Poker theory has become centered around 'GTO.' Obviously these principles about optimal play have merit, they are grounded in the math/solvers/toy games (and they work).

In a RIO video i watched recently coach nick howard said: GTO offers perfect information/solutions, but trying to implement this information in-game is also causing harm and distortions. Which is definitely true in my opinion.

My questions to the forum are:

Where are the flaws in 'GTO' and the ways we are implementing GTO?

And what areas of poker theory are still unexplored and have room to grow and improve?

how is implementing a theory based strategy causing any harm or distortions?

It may be difficult to implement but that doesn't mean it causes harm

it just sacrifices EV if you have enough knowledge to deviate vs a suboptimal opponent

also as you said if you're trying to implement a highly mixed strategy you should pry look at a way to simplify it

Ceegee what your saying is true, generally there no real downside to using a strat based on "game theory" if used correctly.The fact that it is difficult to implement is why it can cause harm/distortions, and this is why im asking if there is a way to improve on the way we implement these ideas in-game.

for example MDF on paper is completely different than defense frequencies used in-game. in PIO a mixed strategy looks great, but finding the appropriate frequencies in-game is difficult (like brokenstars said). or a strategy may dictate that i should call the river because "im at the top of my range" but that doesn't necessarily mean i should.

and what areas can theory still grow? game theory cannot be the end all. in the early 2000's poker strategy was completely different, and look at what it has evolved into.

Everyone's already playing GTO. It looks different for everyone because the underlying game is different for everyone.

The biggest flaw with conventional gto wisdom is it assumes everyone's playing the same game.

The real wisdom is in understanding the specific game each of your opponents is playing.

gto is rather a means than an end. if you are trying to play gto instead of playing max-EV it is harmful to your game. in a matchup between reasonably strong players, max-EV becomes a race for the best application of gto, but taking it into soft games will not maximize your edge.

There is an optimal way to play poker, in the same way there is an optimal way to play chess, Go, or tic-tac-toe. In the early 2000s, we didn't have the "GTO" phrase, but people were obviously trying to improve their play and maximize their profits. They just happened to be much further from "optimal play" in 2006 than we are today. People have got better and better at playing the game, but - just as with chess or Go - humans can't play it as well as computers can. We just have to keep studying, using the technologies that now exist to help us.

Well... it kind of is, actually. If the strategy one plays is GTO, it is all over red over. Just use it and win (over the long run of course). And as has already been stated, only deviate from it when confident it is more +ev to deviate than to stay the course. But the mistake I see so often regarding GTO is, "GTO won't work in [insert player pool]". GTO wins in EVERY SCENARIO BUT ONE, the scenario where all other players are playing GTO. In that case, it is break even (minus the rake). But a GTO strat is NOT breakeven vs non-GTO strat players. It beats them. It beats them all. (P.s. To say GTO has "flaws" is somewhat incoherent unless you are talking about implementing them in non-GTO player pools, but even then I wouldn't say a GTO strategy has "flaws", I would just say it isn't the most +ev in a certain situation).

What do you think this statement means?

No one is going to become a worse player by studying theory. There's this huge misconception that it's theory based vs exploitative based. In order to exploit you must have some sort of clue as to how a spot should be played in theory. Self proclaimed "exploitative" players who don't even know what the theory looks like are just random button clickers.

At the end of the day the best theoretical players will also be the best exploitative players. It goes hand in hand. A player might be doing something apparently strange like betting 23% pot or leading turns and rivers or calling very wide in the BB. Today these are all standard plays but go back 4-5 years ago and you'd be called a fish for doing any of that stuff. Or consider that in 2011 calling a 3bet out of position was a mortal sin. You have to know the theory before you can even begin to say if a situation is exploitable or not.

Some think that gto = breakeven is the goal. I think this is indicative of a deep misunderstanding of where ev comes from. Once the blinds are posted, everyone is playing a positive sum game that distributes ev unevenly among the players depending on position and cards.

IME, I’ve noticed that people who focus on exploitative play only while ignoring GTO concepts have a pretty common flaw in their thinking. It’s very common to see thought processes biased about the hand we are holding and make plays which contradict each other. For example, an exploitative player may reach the river with a hand which is ahead of a capped range and choose a small sizing because they think villain can’t call a large bet with worse. Then they’ll arrive in the same situation with a bluff candidate abd won’t go huge because they think they’ll get called sometimes and a smaller size will work about as much. It’s very common to see this type of contradictory thinking in the strategy forums here.

Humans are notoriously bad at probability. I think that perhaps all but the very best players overestimate their ability to assess opponent’s frequencies and exploit them.

Good point. I agree and I also used to believe this misconception. Also the term "playing GTO" is heavily misused when it often means "attempting to play at equilibrium".

Let's pretend you're playing at the equilibrium. This doesn't prevent players from making huge mistakes and "giving" you EV; "worse" players will lose more vs a GTO opponent than "better" players (even though this GTO player is still not deviating from equilibrium). A player who deviates from equilibrium to exploit these worse players "making mistakes" will win more than a player who does not deviate from equilibrium, but this does not mean that the GTO player will be breaking even.

In other words, when playing versus a player at equilibrium, certain strategies will lose more or less than other strategies.

An easy way to see this is imagine if you're heads up vs an opponent who folds 100% of hands preflop, or a player who folds everything on all rivers except the nut low which he calls, or a player who constructs his preflop range by folding 88+, JTs+, JTo+ and then opens everything else. The GTO player never deviates from equilibrium but these bad strategies are going to lose more or less than other strategies; therefore the GTO strategy is not a guarantee to break even.

The biggest flaw in the way GTO is passed around on the forum is bet size.

Bet size *should* come first, then the corresponding frequencies, then the frequencies become balanced ranges. This process is ‘optimal’.

Game theory only cares for what is optimal. Just because you choose to study a certain bet sizing and create balanced unexploitable ranges for such, does not mean you are optimal.

GTO just says that any deviation from said strategy in a particular spot is less EV.
A GTO strategy in a particular spot is not the same as "always +EV" or even "always breakeven". The GTO play in a particular spot can be -EV.

If you play GTO in every spot then that is only breakeven if everyone else also plays perfectly GTO.

Hi, I think qui nguyen win to Gordon Vayo bc he played more GTO. But Gordon Vayo tought he plays a better GTO, so he avoided big pots. Thinking no need to risk big pot if he plays a better gto. But his deviations cost more in EV. If he really play gto big pot and small pot he win. But he only play small pot gto. And because qui nguyen knows he force big pots. So then vayo just did not play gto. But he s the better gto.

There used to be yadoula user who explained his toughts about gto and I think it made a lot of sense if you read between the lines.

By far the biggest, in my experience, is not understanding the differences between the theoretical properties of optimal strategies in 2-player games as opposed to games with more than 2 players. They are very different.

The biggest flaw is in trying to choose between exploitative play and GTO play. They can be used together in a relatively simple and usable way.

Take MDF on the flop as an example, with a half pot bet the MDF is 66.6%. Apply blindly and you will loose. Start with looking at your preflop range. Did you open, call 2 bet, raise to 3 bet bet, etc.. Your flop range must be a subset of that specific range. But, before you consider removing 33.3 % of your preflop range, consider your hand. You will never actually fold a set because of MDF.

You can easily establish a percentage of your flop hands and draws that you will continue with regardless. Say that is 22% for the sake of discussion. Remove the weakest 22% from your preflop range. Play them for value. Remove 22% from your MDF number, it's now 44% Use the 44% to again reduce your preflop range, removing the weakest hands first. Use this
range to make your continuation decision. The result will be a continuation (MDF) percentage that is closer to the 66.6% number. Still not exact, just much closer. It will take some experimentation to determine your best MDF range. The change in win rates are dramatic.

An interesting approach but this method is an attempt to approximate a GTO strategy rather than using GTO and exploitative together. There is nothing in this example about exploitative play. (Unless you are meaning how the choosing of the value hands is an exploitative method I suppose).

What I posted was a deliberate over simplification just to convey the basic idea of how to practically use MDF.

The key point is that GTO and exploitative play can be combined, it is not an either or. Not combining them is a big mistake.

In my lifetime I never expect to play against players that play a perfect GTO strategy. However good they are they will make mistakes that can be exploited.

By first dealing with conventional play, value or exploitative, and then incorporating things like MDF you can have the best of both worlds. GTO is not the best strategy against players that make mistakes! A combination is.

For example, if a player folds too much that can be easily exploited. If is ranges are either too wide or too narrow that can be exploited.

MDF simply takes advantage of a players mistake of continuing with an incorrect range. It does not preclude exploitation of mistakes. T

The idea that exploitative play and GTO are in any way mutually exclusive is simply wrong. Both can improve your win rate. Both can be developed away from the table and used at the table. No need to develop instinct It can be simplified into something as simple as an MDF range table.

In heads up zero sum poker, GTO will crush such a MDF range table. An absolute bloodbath. That is the point. Unless your exploit is based on strictly actionable biases in your opponent, GTO just wins, wins more, and has less variance. Why spend time studying anything else?

I am trying to separate the theoretical from the practical. You are correct, in theory. But I am not capable of playing GTO poker in real time. I don't know anyone who is.

If I can't play GTO against random opponents, then I need to be able to take advantage of GTO even if it is not perfect. The whole idea is to discover how to use GTO theory to improve my win rate vs. conventional wisdom or exploitative play.

What i am researching is practical alternatives that can be applied in real-time at the table. I can memorize a range that I develop away from the table using simulation. It will not be perfect, just practical.



I have simulated this basic approach and observed improvements in win rates that astonished me. Just believe that there must be a better way.

If you think you have the capacity to critique if someone is playing GTO you are delusional. You'd have to have solved the whole game. So to say "I don't know anyone" who plays perfectly is kind of incoherent.

Second, and I'll say it again because you seem to have missed it initially, you are not combining exploitative and GTO in some kind of brilliant fusion. You are attempting to approximate what you believe to be GTO in a simplified way that can be applied in real time. How is this an exploitative approach?

Sorry, but I don't understand your reply. I did not intend to imply any criticism of GTO. It's just that I don't think there is a person alive that plays GTO. Not talking about pure theory here at all. I believe that it is valid to assume that in a game with real people that all of them make mistakes. If they make mistakes they can be exploited. Against opponents that make mistakes Nash Equilibrium does not apply. GTO may not be the best strategy when players make mistakes. GTO is good but some deviation may be even better for your win rate against players making mistakes.

I thought that this was clear in Game Theory.

MDF exploits opponents who make the mistake of either continuing too often or failing to continue often enough.

This should not in any way be personal. Terms like delusional are personal. Please just explain where i am wrong.

You are trying to approximate GTO play in order to exploit. You are not overtly trying to move away from equilibrium to exploit people. Thus, this is not a mixture of exploitative and GTO play, this is attempted GTO play only.

You wrote "Against opponents that make mistakes Nash Equilibrium does not apply" which is completely contradictory to your whole argument, that MDF is an equilibrium such that when villains deviate from perfect play they lose.

My complaint with your original post is how you argued that GTO and exploiatative play can be mixed together but then went on to say that the MDF is an approximation of GTO play which does not address how they can be mixed together. Rather, you are simply using an equilibrium method that will cause players to exploit themselves.

To play in an exploitative way we ourselves must deviate from equilibrium because it is more +ev. This is not what you are explaining. You are explaining an approximation of what you believe to be GTO play. This is not a mixture of exploit and GTO, do you understand?