Of course.

Of course that is also whitewashing over what it takes for a bot to be "properly nodelocked". That is literally what some people call "playing poker".

The solvers have been very successful in solving certain simple problems, HU pots between good players, by running huge numbers of simulations. That is useful when applied to higher stakes cash games. Although playing GTO is still not optimal in those cases, there are many changes in typical play of good players due to solvers. Solvers are not so helpful with multiway pots in tough games, and don't translate completely to tournament play at different stages. No one has been able to develop anything simulating a game with a bunch of bad lose passive players. Anyone claiming that is an easy problem does not know what he is talking about.

There is an easy solution that ignores the variables you’re listing. The computer gives us those solutions. And after studying pluribus for hours and hours you can see the MW adjustments the solvers make

If it is such as easy solution, why has no one developed it?

The information is out there you just need to find it. Go study pluribus. You can pay upswing $500 or so to do so.

GTO will win in low stakes games. But it's comically sub optimal. The entire strategy is based on not being exploited in a player pool where 99 percent of players aren't trying to exploit you.

So basically the authors don’t have any training in poker gto and are lying or delusional when they make claims about gto

You can’t discuss or compare something you know nothing about, certainly not on a paid and published level. I mean you can discuss it, but what you say has no validity besides guesswork

I think it's hilarious that posters think they are so smart, smarter than the authors, think the authors suck, but continue to post in this thread to l so they can feel good about themselves

If you don't want to read the book, don't

But acting like a pompous douche is not going to stop normal level headed people like me from buying it

Upswing doesn't have any content on Pluribus that I can find. They offer a course where Polk analyzes a match against Libratus, not Pluribus. Libratus is software that played heads-up poker, not multi-way. So what are you referring to?

After studying them both you can see there isn’t much of a difference

So they need Doug Polk to teach them gto?

What if the dude was never born?

I don't think the authors claim to have expertise on GTO. Not sure of the relevance of GTO to 1/3NL. That is maybe partly why they wrote a low stakes book. I am they and most people here know something about GTO. GTO means information based on solvers, not general game theory.

There's not much of a difference between multi-way and heads-up? There's tons of situations where you'd fold multi-way but play heads-up. A heads-up solver isn't going to teach you that.

You sound like a complete fraud. And given the number of silly posts you make in a day, I doubt you'd have time to study solver output or understand them even if so inclined. I don't know what axe you need to grind, but please stop embarrassing yourself and wasting everyone's time.

So you’ve spent the money on Upswings libratus package? And you examined the EV of diff options HU vs MW and you found there to be huge differences ?

Cool. I didn’t find that.

GTO is actually exploitative. People just think it defensive because when you have two solvers play the best either one can do is break even. If you force one solver to make mistakes the other will absolutely adjust to exploit it.

You are wasting your money on GTO packages when you play 1/3.

There is maybe a slight difference between the solvers' assumptions and 5-way to the flop at live low-stakes.

I have no reason to purchase such a course. But you pretend you have, yet you didn't seem to know the Upswing course wasn't based on Pluribus (and who says such a course is the final word?), and incorrectly cited it when trying to make a point. You continually put forth your supposed credentials as some sort of proof you know what you're talking about. Yet you never explain why your view is correct. You just tell other they're wrong, and cite these fake credentials. No one believes you have access to Pluribus. No one believes you could glean anything out of Pluribus hand histories even if you studied them. No one believes you can beat 1/3 NL. The S&M book has many examples explaining their point of view. No such examples from you demonstrating otherwise. And their credentials are well know in the poker community. And their education background is well known. That doesn't prove their recommendations are correct. But criticizing those recommendations without backing up your opinions with explanations beyond your fake credentials is laughable. You're obviously just a troll who gets off tearing down the work of others without contributing anything of worth of your own.

Have I made my point?

Mic drop

Your interchanging solver output with GTO. They're not the same thing. No one knows what the GTO strategy is to NL poker, but the GTO strategy doesn't change based on the strategies of the players. Solvers are a tool we use to try and come close to what the GTO strategy is. But solvers can also be used to find exploitable strategies based on assumptions.

GTO is not exploitable by definition. Solvers, on the other hand, are trying to exploit by design. If, and that's a huge "if", you properly define the ranges of the players and properly node lock how the players will act/react in the game tree, then you can get a strategy that exploits the player's weaknesses (as defined). But if the ranges or node locks are off, so is the strategy, and that strategy is exploitable by the very players that strategy was trying to exploit.

The only time the GTO strategy would change would be if you changed the rules of the game. Not if you changed the strategies used by the players.

Have you read the book in OP? It is a lot cheaper than that Upswing package.

I do use the 2 terms interchangeably but even a static gto balanced strat with zero adjustments. I would argue it more than most humans. While gto isn't making any effort to exploit its also never being exploited and will naturally make money off of mistakes. On a recent episode of solve Berkey and turtle said a gto strat would crush a low stakes live game for at least 30xbb per100.

Part of the reason Pointless can't beat 1/3 is he is spending money on fancy GTO packages when it would make sense only to study the basics of GTO. Spending $500 on some GTO information would only be appropriate for a winning mid to high stakes player. I would suggest he read the book he is criticizing and other books and videos geared towards low stakes games.

I read the book in OP and made some criticisms of it based on actually reading it. I would be interested if Pointless had read it and had comments on specific things in the book.

30 bb/100? Sounds optimistic, especially for a full ring game with a typical rake. If you get 40 hands per hour, that's 12 BB per hour, which is very good. I don't know what the best exploitative players make at the lowest stakes, but it can't be that much more than that. But I agree, GTO would fair better than most players normal strategies would, if there was such a thing. But no one is playing GTO, since it isn't solved and it's too complex even if it was. That said, using solvers to improve a player's game, even with incorrect assumptions regarding ranges and player tendencies, will probably make that player better faster than that player would be left to his/her organic learning curve and other learning methods. It's definitely a good investment for a serious player.

If I am in a ring game where half the players play only 10% of their hands and the other half play 60%, and I know who is, who I will gladly crossbook with a pure GTO bot (that will play the same way against both) in that game.

That number seems a little high to me.

There's also confusion here. What GTO does is to maximize your mimnimum expectation. This is known as the Maximin Principle, and you can read about it here:

https://en.wikipedia.org/wiki/Minimax

"When dealing with gains, it is referred to as "maximin" – to maximize the minimum gain."

And the picture on this page will show what a saddle point looks like:

https://en.wikipedia.org/wiki/Minimax_theorem

So the point to be made here is that assuming the 30 bb estimate is correct, there should now be plenty of spots where correct exploitive play would increase that number.

Mason