(GTO) Where Exactly Does Deviation Lose Money Against An Optimal Opponent?
Posted by AF3
Posted by
AF3
posted in
Mid Stakes
(GTO) Where Exactly Does Deviation Lose Money Against An Optimal Opponent?
My understanding of balanced play in poker is that the best a counter-strategy can is obtain 0EV against the balanced strategy, but that the balanced strategy also actually exploits every counter-strategy which deviates from balanced (it just might not exploit it to the maximum degree). This seems true from experience and I've heard a lot of other people "confirm" it, but I don't exactly see the mechanics of how this is actually happens.
In Rock, Paper, Scissors, the balanced player actually makes his opponent indifferent to playing every other possible counter-strategy, so what is different about poker to where counter-strategies against the balanced strategy can obtain -EV?
I'll give a hand example that might help my question:
Suppose we are playing a pot in position and we are checked to on a final board of:
QT222r
If our balanced value-betting range when we make a PSB here is 2 Value: 1 Bluff then every hand which beats our best bluff will have a call of 0EV. Some of the hands which beat our best bluff, however, are hands like T3s and so on, which seem like they should be naturally exploited by a balanced river betting range. In other words, mathematically we have JJ = T3s = 44 (can somebody confirm this correct)?
If we run the hand from our opponent's perspective though, it seems like he is deviating massively by getting to this river with some of his bluff-catchers (44-55 etc.). So where does he lose money for making his turn-calling range too wide? Is it that he got a lucky river card and he'll have to fold on most? That doesn't seem right either because we'll still be bluffing rivers and 44 will beat our bluffs.
I understand check-calling this river with 44 is massively exploitable, but I'm asking where it's specifically exploited by a balanced overall strategy.
Loading 7 Comments...
"Many people imagine GTO in poker to be like a GTO strategy in Rock Paper Scissors. On the one hand, the strategy is unbeatable; on the other hand, it doesn't win against any other strategy. "
cited from: http://www.pokerstrategy.com/news/content/Would-Chuck-Norris-play-exploitatively-_79418/
this should explain your issue http://www.highstakesdb.com/community/topic/29580-sauce-on-gto/
In Rock, Paper, Scissors, the balanced player actually makes his opponent indifferent to playing every other possible counter-strategy, so what is different about poker to where counter-strategies against the balanced strategy can obtain -EV?
RPS is a 1 street game with static equities for every strategic option (there are only 3)
Tic-Tac-Toe is a multi-street game where a dominated strategy can exist, since if 1 player trips and falls on an early street, he can never catch up with the optimal strategy, which never trips and falls.
in TTT, tripping = not utilizing squares that offer the best value for connecting 3 in a row.
In chess, tripping is making error's like... ignoring basic developmental concepts. If you don't develop your pieces efficiently, you find yourself with a spacial disadvantage on future streets, and your strategy is dominated b/c you can't catch up.
Like sauce said, poker is more like TTT and chess than RPS.
just to elaborate on the chess analogy, finding yourself at a spacial disadvantage is a lot like realizing on the river that you played a shitty range on earlier streets.
In chess, from the moment of your first mistake, GTO villain is swarming in on your positional deficit, gaining EV with every move until the threshold EV of checkmate is reached.
In poker, GTO villain is playing a strategy on each street that can only be neutralized by the corresponding GTO strategy for hero. If at any point hero falls behind the GTO strategy, "positional EV" is lost and cannot be regained.
if you throw chess positions into analysis programs, you'll actually see it denoting the current positional advantage by +(for advantage white) or -(for advantage black), and then a number. so +0.5 means white has a small advantage. If at any point that number rises over 1.0 in favor of villain, the game is pretty much lost for hero unless villain blunders.
I was actually thinking about the chess analogy yesterday and today. It's almost like chess moves are poker combos--For example, you could look at a position and say "Well, if I put the knight over there then he puts his bishop here and that's really bad for me". You could also look at a hand history and say "Well, if I bet my second pair here then he just calls with his trips" etc.
In other words, it's almost like each chess player's position is analogous to his poker range and each chess player's subsequent move is analogous to how the poker range is constructed (calling range, raising range, folding range etc.)
I actually don't see how it's not like 100% certain that NLHE is more complicated than chess.
In chess, for example, there are 18 possible opening moves. In poker, even if you use fixed pre-flop raise sizes, there are 1326 possible opening moves!
However, many of the pre-flop strategies should be able to eliminated by domination arguments, but it's quite hard to do that rigorously.
To answer the question, where exactly does deviation lose money against an optimal opponent? Everywhere, sometimes in the form of winning less EV in a hand, sometimes in the form of losing when supposed to breakeven, sometimes in the form of losing more when you should lose less. But in the general strategy vs. strategy, taking the whole picture into account reversing positions and playing the strategies, the one who deviates against the optimal opponent will lose to him, maybe only slightly so they still both lose to the house, in this case the optimal opponent will be the bigger customer to the house :p who knows, it will depend a lot on the deviations and severity is my best guess.
"In poker, GTO villain is playing a strategy on each street that can only be neutralized by the corresponding GTO strategy for hero. If at any point hero falls behind the GTO strategy, "positional EV" is lost and cannot be regained."
I don't think it is about neutralizing a strategy on each street, since the end result of taking the highest EV line doesn't neutralize one strategy, it just makes him not be able to take a higher EV line. So his strategy won't be neutralized, it just means that in that specific spot he can't do no better than x, but the player in position will still have the higher expectation for the isolated situation, the thing is tho same thing happens when positions are reversed and only then strategies become neutralized because they both play the same strategy in both positions, hence breakeven after switching roles, but the player isn't neutralizing the IP players strategy by playing optimal OOP, he also needs to play optimal like the IP player when positions are reversed.
Right, i agree and i think yours is a better way of explaining it.
Be the first to add a comment