid enough or the modifications not numerous enough.

Hitchhiker's Guide to...

NickyOD wrote:

Bots will not take over, not in 5 years, not any time. Never.

"the personal computer will never take off" CEO IBM circa 1980

always dangerous to make definitive statements about the future!

jimbling

you both took my argument, and tried to rip it to shreds by comparing chess with poker... but thats irrelevant...

do you think there is no mind games in chess... you obv dont understand the game at all... and dont start again, i realise there is far more of it in poker, and its a psychological game etc......

but for years... years I tell you, you had serious chess players all over the world saying exactly what nicky stated...

'Poker bots will never ever be successful at high stakes.' ie... a computer can never ever compete with a human master chess player... but it can and it does.


Compare all you want.... and deny all you want.. there will be a problem. Now wether its the sites that moniter it and dont allow bots etc i dont know. But it most likely will be an issue in years to come.

I have read a bit about neural nets etc, which is an amazing technology and is only in its infancy... check it out if you wish.

jimbling

exactly roundtower....


and by the way... i personally dont think they will take over either... but I do think that it will be an issue that will have to be taken seriously... thats all my point was really....

Vamos

There are bots already out there that have oppenent modelling which would beat most below average -> average players who even have some slight unpredictability in their play and in the future they will only get better.

http://www.poki-poker.com/ai.htm

I think it could mean the downfall of online poker in the future even though current bots arent upto much, and I hope anyone who uses one gets caught and banned from every site. It's cheating and they are total scum in IMO.

NickyOD

It wasn't my intention to tear your post to shreds. I was only voicing my opinion. I have played chess since I was 12 (although not much since poker took over) so I know quite a bit about the game. I don't believe you can compare chess to poker, because in poker there are too many unknowns for a bot to decypher, while all the pieces and possible outcomes are visible on a chess board. That is all.

jem

Vamos wrote:

they are total scum in IMO.

Now Now leave manu out of it

Hitchhiker's Guide to...

The pokerbot idea sounds like it could be useful but only for betting preflop and only for bets up to maybe 4xBB. The bot could be programmed to bet or fold along a simple series of strategies for openers. If the betting preflop goes over 4xBB then the person is called to the table to make the decision. After the flop etc the person makes all the decisions. The advantage of this is that the player can now play more games simultaneously and only has to apply their time to actual play. I don't see how this could be viewed morally as cheating. Preflop betting online where you have no significant reads on people should be as boring as a computer programme. You should have a set or rules for preflop betting online unless you have a read on a fellow player. What do you think?

NickyOD

There's nothing really in that article that wasn't written a year ago and poker bots have not moved to the next level since and show no signs that they will. The part about the Vexbot is a terrible exaggeraion. It is considered to be the top current Poker AI but it can only play Heads up and takes thousands of hands against a single player to adapt. In a multihanded scenario it would be completely useless if it took that long.

Spiritus

NickyOD wrote:

... I saw this happen on Full Tilt already...

What happened?

NickyOD

Spiritus wrote:

What happened?

Guy was using a bot in $20 SnG. It took him almost the full timebank to make decisions every time preflop and postflop, sometimes he actually timed out. he had some basic moves. Standard 3XBB raise, limp or all in. If he reraised it was always all in. Support was notified during the game and he was asked to respond in the chat. He didn't. He actually managed to come 3rd, but that doesn't really tell you anything about the long term capabilities.

padser

NickyOD wrote:

, in poker there are too many unknowns for a bot to decypher, ......... That is all.

Too many unknowns, as in too many variables????? Oh well computers have never managed to overcome the problem of extra variables before so im sure they wouldnt manage this time either!!! (my sarcastic voice).

I really think that its extremely naive to think that because a technology that is in its infancy it wouldnt improve dramatically.

Remember this conversation in 20 yrs time

DannyGT

It is possible to make a bot that is able to beat stakes upto 2/4 limit without introducing AI. If there is no AI needed then it is alot easier. Most players upto this level are making huge mistakes in there game. Playing full handed the bot wouldnt have to get involved in slightly +EV hands, making it easier to program. If you play shorthanded you have to start looking at how other people are playing to determine what you should do.

A bot would not need to be beating it for 3bb/100 to make it worthwhile. Bots can multitable and there is multiple skins for most sites. Even if the bot was break even and you were getting 25%-30% rake back it could make 50-200 a day.

NickyOD

Are you using one?

DannyGT

No.

Know some one that made one and had it running. He never got it past .5/1 though.

NickyOD

DannyGT wrote:

No.

Know some one that made one and had it running. He never got it past .5/1 though.

DannyGT wrote:

It is possible to make a bot that is able to beat stakes upto 2/4 limit without introducing AI.

huzzuh?