Does "Transparent GTO" exist?
Let's say that someone tells you all the information about the cards that will be in play, in the next hand (NLHE). You will be the only player who will know that. You will know all the hole cards of your opponents ( + your own) and you will know all the cards that will appear on the board.
So, before the hand starts, you know all the 17 cards in a 6-handed NLHE game.
My math instinct tells me that there is a big paradox that occurs here - and there is no GTO or optimal path for this kind of situation. GTO can only suggest the best move when your opponent's hole cards are hidden.
Here's an example (6-handed table NLHE):
You know that you (seat #1) will river stone cold nuts and win the pot (if you don't fold). You also know that there is a 3-way bad beat possible - 2 opponents (seats #2 and #3) will also have strong hands in the showdown but quite weak hole cards. The other 3 of your opponents (seats #4, #5, #6) will have great hole cards though - and 2 of them (seats #5 and #6) will hit a great flop, from their perspective.
Now the question is what's better - trying to keep both "bad beat guys" in the hand or maybe the other guys who have very strong hole cards knowing that two of them will hit a great flop. Is it better to play aggressive pre-flop ( + flop) or maybe keep that for later streets ?
In my opinion, there's no optimal solution for this even when you consider all the information (cards, position, pot size, bets, etc.).
You've also got to play in a way that doesn't reveal your magical clairvoyance powers. A good solution to this might be to get blind drunk, or otherwise act like some coked up investment banker type. It's the only way to make 4betting 72o look 'okay'
I'm analyzing this with cards in odds calculator right now. I'm quite sure that there are plenty of unsolvable situations possible, especially when opponents' stack sizes balance possible scenarios making it impossible to choose the best way to act.
Choosing between milking a big stack bad beat player who hits a great looking river and a player with medium stack who holds KK and hits a good flop and turn, seems kind of impossible to me.
I mean, when you play it would be easy because you see all the cards, but the software would have to constantly choose 1 best action for you. My bet is that no software can handle 100% of possible situations.
Maybe only with something like "you should call 40% of the time and raise 60% of the time". But these kinds of solutions would have to appear too often to consider this as a perfect optimal strategy, in my opinion.
With perfect information, the only way you could even attempt to determine the most profitable plays would be to make a judgment about how the others are likely to play.
Theoretically, you could assume the other players would all act GTO based on their imperfect knowledge, and then determine which of your plays would be the most profitable.
Of course, that would take an incredibly large number of calculations and still be less profitable than playing exploitively based on any actual knowledge you had of the players.
I was thinking like this also, but I wasn't 100% sure if solvers calculate normal GTO this way, too. But here, the situation is asymmetrical - only you know all the cards, so it's kind of different.
You'd have to assume that they also use "Transparent GTO" to compare T-GTO to normal GTO and confirm its existence. And you can't do that, so you can only assume they use classic GTO. So again, it's quite a different thing.
They don't have any more info than they would in a normal poker game, so I don't see how the determination of what would be GTO for them would be anything different than usual.
They don't have any more info than they would in a normal poker game, so I don't see how the determination of what would be GTO for them would be anything different than usual.
For them no difference, but if I were using a "Transparent GTO solver" software it would be a difference from my perspective and also programmer's perspective, whether this software fits the true definition of GTO.
Chess is a game of perfect information, and it is generally thought that there is a GTO chess play in any board position. A GTO poker play would still exist given perfect information; it would just be calculated differently.
As much as I am totally loath to admit it today, I guess Deuces has a point.
With transparent GTO you de facto increase the emphasis on more humanistic variables like players responses, given that now your optimal EV is so much more dependent on which direction villain moves over much more marginal and subjective decision points that cascade into infinitely more considerations. i.e. nightmare calcs like: there is an 80/20% swing on whether they call the x/r at this point depending almost entirely on 'vibes', or where they sit on the variance train (would need to formulate entire sessions/mood/weather/socks etc), and literally ALL (or literally nearly all) of the EV of my decision point depends on it.
Whereas with opaque GTO those responses are mere abstract variables spread throughout the gametree. Simple numbers. You solve for potentialities.
But there is no 'vibe equity'. The meta is easier than the micro. In the same way that toy games get increasingly easier to solve with less and less things.
The goal of GTO is not to extract the most possible out of a situation. It is to be unexploitable. Of course there is a GTO solution - probably multiple solutions - when having perfect information.
I think we could fairly trivially say that GTO does not exist because there is no multi-way GTO in general. Even regular hold'em doesn't have true GTO multiway.
Imagine the same situation of hero clairvoyance but HU. Then true GTO still doesn't quite exist because if we make V play HU GTO and we play optimally against that, then V can improve his winrate by unilaterally deviating. But, if we assume V plays optimally taking into account our clairvoyance, then there would be a nash equilibrium from which neither player could improve by changing their strategy unilaterally.
Imagine you're on a walk with a dog. A guy comes to you and asks to show him the way to the nearest restaurant in town.
He expects to hear one optimal path to the target.
If you tell him:
"Please go straight 700m, then turn right, then go the 3rd side street to the left and walk 150m. The restaurant will be to your right.
Or you can go straight 500m, turn right, then go into the 3rd side street to the left and walk 350m. The restaurant will be to your right.
You can also go straight 850m, turn right and walk 120m, the restaurant will be to your left, right at the corner."
It's more natural (and useful) to just tell the guy 1 optimal way to the target. And that's what a good strategy called a "solution" should include, in my opinion.
A problem with the term GTO when used in poker is that it doesn't really mean optimal. It means unexploitable.
Unfortunately GTU sounds like a personal insult. You can see why they changed it.
Loads of perfect information zero sum games have mixed strategy gto solutions ie they have a Nash equilibrium
Sent from my CPH2653 using Tapatalk
If you are playing perfect GTO and your opponent is also playing perfect GTO, then yes.
So it looks more like a defensive shield than a true offensive strategy.
It's not offense v defense. It's exploitable v. unexploitable. If you know someone is folding too much compared to GTO, then you can bet some more marginal hands. The problem is that you never know when they are going to stop folding too much and start calling too much. That's when playing poker begins.
Yes it exists. You know what cards are coming, you know how your opponent's will play, so it's just a computationally trivial EV calculation to find the best line.
You're basically solving a much much smaller game tree.
