Is there a perfect NLHE Bot? Could there be?

[vypremik]

[ QUOTE ]
"Reading" people is nothing more than recognizing paterns in their body language and betting patterns. A computer could theoretically pick up on patterns far more subtle and complex than the human brain could process. Even a player like Phil Ivey, who is probably unreadable to any living person, would have detectable traits and could be defeated. The neural net would be infinitely adaptable to any level or style of play.

[/ QUOTE ]

In my opinion, this does not make the bots better, it makes them more beatable. Although I do not claim that there are any player without tells (I think we all have something), the professionals actually take advantage of them and can use their own tells against you - or even better, against a bot.

For arguments sake, let's say Phil Ivey shuffled his chips with his left hand for three seconds and bet half the pot when he was bluffing. He takes down a few hands this way and shows them every time.

The bot now has made the connection that this behavior is a bluff. So the next time he is sitting on a monster hand, Phil does exactly the same thing and the bot thinks, "Aha! a bluff!" and puts in a large bet to try to push him off the hand.

I think this is a real problem with bots. How do they know when you change things up and when keep them the same? Many pros will be able to overload the bot with false information and take advantage of any tells system that is built into it.

[The Zerojack]

I don't see why not. People are organic neural nets and some irrelevant but fun stuff added. What I'm missing as a factor in this discussion is not if or how it was possible. That's obviously the focal point of this debate.

I tend to look at "When?" Not to scope out into a picture so big it turns this question into a pale suited dot...

A correctly designed system has room to improve on its mistakes. It adapts and learns from errors. It learns to articulate its aims and move towards accomplishing them.

If this sounds familiar, we've all been there. This describes a child moving out of infancy into childhood, repeating itself. A poker player learning the basics, moderate-level nuances, then master-class skills that set him/her apart. Why wouldn't a recursive intelligence with a vast, correctly-functioning memory improve at a faster pace than a human player?

Issue: Would it beat Ivey? I'd consider it a breakthrough for computing not if it beat Ivey consistently, but that it earnt his respect enough that he was willing to compete with it on equal footing.

Kasparov/DB was simply a demostration of the standard of raw computing power against human artistry. That point's been proven. Now thus follows in evolution: Can artistry be repeated in a non-biological intelligence?

No, I don't know either, and I'm straddling the fence on this issue, but I'd like to find out.

[StregaChess]

First there was an earlier post that is incorrect they don't play expert chess. Experts suck, experts I can beat in my sleep...
They play Grandmaster Chess at the world championship level.
That's like Phil Ivey and 2x.



[ QUOTE ]

There is a big difference between chess and poker. In chess there is nothing hidden. A computer can see the board and compute all of the info just as a human would, although much faster and more thoroughly than most of us can.

[/ QUOTE ]


I regret to inform you that this is not that simple.

Chess has too many decision trees for a computer to see everything in a brute force sort of way.

Often they have an algorithm that only explores a list of candidate moves based on the knowledge that computer has been taught.
Its not all brute force programming but brute force does help….

Instead its more like…..
(computer thinking here….)

I’ve know that endgames where a bishop is on the same color as its locked pawns is weak against a knight where both sides only have pawns on one side of the board Even if the piece count in the position is considered equal. This middle game has a good chance of simplifying down to that type of endgame. Do I have a forced win from this position ? (NO!)
OK I most likely should avoid this position and line as it will lead me to an unfavorable endgame.

This other line will allow me to reach a position where I’m a pawn down, however we both have rooks, but my rook will be active, his rook won’t.
Let’s assign a value of 2 for the active rook, OK…. He’s up a pawn (+1) but I’ve got the more active rook (+2), let’s go for this line and reject the dead even line…. as in this line I'm +1....

Ok its that ugly…

[Buccaneer]

If they make a bot that can play the perfect game will it ever go on tilt? This is a serious question. If the bot was adaptive enough to beat the best human players would it not blow a fuse when one of the stupid humans played a 72o and hit a hand? I would think that true intelligence would be unable to avoid tilt just like humans are unable to avoid tilt.

[The Zerojack]

I would think a correctly designed pseudoint would react to the 72o situation as an aberration. If it was sufficiently developed, and capable of self-programming, it is a virtual certainity it would come to the same conclusion after enough processing time's devoted to it, that it is a move designed to "tilt". Would it react wrongly? The first time, of course. If you've had your bullets cracked by that hand, you've tilted. But over time, you've learnt to factor that into your whole poker-playing arsenal. "This will happen, but it is unlikely to happen, so it's not a major concern, merely an overbetting caution." <--- illustrating what, IMHO, a recursive would apply to this situation. That said, they'd assign it a probability value of nonzero, and if they were permitted to remember its opponents moves (Permitted to remember? Jove, that's telling of a biological bias, isn't it? Hums Thus Began Zarathusra and looks for a rock) It'd be likely to assign a probability flag and some processing time for that situation.

Asleep yet? Your pardon. It's just that I find it quite fascinating we're at a point where we're discussing the qualities of emergement intelligences... In... NL... Poker... It kicks ass, man.