Open Side Menu Go to the Top
Register
GTO+/CardRunnersEV? GTO+/CardRunnersEV?

11-21-2021 , 02:04 AM
Quote:
Originally Posted by scylla GTO+/CardRunnersEV?
When playing GTO, any bet size will perform almost exactly the same as any other bet size.
Because of this, it's not possible to determine which bet size shows the most EV; all bet sizes perform the same.
You can test this for yourself by building the same tree, but for different bet sizes.
After that, solve the trees and compare their overall EV (the EV below the table in OOP's very first decision).
As you'll notice, the overall performance remains the almost the same, regardless of the bet size.

If that were the case, why wouldn't you just make flop push/fold for every single situation?

Don't the variations in EV for bet size matter over a large sample? In the SS you gave me, that's a quarter of a chip difference in EV, which at small stack sizes in an SNG might matter quite a lot.
11-21-2021 , 03:14 AM
Quote:
Originally Posted by Kanadian GTO+/CardRunnersEV?
If that were the case, why wouldn't you just make flop push/fold for every single situation?
Don't the variations in EV for bet size matter over a large sample? In the SS you gave me, that's a quarter of a chip difference in EV, which at small stack sizes in an SNG might matter quite a lot.
At a certain stack-to-pot ratio there will be a significant EV drop.
So, you can't push in for $100 if the pot is only $6.
However, any reasonable bet size will perform alsmost exactly the same as any other reasonable bet size.
Even overbetting the pot by 2x will often still show a very decent performance.

Quote:
Originally Posted by Kanadian GTO+/CardRunnersEV?
Don't the variations in EV for bet size matter over a large sample? In the SS you gave me, that's a quarter of a chip difference in EV, which at small stack sizes in an SNG might matter quite a lot.
The difference isn't a quarter of a chip.
It's typically around a hundredth of a chip.

Once again, please just test for yourself by experimenting with some different bet sizes in the same tree, by comparing their overall EV.

The influence of bet sizing is almost non-existent.
However, the influence of using an incorrect strategy (whether to bet/check/fold a certain hand) is huge.
So having absolutely perfect bet sizing makes almost no difference to your overall EV.
Whereas even seemingly small mistakes in strategy can lead to huge EV losses.
So, the importance of playing well outweighs importance of choosing perfect bet sizes by a large margin.
11-21-2021 , 01:53 PM
Hi Scylla, I have this suggestion.

On the main analysis page (with card matrix at center), would you please add a 100% bars chart on the right bottom corner which shows the distribution of bet x/bet y/.../check/fold based on the hand rank?

Or at least, give us the option to extract the distribution based on hand ranks so we can plot it ourselves? The numbers are there already, but there is no way to access them.
11-22-2021 , 02:59 AM
Quote:
Originally Posted by K_of_diamond GTO+/CardRunnersEV?
Hi Scylla, I have this suggestion.

On the main analysis page (with card matrix at center), would you please add a 100% bars chart on the right bottom corner which shows the distribution of bet x/bet y/.../check/fold based on the hand rank?

Or at least, give us the option to extract the distribution based on hand ranks so we can plot it ourselves? The numbers are there already, but there is no way to access them.
Are you referring to the percentages/combos next to the stats?
Switch over to "Percentage" mode (F2) and mouse over a stat to see it at 100%.

11-22-2021 , 09:26 AM
Yes, they are. However, we can't extract these numbers (for all ranks) to analyze. Can you provide the data extraction button, just as with the hole cards?
11-22-2021 , 01:14 PM
Hi,

Have you seen that HRC now accounts for bunching in some of their calculations? They recently released an article here: https://www.holdemresources.net/blog...ching-effects/.

How viable is it to implement something like the for post-flop solves in GTO+? For example, would you be able to allow the user to specify the frequencies of cards remaining in the deck? This could act similar to how you can select "dead cards" in Flopzilla but of course more nuanced as the cards would still be in play and the probabilities of cards hitting would be conditional in a non-trivial way on the hole cards/board cards (at least that's some difficulties I imagine, you possibly have a much better idea of what difficulties there would be in practice).

Very interested in hearing your thoughts on this.
11-22-2021 , 04:01 PM
Quote:
Originally Posted by JackOscar GTO+/CardRunnersEV?
Hi,Have you seen that now accounts for bunching in some of their calculations? They recently released an article here: .
How viable is it to implement something like the for post-flop solves in GTO+? For example, would you be able to allow the user to specify the frequencies of cards remaining in the deck? This could act similar to how you can select "dead cards" in Flopzilla but of course more nuanced as the cards would still be in play and the probabilities of cards hitting would be conditional in a non-trivial way on the hole cards/board cards (at least that's some difficulties I imagine, you possibly have a much better idea of what difficulties there would be in practice).Very interested in hearing your thoughts on this.
Ok, we can consider it for future releases.
11-22-2021 , 04:01 PM
Quote:
Originally Posted by K_of_diamond GTO+/CardRunnersEV?
Yes, they are. However, we can't extract these numbers (for all ranks) to analyze. Can you provide the data extraction button, just as with the hole cards?
Ok, I'll see if we can add a text output button.
11-22-2021 , 06:05 PM
I've been having an ongoing problem with GTO+, but not all the time. I can't seem to intentionally replicate the issue. Maybe like 5-10% of the time I run a sim after it finishes "finalizing" the solver crashes and reopens. This can also happen if I start a sim and a soon after press "Stop Solver". Sometimes the exact same sim will crash again, then the 3rd time I run it it's fine. It doesn't seem to be an issue with the size of the tree since I've had very large trees work fine, while smaller trees can still crash. This also never happens on the turn, it's only when I initially solve the flop.

Usually I don't mind so much, but it's very frustrating if I spent some time building a strategy and nodelocking only for it to crash and have to do it all over again.

It did it again as I was typing this

I'm running v 1.3.6, but this has persisted through a few versions. 64bit on a Ryzen 5 3600 with 16GB ram.

Is there a way to enable logging, or is it done automatically and saved somewhere?

Oh also, just as a request for something that could maybe be added in future versions. Could there be a hand class of 1-card backdoor flush draws? It's a pain to sort through all of those offsuit hand combinations locking and unlocking to select hands that have a spade blocker or whatever for flush runouts.



Edit: I've actually managed to make the solver crash several times now with a specific locked range. When I unlock it, it's fine. When it's locked, no matter how long I let the solver run, as soon as it finishes finalizing it crashes.

Last edited by CameronN; 11-22-2021 at 06:30 PM.
11-22-2021 , 06:32 PM
Quote:
Originally Posted by CameronN GTO+/CardRunnersEV?
I've been having an ongoing problem with GTO+, but not all the time. I can't seem to intentionally replicate the issue. Maybe like 5-10% of the time I run a sim after it finishes "finalizing" the solver crashes and reopens. This can also happen if I start a sim and a soon after press "Stop Solver". Sometimes the exact same sim will crash again, then the 3rd time I run it it's fine. It doesn't seem to be an issue with the size of the tree since I've had very large trees work fine, while smaller trees can still crash.
Usually I don't mind so much, but it's very frustrating if I spent some time building a strategy and nodelocking only for it to crash and have to do it all over again.
It did it again as I was typing this
I'm running v 1.3.6, but this has persisted through a few versions. 64bit on a Ryzen 5 3600 with 16GB ram.
Is there a way to enable logging, or is it done automatically and saved somewhere?
Can you try upgrading to v1.3.9?
We made a few changes that may resolve your issue.

Quote:
Originally Posted by CameronN GTO+/CardRunnersEV?
Oh also, just as a request for something that could maybe be added in future versions. Could there be a hand class of 1-card backdoor flush draws? It's a pain to sort through all of those offsuit hand combinations locking and unlocking to select hands that have a spade blocker or whatever for flush runouts.
Edit: I've actually managed to make the solver crash several times now with a specific locked range. When I unlock it, it's fine. When it's locked, no matter how long I let the solver run, as soon as it finishes finalizing it crashes.
Ok, I will consider it for future releases.

Quote:
Originally Posted by CameronN GTO+/CardRunnersEV?
Edit: I've actually managed to make the solver crash several times now with a specific locked range. When I unlock it, it's fine. When it's locked, no matter how long I let the solver run, as soon as it finishes finalizing it crashes.
Can you send a savefile to support?
I'll see if we can reproduce the issue.
11-22-2021 , 07:01 PM
Quote:
Originally Posted by scylla GTO+/CardRunnersEV?
Can you try upgrading to v1.3.9?
We made a few changes that may resolve your issue.



Ok, I will consider it for future releases.



Can you send a savefile to support?
I'll see if we can reproduce the issue.

Yeah I'll upgrade, but it will be hard to tell if it resolves it or not due to the unpredictability of this issue. I sent an email with my savefile to support, hopefully I sent the file you were looking for.

Thanks!
11-22-2021 , 09:57 PM
Is there a way to force the same rounding used in the trainer pre-solve, so that the rounded outputs are available to study?
11-23-2021 , 03:35 AM
Quote:
Originally Posted by sublime GTO+/CardRunnersEV?
Is there a way to force the same rounding used in the trainer pre-solve, so that the rounded outputs are available to study?
To study a solution, I would strongly recommend looking at the unrounded solution.
It's entirely possible for us to round the solutions in the navigator, but this solution would simply be incorrect.
The rounding in the "Play against the solution" feature is there because if a solution is 99% check vs 1% bet, then many people would prefer this to be rounded to 0%/100%.
This same approach does not apply to performing analysis.

Last edited by scylla; 11-23-2021 at 03:40 AM.
11-23-2021 , 10:21 PM
Scylla I'm curious what is the best practice you suggest is for a database situation.

If I am looking at a hand in a database and like the custom advanced tree but want to change the flop, what is the best way to do that? The tree is not saved as a profile for the tree editor.

Thanks for your time answering this question.
11-24-2021 , 12:18 AM
What am I doing wrong with this very simple tree? I set OOP to "Don't donk", yet it still ends up donking.
11-24-2021 , 03:38 AM
Quote:
Originally Posted by 1BigOT GTO+/CardRunnersEV?
Scylla I'm curious what is the best practice you suggest is for a database situation.
If I am looking at a hand in a database and like the custom advanced tree but want to change the flop, what is the best way to do that? The tree is not saved as a profile for the tree editor.
Thanks for your time answering this question.
For this:
1) Change the flop
2) Go to "Build tree"
3) Select the tab "Rebuild"
4) Click on "Rebuild with current settings"

11-24-2021 , 03:42 AM
Quote:
Originally Posted by RiverJohn GTO+/CardRunnersEV?
What am I doing wrong with this very simple tree? I set OOP to "Don't donk", yet it still ends up donking.
This option only applies to the turn/river.
This is because on the flop, we don't know whether the bet is a donk/probe/cbet.
To leave out the flop bet, enter 0 for "First bet".

11-24-2021 , 08:54 AM
Hi Scylla ! Thanks a lot for updates, esp for EV accuracy when tree solved in bb.
Could you clarify what this feature is about ? [IMG][/IMG]
As i understood - now we can edit range just w/ a changing of %% digits in text. As well i can make some editing in FZ and paste text into GTO+, what would be amazing for the purpose of node lock studying.
But i doesn't work so. Changing %% numbers doesn't make any effect
Yesterday , 04:03 AM
Quote:
Originally Posted by relite GTO+/CardRunnersEV?
Hi Scylla ! Thanks a lot for updates, esp for EV accuracy when tree solved in bb.
Could you clarify what this feature is about ? [IMG][/IMG]
As i understood - now we can edit range just w/ a changing of %% digits in text. As well i can make some editing in FZ and paste text into GTO+, what would be amazing for the purpose of node lock studying.
But i doesn't work so. Changing %% numbers doesn't make any effect
It means that the strategy can also be edited by clicking on "Output" and editing the percentages.
See the screenshot below.

Today , 10:25 AM
Quote:
Originally Posted by scylla GTO+/CardRunnersEV?
Can you try turning OFF "Settings->Scale small stats"?
Scale small stats - disabled.
River strategy QTo 3.9 combos, but visually looks like 8-9. Is it possible to visually match the percentage of combinations?

Today , 06:26 PM
Hello Scylla. Is it possible to play against solution, only some par of range. ect hero is only get second pair.

      
m