[GTO] Is it possible to solve a flop without calculating a river? - Poker Theory

Two Plus Two Forums Poker Strategy Poker Theory & GTO

[GTO] Is it possible to solve a flop without calculating a river?

Post Reply Subscribe

...

06-24-2021 , 10:51 PM

tombos21

veteran

Join Date: Sep 2018 Posts: 2,048

Is it possible to solve a flop without calculating a river?

Let's say you know the ranges, rake, bet sizes, etc. You had an accurate enough subgame. Can you estimate the GTO strategy based on the game state, without working backwards from the river?

Last edited by tombos21; 06-24-2021 at 10:57 PM.

Quote

06-25-2021 , 05:37 AM

aner0

Verified Coach NLHE

Join Date: Oct 2013 Posts: 3,673

It's possible to abstract the game and do this with some sort of machine learning.
Of course it will be far less precise than solving the whole thing

Quote

06-25-2021 , 10:08 AM

Brokenstars

Verified Coach NLHE

Join Date: Jul 2012 Posts: 16,104

Short answer is probably no, long answer probably something along the lines of what aner0 said.

Quote

07-02-2021 , 01:54 PM

tombos21

veteran

Join Date: Sep 2018 Posts: 2,048

Let's change up the question. Is there a human-implementable method to estimate GTO strategies on the flop?

Let's say you know the exact ranges, equity, SPR, etc. But you don't know the EV. Could you create a pseudo-GTO strategy, without calculating all 2352 runouts?

Quote

07-02-2021 , 05:42 PM

Brokenstars

Verified Coach NLHE

Join Date: Jul 2012 Posts: 16,104

I'm going to go with no. Turn/river is a huge aspect of the game and is more important than the flop. Without conceptualizing or thinking about future streets you're not going to be playing the same game at all.

Quote

07-10-2021 , 01:51 AM

tombos21

veteran

Join Date: Sep 2018 Posts: 2,048

Quote:

Without conceptualizing or thinking about future streets you're not going to be playing the same game at all.

This is an interesting point.

Let's imagine we abstract their raw equity into a [0, 1] game type solver. It assumes equity is fixed, but it *can* calculate future streets. How badly would this abstraction lose?

What if, instead of raw equity, you use an equity distribution based on all future runouts? That way you could represent the polarity, (e.g. a draw that's gonna be nuts/air by the river, or a pair that's gonna be stable equity). This is one way to "conceptualize future streets".

Is this kind of abstraction useful?

Quote

07-10-2021 , 04:50 AM

Brokenstars

Verified Coach NLHE

Join Date: Jul 2012 Posts: 16,104

It'd be helpful, but honestly equity realization is more important.

Quote

Post Reply Subscribe

...