I've never seen a single person on the internet say you should only try to play GTO

Nope, Game theorists aren't the ones who can play closest to GTO, the same way a given sport coach isn't necessarily one of the best players. People who grind solvers for hours on end every day are, and they are the ones crushing high stakes at Vegas and everywhere else.

Here is one quote from a post in this thread, on page two: "Playing anywhere near decent GTO will make you a huge winner in nearly every poker game.
In and by itself, no questions asked".

However, I suspect my problem is the following: since I do not understand how GTO works, I might take statements like that at face value while they might not be meant to be interpreted literally, and that's why it does not make sense to me. I have posted a new thread under "Beginner Questions" entitled "Question about GTO". When you have the time, could you please read it and tell me what you think? It would be very helpful to me, and I thank you in advance.

That quote is true. He is not saying you should do that.
Humans can't play close to GTO and never will

There is the descriptive statement that GTO beats every human to a pulp, that's just a fact.
And then there is the prescriptive statement that you should only focus on playing as close to GTO as you can and ignore everything else, which I've never seen anyone say.

Well that was the problem. I understood the above quote and other similar ones as a claim that playing close to GTO is perfectly achievable and assumed there were players doing it already.

When you say GTO beats every human to a pulp, do you mean a computer with software that has incorporated only the GTO library or the GTO library on top of all the other poker moves available?

The "GTO library" alone would beat every single person.

I don't understand what you mean by other moves available

Sorry, never mind. I kind of remembered that you said in my beginners thread: "GTO: Every aspect of a strategy that uses its range composition to deceive Villain from knowing the hand we're holding.
Things like using the same sizing with certain parts of your betting range on a given spot, being careful to neither bluff too much nor too little, being careful to not station nor overfold... ", and I was thinking the computer cannot win by doing only that, but now when I went back, I read your statement on top of that and got it, the library contains all the moves from the simulations, not just the ones categorized as GTO-type. Thank you for all your help!

i post this in basically every thread but think of GTO in poker like Chess rather than RPS. The difference is chess doesn't have mixed frequencies, however the similarities are that it is extremely complicated, too complicated for any human, but anyone who can play close to it will crush

Wouldn't chess GTO still have different openings available?
Could call that mixed strategy.

Probably no. I imagine as computers get better it would eventually converge into one opening having slightly higher EV than others.

Even if there were multiple max EV openings, it wouldn't really be a mixed strategy in the same sense that poker is, because every game in chess is completely independent while in poker, ranges are determined by how you play multiple hands

There's no "higher EV" in chess GTO in that sense, it's purely deterministic. Any given position either results in a win, a loss or a draw for white if both players play perfectly from that point forward. It's just one single outcome 100% of the time.

If there is any opening that can force a win with white then all openings that can force a win with white (against a perfect opponent) are also GTO and any mixture of these openings is also GTO.

Ya there are no are mixed strategies unless its like "ok if i choose option A, its a draw, and if i choose option B, its a draw" and the computer random picks one in that case

There is no imperfect information like in poker so it is unnecessary to have mixed strategy

I know very little of chess, I'm probably like a 900 elo player (since I basically almost never play)

I do like reading about it though, currently reading a biography on Magnus Carlsen.

From what I understand there are hundreds if not thousands of openings available to the currently top rated human players.

I'd imagine at least a few of those, if not quite a lot, would be capable of forcing a win with white in a GTO match.

Isn't GTO in poker mixing just as a result of each specific hands EV?

I know GTO doesn't try to balance.
The balancing (mixing?) just results from playing each hand at it's maximum EV, knowing your opponent knows exactly how you will play each hand.

I think chess GTO would be similar, as in a lot of equal options being chosen at random, because they all lead to the same result.
But like I said, I'm bad at chess.

I think the chess comparison is kind of flawed as we are comparing a perfect information game vs imperfect information. Mixing is most likely required in poker GTO, in chess it is definitely purely optional.

It does, that's why specific hands have specific frequencies for same EV actions. If not, frequency wouldn't matter

Quote from Modern Poker Theory:

Maybe the most profitable way to play a hand is to do something 60% of the time and something else 40% of the time, each resulting in the same EV. You do it not to be balanced but to not be exploitable.

GTO isn't trying to be non exploitable either
It's trying to maximally exploit it's opponent, but in practice that ends up pretty much exactly the same.

That's exactly what I meant in my previous post though.
GTO isn't trying to balance, the balance is just a result of trying to maximize value with each individual hand against a GTO opponent.

I would be careful with trying to see it that way. You will eventually hit a wall, in my opinion. GTO assumes the opponent is also GTO. If this prior is not true, then your statement becomes false. If the opponent is way off, then it can be said that GTO is “trying” to be balanced, and unexploitable, with the bonus that the opponent can not even examine your strategy and deduce a counter strategy, other than “git gud”.

Yeah true, but not because GTO is trying to be balanced.
GTO is balanced and unexploitable as a result of trying to maximally exploit a GTO opponent.

This is not my view btw, it's literally stated like that in Modern Poker Theory.
Unless I'm interpreting it wrong.

Note that when I say GTO I mean NE, but I think you got that.

But yeah I agree, against a non-GTO opponent, GTO is balanced and unexploitable.
Well it's also balanced and unexploitable against a GTO opponent for that matter :P

Just to be clear, I disagree with the author on that point. So long as the GTO is unchanged, it does not matter how you attribute intent to the strategy. However, since GTO does not change for any reason no matter the opponent, I do not see how it is rational to attribute the intent of GTO to maximally expoit anything.

GTO just is GTO.

Re:chess. I am an international master with 30 years of experience in trying to improve at chess while making a living with poker.

Well, the top engines are so good nowadays that the openings have evolved a lot even in the past 10 years. Knowledge in every sensible opening has grown so big that I'd describe it as a bluff if someone plays an opening that leads to a clearly worse position, assuming best play by both sides. Chess is an extremely drawish game (because it's a draw, duh) and it usually takes several small mistakes or one big mistake to actually lose. Humans lose because they make a lot of mistakes, same as in poker.

All in all, I'd say a tiny portion of modern openings might be winning for white with best play, but we are talking about 1% or so. Of course also depends a bit on what we define as an opening. But lets say positions after 12 moves that have arisen more than twice in the past 3 years between top players.

Yes, the starting position is a draw, but even if you had a perfect solution, all moves leading to a draw are not equal. Some drawn positions are such that one side needs to make 10 exact moves in a row to get the draw while the other side has no legal moves that lose the game. Thus I think it's better to assess chess positions not with "white wins, draw, black wins" but with something that describes the equity both players have. For humans, this is, perfect play vs perfect play is not interesting as it's always a draw.

Agreed. In chess top players mix, because they don't want to be predictable, but instead test the opponent's knowledge in a different branch of the game tree. Thus mixing is great in practice, but has no meaning in theory.

So all the top AI's are just drawing vs each other mostly?
You're saying a perfect GTO strategy (most likely?) wouldn't result in a win for white, but a draw?

They have stopped AI vs AI matches starting from the initial position, years ago. Probably two reasons:

1. no need for a pre-built opening library which is a bit problematic as it's human built anyway. I guess they could force the engine to build the opening library by itself or ban any use of prior opening knowledge.
2. probably the percentage of draws would be huge, like 99%.

So AI vs AI competitions are today organized so that they play from a few pre-determined positions (with reversed colors) which sometimes favor white greatly, just to get some wins.

I am totally convinced chess is a draw. Or let's formulate it like this: everything I've seen indicates that perfect play leads to a draw, and that there are several ways for black to achieve it.