Split discussion on bots (from 19 hands)

Originally Posted by Yakuman

If a mechanical system could beat the tables, someone could program a bot, set it loose, and rake in the dough.

If you don't think it's possible to program a bot that can beat the lower stakes, that's just delusional.

Those who believe that the WHOLE POINT of poker is that a system cant beat it. It is a game of incomplete information, and no system can beat such a game in the long run, period.

Sorry man, my thesis is in Artificial Intelligence, and regardless of what this might implicate about Aok's system.. this is just wrong. Every human uses a "system" to play aswell. The hunches and intuitions you use to play, are in fact subconscious odds calculations. An AI can be programmed like this too. It's no doubt a heck of a lot more complex than 19 hands to make a winning poker AI, but that's beside the point I'm trying to make here.. there's no doubt in my mind that it's possible.

there's already been bots found playing sngs (the lowest form of poker) at Party

Lol ya.. I use to play those sngs.. sooooo knew some of those guys were bots way before they got busted.. I just didnt care.. they were like a free double up if you were to their left.. it gets 4-5 handed.. they push on your BB with any 2.. over and over.. ty for the chips fishy : )

thing is they only got busted cos of the stupid bot raise they made.

ya.. thats what gave it away to me too.. and some other random things.. like when they did that dumbass raise.. like 97% their stack or w/e.. when u reraised allin for their last 30 chips.. they would take 5-10seconds to call.. like they had something to think about

Read this:

http://www.thisismoney.co.uk/money-s...3&in_page_id=5

Here's a quote:

"I'm doing pretty well. I have two computer systems, and each one can run four poker bots, and each of those four can play up to five tables at once. At worst I make on average £2.90 an hour at each table. That's a minimum of £116 an hour if I can get all the bots running at once."

Poker sites do everything they can to combat this new problem too. Here's another quote:

"I play Partypoker, pokerstars and other poker sites often. However, I have read on numerous msg boards that certain poker clients (PartyPoker in particular!) take screenshots of your desktop that are saved to a local folder (ex: c:\temp) and then they are sent to partypoker!

I read a letter from a former employee of partypoker who quit because of how unethical the behavior was. Apperently $5/hr costa rican employees would pass around funny screenshots of peoples desktops which may include things like private bank acct info!!"

Another interesting read:

http://www.thisismoney.co.uk/money-s...5&in_page_id=5

quote from the article:

"I got banned and had my funds taken when one of the biggest WinHoldEm botters in the world - a Vegas guy - got busted and it led back to me. He got lazy and had 50 accounts seized - enough money to buy a few new cars. The accounts were linked to bank details and then to another account that belonged to a pro poker player; I'd traded him an account of mine and the trail came back to me. Now I need two new identities to replace my bots. After the thousands of hours of programming and testing, I suppose this is the only hard part."

has anyone seen that video where the guy shows a wall full of PCs which are using bots?

no.. link me?

anyway, the difference between human "systems" and bots/19 hands is that we as humans can adjust accordingly.

Originally Posted by midas06

anyway, the difference between human "systems" and bots/19 hands is that we as humans can adjust accordingly.

If programmed correctly, bots can too.

Originally Posted by jackvance

Originally Posted by midas06

anyway, the difference between human "systems" and bots/19 hands is that we as humans can adjust accordingly.

If programmed correctly, bots can too.

At which point.. its hardly following a system such as a play chart

Originally Posted by Bailey

At which point.. its hardly following a system such as a play chart

Like I said earlier, I'm not saying this to defend 19 hands, just that bots can own online poker.

Originally Posted by jackvance

Originally Posted by Bailey

At which point.. its hardly following a system such as a play chart

Like I said earlier, I'm not saying this to defend 19 hands, just that bots can own online poker.

without a doubt.. but its nothing to fear.. you have to be good at poker inorder to make a winning bot... and at highstakes they just dont exist really.. cause anyone that is good enough to make a profit at those stakes.. wouldnt invest the 100s of hours it requires to code a poker bot : )

Though im sure a bot could beat the lowest levels of nl games eventually, once I figure out its a bot, I know exactly what level it's thinking on (1st) and can easily destroy it in pot after pot after pot after pot after...

what about a bot hooked up with PT so it knows what your ranges are?

Like Rilla says - if you're a decent player and you know you're playing a bot you can adjust to the bot like you would adjust to a regular player but with even more advantage. If the bot sees you're playing 35/20 and tightens up, then you play accordingly. It's like another player, only slower and more rigid in the adjustments it makes.

The only way to do it would probably be with an expert system, the kind that is used together with fuzzy logic to include expert knowledge.

But this means in order to even start working on such a programm, you would have to find a really really good player who is willing to share all of his thought process when he is playing poker, and it would take forever to do it.

Now that I think about it...definately an interesting topic, because you could combine it with a learning programm. If the bot has learning routines, it could learn while playing human players and the more it plays, the better it gets. This would also mean it's play always changes and it is not that easy to figuer it out.

estrop went bust

ABSOLUTE PROOF bots cant beat humans

end of discussion :P

Wyvver, You could probably get away with just tuning a basic bot for specific holes in it's performance. If you took the 19 hands as a base and found that it was always making the same sort of error in the same sort of situation, then adding logic to change the leak _should_ be possible.

True, it would not be perfect and this would also lead it to be more identifiable, but it would work.

If you had a DB at the back end that recorded results and introduced logic in the same vein that HoH encourages (raise x%, call y% etc) then it would be a lot harder to detect and would [probably] be more profitable in the long run.

J

Originally Posted by jackvance

Originally Posted by Bailey

At which point.. its hardly following a system such as a play chart

Like I said earlier, I'm not saying this to defend 19 hands, just that bots can own online poker.

The original statement was made in a thread about NLHE cash games. Please give me evidence of a bots beating NL ring games.

Originally Posted by jackvance

Originally Posted by Bailey

At which point.. its hardly following a system such as a play chart

Like I said earlier, I'm not saying this to defend 19 hands, just that bots can own online poker.

I totally agree, at least at the low levels.

What was your thesis in? You can PM if you like.

Originally Posted by DaNutsInYoEye

The original statement was made in a thread about NLHE cash games. Please give me evidence of a bots beating NL ring games.

Well, I researched the topic a bit earlier today.. you can buy the bot for $100/year but it's a blank slate. You have to program the AI in a C-like manner. The market value of such AI scripts is 5000 times the $ it can make per day per table. So if you can write a bot that can make $2/day/table, it'd be worth $10k. Basically, if you can make such a winning bot depends entirely on how good you can code one.

What was your thesis in? You can PM if you like.

It's still going on. The subject is the japanese board game go. Go, due to its complexity (19x19 board, as opposed to for example chess which is 8x8), doesn't have very good computer AI yet. So we're working on ways to speed up the computations. More specifically, I'm using relevance zones for this effect.

This has opened my eyes, there is another read here
http://www.wired.com/wired/archive/13.09/pokerbots.html
and there is a forum dedicated to bots with currently 1885 members
(winholdem)

Would it be worthwhile when at a table to post something like "bot count all say hi if your not a bot". I am a winning player but dont like the thought of playing bots which are evolving while i am staying relatively still on a learning curve, and remember bots dont get tired or go on tilt.

Yeah it was quite a shocker for me too that there is a whole community into this kind of thing.

But I'm a winning player, so I don't really have anything to complain about in online poker, heh..

and remember bots dont get tired or go on tilt.

..but they get caught though

I read a post about a guy who opened his bot for 2 seconds while his Party Poker software was running.. and he had his $500 something BR confiscated.

But I'm a winning player, so I don't really have anything to complain about in online poker, heh..

thats not what i said

no, it's what I said..

in that case i agree

Originally Posted by jackvance

What was your thesis in? You can PM if you like.

It's still going on. The subject is the japanese board game go. Go, due to its complexity (19x19 board, as opposed to for example chess which is 8x8), doesn't have very good computer AI yet. So we're working on ways to speed up the computations. More specifically, I'm using relevance zones for this effect.

Difficult problem. Very cool.

My course work has danced around AI. I've only had an introductory course in AI so I'm not familiar with relevance zones specifically.

Good luck

...