GTO not even close to optimal

[BelFish]

I came up with another illustrative example with blackjack:

For example, the rules are such that you don't see any of the dealer's cards. But the dealer announces whether his first card is higher than the nine, or there is a card up to and including the nine. Then according to the GTO, it would be correct for the player to stop taking cards at 15-16 points, when the dealer announced that his first card was not higher than 9. And you need to take another card if he announced the opposite.

But if the dealer gives out by some action, for example, a smile, that his card is not from 2 to 6, with which he will bust more often, but from 7 to 9, then it will be more profitable to deviate from GTO and to exploit this leak of the dealer ))

 

[German629]

GTO not even close to optimal?.. I'm fully agree with this affirmative! I think, all is more difficultly and sophisticatedly...
And therefore, when I have a good concentration in Game, I more believe the my notes and intuition..,and then I more often winning!..

 

[Axmanace]

Ed

But if the dealer gives out by some action, for example, a smile, that his card is not from 2 to 6, with which he will bust more often, but from 7 to 9, then it will be more profitable to deviate from GTO and to exploit this leak of the dealer ))

Wouldn’t this follow GTO theory though?

If you can narrow the hand down to a specific set of cards GTO solvers would also adjust for that.

Intuition in this case is simply attempting to mentally change the ranges as needed.

 

[TeUnit]

Exploit them when you can, GTO them when you cant.

 

[Rockyfour]

I came up with another illustrative example with blackjack:

For example, the rules are such that you don't see any of the dealer's cards. But the dealer announces whether his first card is higher than the nine, or there is a card up to and including the nine. Then according to the GTO, it would be correct for the player to stop taking cards at 15-16 points, when the dealer announced that his first card was not higher than 9. And you need to take another card if he announced the opposite.

But if the dealer gives out by some action, for example, a smile, that his card is not from 2 to 6, with which he will bust more often, but from 7 to 9, then it will be more profitable to deviate from GTO and to exploit this leak of the dealer ))

This is my friend Belfish, he is a cat.

 

[Axmanace]

This is my friend Belfish, he is a cat.

The thing I don’t get - if you picked up a read to put a person on a more specific hand… wouldn’t that modify the GTO solver to do exactly what your “intuition” told you to do?

Most trainers provide % ranges for your opponents.

I’d argue if you have improper inputs for GTO - it’s not actually GTO.

In your examples you didn’t actually deviate from GTO theory - you simply changed the inputs.

 

[BelFish]

The thing I don’t get - if you picked up a read to put a person on a more specific hand… wouldn’t that modify the GTO solver to do exactly what your “intuition” told you to do?

If we know that a certain player is a nit and that he pushes only AA/KK, then unlike us, the solver does "not know" this and according to the GTO there will be calls with some kind of range in accordance with the Nash equilibrium. And if we fold against this nit, we will exploit him ))

Chips saved are also chips won...

 

[Axmanace]

If we know that a certain player is a nit and that he pushes only AA/KK, then unlike us, the solver does "not know" this and according to the GTO there will be calls with some kind of range in accordance with the Nash equilibrium. And if we fold against this nit, we will exploit him ))

Chips saved are also chips won...

You’re the one they provides the solver with hand ranges…

So why wouldn’t it know this?

 

[Rockyfour]

You’re the one they provides the solver with hand ranges…

So why wouldn’t it know this?

Even if you put a range of hands for your opponent his actual range is going to be different than what the solver thinks his range is gonna be. Like if you put in your opponents VPIP of 22%, it's going to make a lot of assumptions and rank hand strength differently than a computer.

So for example, A high suit plus any pair = a raise from any position w/ the solver (6max Omaha), but most people aren't going to be playing A229. However a hand that is bad like 3-4-5-6 might be valued higher to a human than the solver because the human is over valuing the ability that hand has to make straights.

Anyways, it's not really possible to input your opponents true ranges into the solver. Especially because they are dynamically changing more than likely, and also I believe you would have to create a sim and then run it for a few weeks to figure out all the ways you can exploit that very specific play style, for him to turn around and change his play style after he figures out your exploiting TF outta him.

It's just not realistic atm. Maybe with AI. Solver also doesn't value timing tells.

 

[BelFish]

You’re the one they provides the solver with hand ranges…

So why wouldn’t it know this?

I'm not sure about the solver, probably you can set ranges there, but as for GTO, then if this is really GTO, and not an exploitative strategy, then the ranges for opponents are not set based on their statistics, but ranges that give Nash equilibrium.

As soon as you start substituting known data on the opponent into the solver, GTO ends and exploitation begins ))

 

[Rockyfour]

I'm not sure about the solver, probably you can set ranges there, but as for GTO, then if this is really GTO, and not an exploitative strategy, then the ranges for opponents are not set based on their statistics, but ranges that give Nash equilibrium.

As soon as you start substituting known data on the opponent into the solver, GTO ends and exploitation begins ))

i think he is wondering about having a computer get the optimal results to counter an opponents play style. Like could the computer output the true optimal actions versus an opponents leaks.

And I am no expert, but it sounds like it'd be too confusing and complicated to do.

 

[quick]

I think I'll defer to Phil Galfond on the matter...

"You don't make money based on how close you get to executing a GTO strategy.

You make it by fully leveraging your opponents' weaknesses. Study solvers, by all means, but don’t lose sight of where you get your edge".

He's spot on. And it's ironic because based on his correct statement, GTO strategy and solver study are just another tool in one's toolkit to play exploitative poker. At the core of the game, exploitative strategy is the strategy. You profit off others' mistakes and misplays , the solvers and GTO strategy are only a mathematical guide on the optimal play for the cards, stacks, and betting action, and are by nature defensive (limiting you being exploited and you making mistakes) not offensive.

 

[BelFish]

GTO is a set of actions against ideal opponents. Well, or against the unknown.

But there are few ideal opponents, so GTO against most opponents will give a shortfall in profit compared to exploitative strategies...

 

[kitchy65]

GTO is a set of actions against ideal opponents. Well, or against the unknown.

But there are few ideal opponents, so GTO against most opponents will give a shortfall in profit compared to exploitative strategies...

They say the best sell for GTO is you cannot be exploited. You said this yourself in another thread too....technically correct given the assumption you are playing an honest opponent.

But this is poker and your opponent isn't honest.

GTO has a crap preflop strategy....meet resistance to your 3-bet and you fold way too many hands.

I'm guessing this is the 'can't be exploited' claim in action and you certainly couldn't disagree with it.. but you must ignore that you're passing chips to opponents uncontested and believe you did the right thing.

Disclaimer...I'm speaking from a position of ignorance.

Here's my example:

Hero SB with QJs
Villain MP with ??... 2x the pot

Hero 3x pot
Villain comes over the top with 4-bet

Hero folds 100% of the time.

Villain is a maniac and was robbing you with Q9o

 

[jordanbillie]

Here's my example:

Hero SB with QJs
Villain MP with ??... 2x the pot

Hero 3x pot
Villain comes over the top with 4-bet

Hero folds 100% of the time.

Villain is a maniac and was robbing you with Q9o

It's pretty rare to get a 70/30 with QJ, so that example is an outlier.

I simply wouldn't open any hand I wouldn't be willing to jam against that opponent. Wait for a premium hand and win a lot of chips.

 

[Luvepoker]

following this thread had been interesting. Many people have many strong ideas on it and many don't honestly know GTO at all and that's fine. The thing is GTO is a theory that many do use with success. The problem is many look at it as this is what you must do in any situation and that's it. That is not true at all.

GTO is assumes your playing against another player who is playing GTO. Nobody plays GTO perfectly at any level. GTO is useful but not a road map to just winning. Its really more of a guide on best options in a perfect world.

 

[Pulse_69]

I think that GTO is a theory and initial practice, in a pro it does not have any great importance, it is the basis. Then psychology and strategy.

 

[John A]

@John A, you're a coach, right? Your job and livelihood is NOT threatened by solvers buddy. Your comments in this thread is borderline misleading. Don't think everybody will just keep quiet because you have a few badges on your avatar accumulated since 2012.

lol, no... I don't coach anymore, except for the occasional celebrity or high stakes player that asks. Maybe you should look up my name before just making random attacks.

My comments are accurate. I'm sorry that you don't understand what GTO is, or why it's not helpful for micro and small stakes players. It's not a baseline to use either. It's something to consider as a THEORY, but when your opponents are making tons of mistakes, and leaving themselves exploitable, you should exploit them. Period. I can't tell you how many small stakes players I've seen make absurd plays in the name of "balance", that left tons of EV on the table.

 

[Axmanace]

lol, no... I don't coach anymore, except for the occasional celebrity or high stakes player that asks. Maybe you should look up my name before just making random attacks.

My comments are accurate. I'm sorry that you don't understand what GTO is, or why it's not helpful for micro and small stakes players. It's not a baseline to use either. It's something to consider as a THEORY, but when your opponents are making tons of mistakes, and leaving themselves exploitable, you should exploit them. Period. I can't tell you how many small stakes players I've seen make absurd plays in the name of "balance", that left tons of EV on the table.

Can’t you adjust GTO for theory for unbalanced players?

Pretty sure the advance GTO solvers allow full flexibility in terms of opponents ranges.

 

[John A]

Can’t you adjust GTO for theory for unbalanced players?

Pretty sure the advance GTO solvers allow full flexibility in terms of opponents ranges.

You can node lock and add whatever ranges you want... but then, you might as well be using an equity calculator at that point basically.

 

[Rockyfour]

Told you lol (Pro Poker Tools)

 

[Axmanace]

Told you lol (Pro Poker Tools)

Pro poker tools is based on GTO theory.

It is GTO.

 

[Rockyfour]

Pro poker tools is based on GTO theory.

It is GTO.

It's not it's just an equity calculator.

 

[BurnTurnRun]

Gto is only optimal if you are playing against another gto player.

I find it quite laughable how players advocate shoving with trash SB into BB.

 

[John A]

Pro poker tools is based on GTO theory.

It is GTO.

No, it's an equity calculator.

And just for the record, not that it really matters I guess, but I actually created the first GTO calculator in this space over 12 years ago. Myself, and a friend, built a competition bot based on nash equilibrium to compete in the AAAI poker bot competitions and we won first place, beating teams from MS, google, etc... We turned that into Ace poker coach, and it was the first GTO advisor that would automatically analyze your played hands. BUT, and here's the funny thing, people at the time didn't understand why you'd choose to balance and vary your actions with different hands in your range. It was seen as unreliable because it was advising to Check X% of the time, bet Y% of the time, and so on. The poker community didn't get it, so it was literally the only piece of software I've developed that wasn't a hit. Ironically, it was way ahead of it's time.

You can still download it here, but it's not been in development for years.

Ace Poker Coach for Windows - Free download and software reviews - CNET Download

Download Ace Poker Coach latest version for Windows free to try. Ace Poker Coach latest update: September 15, 2016

download.cnet.com

So I understand GTO play pretty well. And I understand its limitations, and the group think of the poker community pretty well also.