OMG the meats have it.

I predict a long period where the top players can't figure out why they can't beat this "new breed" of hot shot young pro (which would be a bot assisted human). And a longer period where sites pretend like the problem doesn't exist. And then a very rapid dying of the games online as it becomes completely unbeatable at every level. And lots of jumping from high buildings.

Becausec, as far as I can understand, computers are strictly numbers-based. If you're completely randomizing play and displaying zero tendencies from limping to time-tells, than how could it develop and optimal approach to you?

If a computer doesn't have a strategy aside from what it develops to oppose you... then wouldn't the best strategy be to not have one as well? Wouldn't you be at an advantage by default because at least you'd be conscious enough to realize that the bot has no strategy, since you don't have one to develop itself off of, while the reverse can't be said?

I'm asking questions here, so someone please enlighten me. If a computer has a sample of one hand (or less) how can it adjust to exploit your strategy (that you are rapidly changing and randomizing)?

lol

If you are displaying absolutely zero tendencies, that means you are taking every possible action an equal % of the time regardless of your hand (which I guess is a tendency, but I think that's what you mean when you say zero tendencies).

Think about what your play would be like if you did that. You'd be open folding AA 1/3 of the time you get it, you'd be raising 72o 1/3 of the time, etc. You'd be playing so unbelievably poorly, you wouldn't stand a chance. The computer could still calculate that you're taking every action an equal % of the time with any two cards, and come up with a (fairly straight-forward) strategy to combat that.

Why is this thread even in NVG lol

-Mike

If you apply reductionism, you'll find it's only a matter of +EV or -EV decisions. Making a 52o shove when you sense your opponent can't call isn't creativity, it's being able to discern a +EV decision better than others (what we can train a computer to do even if the number of variables is huge). +EV encompasses your definition of creativity.

The computer can too, but like you said it would be foolish and a losing play, so why would it? You say the computer is bound to play optimally like it was a bad thing.

No it's not a bad thing at all... but if we know it only makes optimal plays we can counter it.

Also Is this a heads up bot,6max or FR.
I would think the heads up bot would be the best of them

In poker, you can check, bet, call or raise. In each situation, one of them is the most +EV. What is new or creative about each of these choice a computer can't see? I thought the 52o shove by Ivey was a good enough example: you can argue it's an unintuitive solution, but a computer built to play optimally would have seen it, given enough computing power to crunch all the variables and metagame between the players.

Without wanting to sound to too hurtful, I guess you haven't done much game theory. Search for the definition of optimal and see what is wrong with that statement.

I agree. I think it already exists:
http://forum.winholdem.net/wbb/

Not hurtful...bad guess
The computer isn't going to open with 106os under the gun because that isn't optimal.
It will never call if the pot odds are not there on it's draws...not optimal.
How does the computer make a hero call?Especially if it is not optimal?

I understand that playing an optimal style is one, that in theory can't be countered...but playing optimal vs one opponent is different than playing another.
Every style can be countered...every game has a leak. To dominate one player leaves yourself open to the other.

A bot will open fold KK on the button to a three-bet if this is the optimal play.
A bot will open 2-7 utg if everyone at the table is a nit.
A bot will call with ace high if it senses it beats his opponent's range.
A bot will call with bad odds if it thinks it has great implied odds and is playing deepstacked.

You have a very flawed vision of optimal play, and seem to equal it with boring and orthodox play, which is far from optimal nowadays (high stakes at least). Optimal play is what it says it is: the best response to an opponent's specific strategy, be it a donk or Ivey himself. (I'll drop the +EV because it seems to be confusing too )

Interesting...

"A bot will call with ace high if it senses it beats his opponent's range."
How does it sense?? also that doesn't sound optimal.

How does it know if it's playing Ivey or a donk.

The fact is a computer could never have the ability to produce something new through imaginative skill, whether a new solution to a problem, a new method or device, or a new artistic object or form.

The bot will always have a flaw...the programmers.
How can it play optimal if we always change gears?
How does it know when to change gears?

It doesn't sound optimal to call with the best hand when you beat your opponent's range? A computer wouldn't be able to tell a good player from a donk? Is this a level?

I didn't read the topic, but I looked at the poll results and 49.36% of you guys are ******ed.

The same way every human player assesses an opponent's range: it remembers previous play with that particular opponent. Why do you think everyone has a HUD nowadays? Because obviously a human isn't able to remember that much information by himself.

Huds are even limited on the amount of info they display. You can see the bot as a "superhud" which has access to everything about your play and is able to process the info/find patterns much more efficiently, without cognitive biases or tilt.

It knows how to change gear by analyzing your play and predicting it. You don't play a donk the same way you play a good players. How do you know which player is good and which isn't? By looking at their stats. Looking at his play without a HUD and remembering it still is stats, only in a different format.

If you're still asking if it can bluff, you haven't really understood our little optimal play and 52o shoves discussion.

yes but I doubt they will exist in this gen

Well now you're raising 74 and 64 and folding A7. Maybe we can raise T3 and fold KQ too. Kinda seems like this "tricking" the computer thing is more the player starting to play badly .... and I don't think that would work.

Chess has not been solved. I don't think you know what solved means.

Lots of people in this topic don't understand that it possible (in any game) to have a strategy which cannot be exploited by any other strategy. A bot could play in such a way that it doesn't need to even think about what his opponents strategy might be and be very profitable.

Partially solved, I will admit. The main point is that now bots dominate the game and are the highest ranked "players" in the world.

To beat the best players (what this thread is about), it needs to adjust.

No, it doesn't. That was my entire point. There exists a strategy which is impossible to beat, no matter what strategy you use against it. If the bot used this strategy, it doesnt matter what you did against it. At best you could draw against it. Once this strategy is found, poker will be classed as solved.

What if no such strategy exists in no limit Holdem? Not all games have an optimal strategy which works for every situation. Oh and you wouldn't need a bot to discover it btw. This is an entirely different discussion.

It does exist for Hold 'em, and it does exist for all games.
http://en.wikipedia.org/wiki/Nash_equilibrium

With "optimal" (for two-player games) defined as "does not (expect to) lose against any other strategy."

A NE would only be useful if the bot was playing against himself, and even then they would both lose money through rake, so it isn't really a winning strategy, as you're better off not playing at all.

And this is slightly different than what Karganeth mentioned: a strategy which doesn't need to adapt and can't lose against any other one. He is talking about a dominant strategy (http://en.wikipedia.org/wiki/Strategic_dominance) rather than a Nash equilibrium, which is obviously not always present in a game (it rarely is tbh, and certainly not in poker).