I have a question to which the answer I've been wondering about for a few weeks. It seems like a weakness of GTO play is to be found in range prediction, since so much of the playstyle depends on correctly establishing and weighting opponent ranges and it seems like it may be profitable to confuse GTO players by making range corrections by shifting them to "another level," splitting them differently, and reversing them in certain ways.


Generally, this would make it difficult on the typical GTO player to make correct plays post-flop due to his misrepresentation of my range. I am speaking here of some fairly dramatic changes. I guess I'm wondering, in essence, if many of you have rebuilt your ranges specifically to combat this weakness of GTO-style players and, if so, what type of results have you had?


It seems like building anti-GTO specific ranges to deal with GTO-style players, at times, would be intelligent since many root their analysis by building opponent ranges out of more common preflop ranges. I mean we know how GTO players are playing, why not build ranges that make it harder for them to play game theory optimal poker? Am I just dreaming here? What is wrong with my idea?

Will it be more expensive than it is profitable?

A few problems I can think of:

1. A limited number of hands are actually profitable to play preflop based on position, player tendencies, and cost to enter the pot so there is a limited pool to build from.

2. Knowing a range of hands isn't the same as knowing an exact hand as there is still room for information hiding.

3. The gain in EV from hiding information has to be equal to or greater than the EV lost for making an abnormal play.

4. Postflop is a better place for exploitation than preflop.

You should be exploiting players if you find issues with their play, but I think to focus intensely on preflop is somewhat a waste of time.

Sent from my SM-G900R4 using Tapatalk

The money making decisions are post flop. OP wants to throw a wrench in the works.

Very risky proposition.

First step is to actually know the ranges the gto players are likely playing. There is already several different flavors of gto ranges.

If OP does not tailor his exploit to counter the gto ranges, he will just be missing some combos and adding others. GTO will still expect you to have a mix of value and bluff, and bet frequency accordingly. So, OP needs to have extra value when he should not, (like extra littles maybe), but will have to eject some +EV hands preflop to make up for the added range, otherwise its just LAG.

Yeah, that's what concerns me. It may be that GTO can still counter these ranges well enough anyway and that the adjustments might not produce a serious deficiency in my opponent.

While I know I can do simple things like switching hands in and out different ranges or playing an overall strange range, I wonder if more serious and pointed changes could produce prominent results.

I feel like range assumptions are such a weak point of certain players' games and that they can create systemic problems resulting in substantially larger reverse implied odds.

It must be. A GTO strategy's best response (i.e. what you're calling "Anti-GTO") is itself, or itself among others, in a symmetric game. Any deviation you make from it will be losing value unless it happens to be part of another GTO strategy, if there even are multiple GTO strategies for NLHEHU, which would probably be extremely similar due to the game's nature.

Any value you gain by arriving at a point in the game tree with a combination you shouldn't have and thus your opponent will not consider is more than compensated by the amount of value you lose at the previous parts of the game tree with that combination.

The bolded is probably true, but says nothing about a GTO strategy. A GTO strategy has no weak points.

It sound like you're using "GTO" to mean something other than what it is--a strategy that's part of a Nash Equilibrium profile:

There aren't any typical GTO players; there aren't proven to be any at all for commonly played poker variants, and nothing would be "typical" of them because they'd (probably) play exactly the same. GTO isn't a "style".

K, GTO is most definitely a style, but I'm not really up for a debate with someone that has grossly underestimated what I understand about GTO. Many players idealize game theory optimizations and then fail to realize that the implementation of game theory into poker falls short of its idealization, then you rudely come and argue with me about stuff that I already understand.

This thread is about a potential path of exploitation to thwart the non-ideal implementations of game theory optimizations and not a thread about combatting GTO styles with a GTO style. You should have realized that anyone that would ask the questions that I have asked in this thread in the manner in which they have been asked would already be aware of the most obvious way of combatting GTO players, which is to play GTO oneself. That is not what this thread is about.

Imo the obv weakness of a lot of weak GTO strategies is that they have often strict tree limitations. You can try to attack this weakness by manipulating the tree in your favor (by using uncommon sizings etc). You can also investigate if you could improve your EV slightly by adjusting your preflop range but it's more complex bc most of the time you have to deal with more than one player.

GTO doesn't care about your range.

Awesome question! I'm liking you more and more Hrmmm!!

I am in a bit of a rush today, but I had to make a real quick post.

Yes this is very possible and very useful in todays games. It's all about finding the most difficult thing for a GTO player to calculate. It's all about making life hard for them.

The vast majority of these guys are trying to memorise a whole strategy rather than working out the play as they go along. Its easy enough to for them memorise a preflop range because there are less variables. Preflop is the beginning of the 'strategy' and so its the easiest section to memorise. Its the flops where we can really start messing with them (Although we can and probably should pre-adjust pre-flop) the flop adds a shed load more variables and so its only the more advanced GTO player that can balance here. The turn I think is the best place to make your move, because by the river many of them are able to calculate their GTO play rather than memorising it.

I try to make weird lines on these streets to really make their heads explode, lines which they wont be familiar with. I make weird betsizes too. Anything they are not familiar with will make it difficult for them. These guys memorise plays and ranges and that is something we can use to our advantage.

But the human trying to implement gto does care about your range. Its actually a big part of the implementation.

Does OP think he could beat a GTO bot by deviating from GTO? (Clue: You can't "confuse" a GTO bot, because it knows - or can calculate - the solution to every spot. Note that when the best HU players in the world went up against Libratus, they tried everything they could think of to find its "weaknesses". It destroyed them.)

The whole point of a GTO strategy is that it will at least break even against every possible strategy you try. Most strats you try will simply lose.

Well said Arty. It should be intuitive that you can't triple stamp a double stamp.

Op, sure you can try this stuff and you'll probably get away with some spew now and then, but I think that in the long run you will do best by keeping it simple and just trying your best to play solid ev maximizing poker in this hand vs this opponent on this street on that card.

Why would you think I'd be talking about GTO bot? That just goes to show how biased that players are against any type of exploitative card playing. Clearly, I'm not talking about a GTO bot.

I understand what GTO is, Arty. The theory cannot be perfectly implemented for more reasons than almost all people understand.

The point here is to avoid taking a side-step towards saying "just play GTO against GTO." I know I can play GTO against GTO. I, also, know that there are some methods of exploiting GTO players that have yet to be fleshed out because players keep conceding to the fantasy that GTO implementations are perfect. They are not perfect. They aren't near perfect and humans are even less perfect. My point is too look for supplementary methods of exploiting GTO players in addition to playing GTO against them, if that eases anyone's mind. They exist, they are real, blah, blah, blah...

I get it, "play optimal poker." Well, that is what I'm trying to achieve. Just saying it doesn't do much, one has to figure out how, first.

Then in that case you should stop calling anyone by this name:

You're right that we all have leaks. I have tons. I'm exploitable. So are you.

Look, the term means one thing to you and another thing to me. For me, it is an imperfect term, and you apparently project perfection into it.

http://forumserver.twoplustwo.com/15...ology-1180391/
^ The sticky at the top of this subforum.

"Unfortunately, many words commonly used in game theory have very specific, exact meanings which may or may not line up with their usage in everyday speech. In order that we are able to communicate effectively, it is important to understand the technical meanings of a few terms. "

"nash equilibrium or (game theory) optimal strategies or unexploitable strategies

A Nash equilibrium is a set of strategies (one for each player in the game) with a couple properties. These properties are equivalent, they're just different ways of looking at the same thing:
- No player in the game can unilaterally change his strategy to improve his expectation.
- Each player's strategy is maximally exploiting those of his opponent(s), at the same time."

It doesn't mean what you're saying it means, and you shouldn't get upset if you're using the term counter to its specific technical meaning that posters on this subforum abide by as per the sticky, and all game theory writing in general.

This is getting so perturbing. Strategies are perfect... those strategies cannot be implemented perfectly and become imperfect when introduced into many formats. I'm not going to argue about this, anymore.

You don't get to make up definitions.

I'm not. Bob asked me to stop using the word "GTO player" since it implied that GTO players were perfect and I should not use words like that when applying to it players. I said that I will not stop using the term since I clearly realize that the term "GTO player" implies imperfection since game theory is impossible to implement into hold'em especially by humans. I didn't make up any definition.

This discussion is about possible ways to exploit game theory type players. If you are not going to stay along the lines of the discussion, then I would prefer that you did not interrupt the discussion by trying to remind me of the naive assumption that such players are truly unexploitable.

Like I said, I'm not going to argue about this anymore. We all know that game theory players are exploitable.

I think we all agree with that the theoretical GTO player/computer is unexploitable and unbeatable.
I think you're using the term "GTO players" to mean players that strive to play optimal poker, but aren't quite perfect and may have some leaks/imbalances.
The only way you can beat those players is either by playing perfect GTO (because it beats all strategies), or by identifying and targeting their individual imbalances and exploiting them.
e.g. Maybe Endboss#1 calls 'slightly more than GTO' when facing overbets on the river, but Endboss#2 calls 'less than GTO' in the same spots. The only single strategy you could use that is guaranteed to at least break even against both players is to play GTO yourself, but if you were trying to make more money, you'd exploitatively alter your betting frequencies against each player. (i.e. You play differently against each opponent).

The problem for you is that the very best players are capable of playing close enough to optimally that you just won't be able to find big exploitable edges. (It's partly why nosebleed games are dying). I mean, the question you're practically asking in this thread is "How do I beat players that are better than me at poker?"
The answer is you can't. You need to play better than they do, or play against much worse players whom play a style that is much further from optimal and is therefore more exploitable.
There is no such thing as an "anti-GTO" strategy, unless you mean a "losing strategy".

Much respect for Arty from me.

However, I think in all fairness to the OP, the question is more specific than that.

It is about preflop ranges, in particular, and if an exploitive range can be contrived that will throw of human gto calculations.

If we all think the answer is "no", then so be it.

Lets not get sidetracked by the terminology that we here in the theory subforum, and turn the thread into anything more than that.

K, this is getting really old. I think I may just have to give up and admit that 2+2 is not a place where a discussion like this is possible. I've had to repeat myself now several times and I'm just not into doing it once more.

Here's how I understand it, please correct me if I am wrong.

A true, perfect GTO strategy would by definition be unexploitable. The only exploit (if you can call it that) would be to employ the exact same GTO strategy, and by doing such, you would achieve your best possible result of breaking even.

A true perfect GTO system would have a dynamic situational preflop range constructed with the most profitable hands/the least negative hands. Trying to counter this by doing whacky stuff preflop ( I mean, what can you REALLY do besides wacky ****, like replacing 56s with T4 suited? etc...) would only result in losses.

If GTO were perfectly implementable, but it is not. There are many things that are being left by the wayside here because people will not get over the assumption that a GTO player is unexploitable. What we really need to be doing is looking at how GTO-styled players are exploitable. We all know that we can play GTO against it, but if we can temper our GTO-style with a slightly exploitative style we can actually raise our EV rather than being foolish drones that think it is impossible. That is my perspective anyway.

GTO as it is played by human beings and in the way that it is practiced is not a mathematical certainty.

So, whatever, I was hoping that people would spitball ideas, but this has become an exercise in futility that I am close to conceding. Admittedly, the initial post represents a more dramatic change, but this was really meant to be a speculative thread.