Most poker players recall the crazy Brad Booth all-in , bluff shove on Phil Ivey on high stakes poker. In reviewing that hand, Booth said he was getting lifetime implied odds for future good hands, since he knew the hand was going to be televised and people would think he was crazy.

Assuming that there is such a value to these types of plays, does game theory probability theory factor it in? If so, how?

No. Because game theory does not take into account previous hands and it assumes you don't either.

Isn't that a flaw? Well, not necessary a flaw, but a limitation rather. If game theory, as it regards to poker at least, could account for the ways previous hands were played and what these plays mean regarding opponent strategy, then it could only benefit itself, right?

Now if that is true indeed, which I think it is. Why does game theory do not account that?

Because it would then set itself up to possiblybe double crossed.

Game Theoretical Optimal doesn't mean what people usually think it means. It does not maximally exploit a given opponent. Instead, it is a way to play where it is not possible to get exploited yourself. So, it doesn't matter what your opponent does or what cards he gets dealt, he won't be able to gain an advantage over you that isn't inherent in the situation already.

So because of this, it doesn't matter what your opponent has done, it's irrelevant.

You could make more money by deviating from GTO and exploiting your opponent. Doing so, of course, opens you up for being exploited yourself. There is a terminology sticky in this forum that might help you out some.

Game theory optimal makes a lot of sense in the realm of bots or AI approach to the game. But when we are talking about poker played by humans, isn't it totally silly that "gto" can not realize that there is a lot more value to be taken out of the spots by just perceiving and accounting for the opponent strategy flaws?

What I mean is that when a player (human, bot or alien) which is capable of playing the optimal strategy indeed, well, seems quite obvious to me that one could easily perceive flaws and inconsistency upon different opponent's strategy's and therefore unilaterally increase his expectation to the maximum when facing an opponent who do not play optimally himself.

Would be non sense to play "optimally" if there is a strategy that makes more value. And the poker theory optimal should account for that. In that sense, playing GTO would only make sense when facing a GTO opponent.

OK? I'm not telling you what you should do. I'm telling you what the term Game Theory Optimal means.

In terms of a GTO solution, metagame has no value. So that's the answer to the OP, unless he'd like to amend his question in some way.

I'm not asking what I should do.

I'm just saying that GTO is not really the very best solution to the game if it can't perceive and account for what is the opponent expected strategy. Instead, the perfect player would definitely deviate from GTO to go for the strategy that has the greatest value against his opposition every time one has reliable data to do so, and play the GTO strategy every time it has insufficient data.

In other words the best solution to the game is a GTO strategy that is able to deviate from itself to maximize value when flaws on opposition strategy are known. And if a GTO solution can't do that, then it is not really optimal in stricto sensu.

And although I do understand Sklansky statement that GTO do not account previous plays because it would open the possibility to get double crossed, I can't see how that is enough reason to not really consider it. Just because there is a possibility to get rolled over, it does not mean the risk is not worth it in terms of value.

I was not thinking about it as metagame, but rather as assumptions regarding opponent strategy given his previous plays. And I've picked on what OP dropped to go on on some thoughts about it...

On 2p2 people always stay very attached to what stated on OP? In my experience on forums, OP is just a start and quite often a thread go pretty far from the original OP scope. That can be bad or good, depending on how interesting topics are considered by viewers.

No one who studies game theory has ever denied that if you are better than your opponent, you should deviate from GTO. That's completely obvious if you spend more than a few minutes thinking about it. Knowing what GTO is is still useful in this case, though, because in order to spot deviation, you have to know what a person is deviating from. And in order to understand your own exposure in deviating, you have to know what you're deviating from.

Don't get hung up on the word "optimal" here. It has a specific meaning in this context, that does not match it's general usage. Also, it's usage in poker doesn't *necessarily* match it's usage in the wider world of game theory, although it is used that way in general game theory sometimes also. Optimal in GTO does not mean "the best"

There's more to it than meta. If you view every single action as 2+2 = 4 you do not hit anywhere near game theory optimal play, and meta play one of many roles in this process. I could go into detail, but I'm not in the mood, so no.

Can't game theory subsume some of this meta game stuff by analyzing this as a repeated game? just because the state space increases in dimension doesnt mean we cant account for it...just may be that we need extra tools to attempt a solution (subgame perfect eq. maybe?) I dunno...

I am playing nowhere close the best poker with gto alone.

Trying to trick others to adjust, looks like a good idea, starting by adjusting to the image one might have. Tricking actively is a known starategy and works vs unaware and sometimes vs most players. It is up to one how well one does that. Adjusting, setting up, exploiting.

The ranges can be different vs some and sometimes with up to no statistical tilt nor lose. So, something can be changed and it doesnt matter up to at all if it doesnt work. But if it works, one does better than gto.

As others have said GTO means unexploitable and impossible to beat. It does not mean wins the most in any situation.

Interesting. The spoil is for other posters, not RoyalRumble.

What did you mean by that DS?

Also, doesn't account for live or online tells...and reverse tells...leveling of tells, etc.

There's still a lot of stuff that doesn't seem quantifiable at all by mathematics that affect the game.

Of course , how do you mathematically model if someone is tilted and about to blow up

1. Probabilities
2. Assumptions
3. Meta

All that online intuition crap is non-sense and everyone knows it.

I presume he means that as soon as you deviate from GTO to target your opponent's tendencies you open yourself up to the possibility your opponent may change their tendencies to counterexploit you... set you up and double-cross you.

That's why GTO strategy is static.

.

So, what's the final analysis of the usefulness of GTO?

If people agree that there are non-quantifiable aspects of the game that GTO cannot factor in, then how do people make decisions about whether to deviate from GTO or not to try and exploit/take advantage of a weakness they perceive in an opponent that might not be factored into GTO?

Do the best players always stick to GTO or do they selectively choose spots to deviate from it, where they think it gives them advantage (based on their reads of things like tilt, tells, leveling, game flow dynamics, etc.)?

It sounds like what you guys are saying is that GTO is kind of a baseline strategy for not being exploitable, but not really the best way of playing to win optimally if you have good reads of situations?? Is that correct?

yes

Remember the T in GTO stands for "theoretically". Reality is often very different.