Note: Mods please move if this is the wrong forum, Ive never posted outside NVG so no idea where this is supposed to go.

Every serious player is aware of GTO these days, but players have differing opinions on how much GTO and how much exploitative you should try to play. I have a reg acquaintance who is heavily into the GTO camp, where he argues that the best thing to do is to try to memorize ranges and frequencies as well as you can in all spots, and that trying to exploit is a waste of time "as you open yourself to being counter-exploited". He is the kind of person who will still call off against a nit because "Its a GTO call".. and argue that its the correct thing to do. This is not a thread about that specific argument, as Im pretty sure most will agree that theres more EV to be gained by playing exploitatively off a GTO baseline understanding rather than "pure GTO".

His argument that Im wondering about was "As long as I follow GTO, I cant make -EV plays, thats how GTO works". Im pretty sure this is wrong, but I wasnt able to come up with a clear argument...

I made the following example:

BU v BB 20bb eff, BB plays a 3b jam or fold strategy and only jams AA/KK. Calling v jam here is clearly -EV if following GTO ranges. He countered by saying that "The EV I lose by calling his unbalanced jam is recovered by all the EV he loses by overfolding to my opens, so its not really -EV to call".. which I guess is correct even though folding to jam is clearly superior with almost full range. So my example wasnt really a good example of GTO being -EV in certain spots.

But surely there are spots where if you assign a certain range to V, following the GTO range will be -EV in that specific node? Like CO v BU 25bb eff, BU 3b to 5.5bb ... some people seriously only do this with AA/KK. If you call GTO range and then get into a flop spot where Vs range is exclusively AA/KK, can you still follow GTO (without nodelocking to Vs real range) without being -EV on the further streets? Once you get into this node, where Vs range is so unbalanced and strong, is GTO still able to be +EV in that specific node?

I am fully aware that you cant exploit GTO, Im not trying to say that this is an example of profitably exploiting GTO as GTO would still win overall vs this unbalanced strategy. I just got into this argument and realized I dont understand how GTO strategy plays out in spots like this.. not a solver student by any means either so apologies if my question doesnt make sense.

Not being nitpicky about multiway dynamics, he is right that his strategy will still beat the nit and won't be -EV as a whole, and that's how it works in basically every spot.

In your second example, the node facing 3b or postflop you're ****ed yes, but the node of you opening is way higher EV due to more equity realisation, so the line and range as a whole wins.

All in all, GTO "can't lose", but saying "the call vs AA/KK isn't really -EV because i make it up with other parts of my range" is silly and not how it works. The call is really really -EV.

Essentially, you could just play postflop ignoring the fact that the guy only has AA/KK and you would still win due to spots where you don't end up in that situation, but it's as stubborn and dumb as throwing 33% rock vs a guy who only throws or will throw paper.

And to top that off, I guarantee he can't memorise or play 10% of the way GTO plays, so past a certain point of diminishing returns it's not even worth trying.
That's not to say GTO is not extremely useful though.

Example: You have accurate ranges for both players, and you know and accurately built the tree the exact way you and villain would play the spot. So you plug it into the solver with the flop texture that was dealt, and the solver gave you an EV number for you. You're guaranteed that, no matter what villain does from this point til the end of the hand, your EV can't be lower than this value. But in a specific node, let's say river, villain plays suboptimally and yes, you can either be losing EV or gaining EV, and that loss or gain has to be compensated by other nodes.

How visible would be a given exploit, and how easy would it be to get counterexploited, are the main questions IMO. Why would your acquaintance want to calldown a nit, weaker player who rarely adjusts unless you're hitting him with a hammer, and sometimes not even after being hit by a hammer, is beyond me.

Thanks for quick answers. That is basically what I thought. Even though technically your strategy is not losing EV because V's nittyness is offset by how much EV he loses in other lines with the rest of his range, the isolated post flop spot is -EV and its just an unnecessary rigid way to look at the argument "GTO cant be exploited".

I have no idea and I dont even know why I tried to pursuade him because hes a misreg and I dont mind that he punts his stack against nits. He just kinda butted into a discussion I was having with a friend and we got into the argument and I got annoyed at how adamant he was being when his argument is so muleheaded.

In HU gto can't lose in mw it can


In your example with COvBU if CO 3bets wider and blinds cold 4bet wider you just end up losing because you are forced to fold all the time.

If GTO 'can't lose', how does GTOw AI crusheth the other gto based bots?

Better approximation of true gto

Where's the line dude??

Won't solutions with less sizings always run faster (to the same degree of accuracy), or be more accurate (for any time limit it has to run)? So if GTOw is more accurate with choosing the best possible sizing, instead of either less accurate, or running multiple sizings and using some computational heuristic to approximate things when sizings don't match their own solutions, I can see why it beats the others.

Never forgetting all of these are only guaranteed to be pseudo-GTO, right?

Your example of a BB playing fold or shove with a shoving range of KK+ is a good example of what GTO is and is not. A GTO strategy cannot be exploited by a player using this BB strategy. Playing a GTO strategy would bet you exactly the same EV vs this player as it would against a player who plays a GTO strategy in BU vs BB. What GTO does not do in spots like this is optimize your EV. That’s why I constantly fight against the terminology “GTO” (game theory optimal)* -it is not necessarily optimal. An optimal strategy against that player would be to pretty much open ATC and fold to the shove unless you have AA. A pure GTO strategy will be +EV, but the strategy I gave would be higher EV against that player.

The problem here is that we typically don’t have our opponents’ strategies. If we are mistaken our “optimal” strategy could do far worse than GTO. We can use a non-GTO strategy to exploit our opponent, but if we are wrong about his strategy, he can use a non-GTO strategy to exploit us.

* The technical term “Nash Equilibrium” is much better IMO. That’s what GTO really is, an equilibrium strategy that results if two players start with arbitrary strategies and adjust and counter-adjust to exploit each other. They will both eventually reach an equilibrium state where further adjustments cannot increase EV - the Nash Equilibrium, which is what is really meant by GTO.

Things are harder multiway, your friend is a complete simpleton. In HU you can't beat a Nash equilibrium strategy. In multiway, if the table is playing in some equilibrium one player alone can't gain EV by switching his strategy (by definition). Let's say one of the players plays strategy A in that equilibrium (which has EV 0 if all are playing in THAT equilibrium by assumption). There is no guarantee that strategy A has EV>=0 if all the others (or even just two others) switch strategies. To think that you can't lose money by playing some strategy which is part of an equilibrium 3+ handed means you have NO IDEA of the math behind things. Trying to memorize some specific game tree to replicate it exactly is the act of a FOOL (a fool with a tool is still a fool). (I am in no way stating that studying GTO strategies is bad.)

So to be clear, the sentence "As long as I follow GTO, I cant make -EV plays, thats how GTO works" IS COMPLETELY WRONG and can't be further from the truth assuming it's not HU and assuming he means with "As long as I follow GTO" that he plays some fixed startegy A which is part of some equilibrium (which it probably means). It is true in HU though. To be even more clear: Assume you have 4 bots all playing some strategy which is part of a Nash equilibrium and 2 other players playing 6max (non of the bots/players colluding). Assuming just this there is NO guarantee that your 4 bots are breaking even or winning. If there are 5 bots playing strategies of ONE Nash equilibrium and one other player they are guaranteed to at least break even.