Ok, I'll take it under consideration.

Hi Scylla

is there a feature already present that calculates the average EV for e.g. 10 different flops? For either CREV or GTO.

As you said, the frequency % are not very useful. But I think knowing the EV can be the start to preflop solving.

Yes, GTO+ offers this feature. With it you can create a database of multiple flops and then see the aggregate data for the entire database. This feature is described in the second video here: http://www.gtoplus.com/videos/

Please let me know if you require further support,

Scylla

Feature request to add the ability to use Big Blinds when building a tree instead of only % of pot. Currently, this feature is only available when editing a tree.

For the advanced tree builder it's only possible to enter bet sizes as a % of the pot. The reason for this is that the pot may be a different size between different scenarios. So a solid size might not be applicable to the situation, whereas a % will always be correct.

Hey Scylla, Quick question for GTO+. When on the 'run solver' window and we are deciding the accuracy of the solver what is the percentage a percent of and what does the percentage mean? For example if we input 0.5% into the solver and get it to run till completion then is our strategy accurate (whatever that means) to 0.5% of the pot or the effective stacks or 0.5% of one big blind? Thanks for your help

Also if I'm on the wrong thread please direct me to the GTO+ thread.

It's to 0.5% of the pot.


It means that if either player were to play max exploit as opposed to their current GTO strategy, their overall strategy would only improve by at most 0.5% of the pot. So, for example, if the pot is 30, then even if they exploited every single mistake by their opponent, then their overall EV would only increase by 15 cents.

Hello Scylla,

I am having trouble running turn reports for certain parts of the tree. For example, OOP 3B IP, OOP bets flop, checks turn and faces a stab. How do I make a report for OOP's defense versus the turn stab?

There's a few issues in displaying frequencies for spots that are not the first decision. Most importantly, both OOP and IP will have a different range for each line. So in other words OOP's checking frequency will be different for each turn. And IP's response to that check will also be different. Although it's possible to plot the frequencies beyond such a point, the data for each turn is measured under completely different circumstances. As a consequence, this data isn't really suitable to be plotted in a graph/table. We can consider doing so anyhow for later releases, but right at this moment, we have left it out for this reason.

Makes sense. Thank you.

Hi, I'm very noob, but I was curious then I downloaded the trial version of your software to take a look of how it works, then I did a test, I solved a spot where sb open raise and bb coldcall 100bbs deep(using 3different sizes for bb on postflop tree), my two questions are:

1: To solve this spot it takes almost 1 hour, I know a big part of this is cause my pc specs isn't good, but solving trees using 25bb deep I could solve it much faster and in a very aceptable time, I know that most wider the ranges and most deep stacked the players, more time will take to solve a spot, but the flop texture can interfere the time to solve aswell? T96 double suited is a board with a lot of draws, if the board was for example KK2 rainbow, the time that my pc would take to solve the same spot 100bbs deep would be smaller than on that wet flop?


Question 2: I build a tree with bb using three different sizes post flop 50%, 66% and 75% then I checked how sb should react facing a bet from the size 66% on the flop, but if I had solved the spot using only one size for the bb (66% only), when the sb face his 66% bet on the flop in the tree that I solved for only one size his strategy would be different than from the three that I solved using three different sizes?

Can you perhaps send a savefile to support? I will run it on my own computer and let you know how long the solve takes over here. One hour sounds way too long though. Unless the tree is huge or you're solving to a tiny dEV most solves will take at most a few minutes. So please send a savefile and I'll take a look at it.

If in the first tree bb did not bet the 66% line 100% of the time, then this will mean that in the second tree sb will be up against a different range from bb. As a result, he will need to choose a different strategy. Sb's overall EV is not expected to be significantly affected though.

I sent u a pm, I solved with a nash distance of 0.50%

This tree is somewhat large, and takes my own computer about 4 minutes. To be fair, this exact computer is rather fast; my previous computer would probably have needed about 10 minutes. So one hour does indeed sound a bit long. Is it possible that your computer is 5 years or older? GTO solving requires quickly performing operations on large chunks of memory. Older system are far less efficient at doing this. Another important factor is the number of cores in your system. The standard nowadays is 4 cores. If you only have two, then the solve will take twice as long. I do notice though that you are using multiple bet sizes for the turn and river. The tree will become half its size if you just use single lines there, and solve faster. The flop solution should not be significantly affected by using a slightly simpler tree for the followup play.

Yeah my pc is a dual core from 4 or 5 years ago, I did another tests here following your tip and using only one size for flop and turn, for single raised spots that wasn't bvb the time was aceptable 15-20 minutes, thanks for the support, one more question, how to cap the number of raises post flop? to something like never 3bet or if 3bet only 3bet shove

We have not added such an option, given that this would not reflect real life play. Should you want to remove certain lines anyhow, then the tree editor can be used to make changes to the tree as you see fit.

Hey Scylla,

Thanks for answering all the questions here. It's very helpful.

A few more for you:

Could you elaborate on or point me in the direction of how I can get GTO+ running as fast as possible on different machines? What type of machine would give me the fastest solving speeds with this program?

Also, do you know of any pre-solved solutions that are available for sale. I'm thinking something like PIOCloud where someone has run multiple huge scripts and is offering the database for a price.

Thanks for all the help.

I'm interested. I haven't been able to use PM's so please go ahead if you're interested.

glgl

I like editing trees directly instead of using the tree building function in the advanced settings. Is there a way to save it? It looks like the only way is to input all the options in the advanced setting and then going to "Edit Profiles."

The solving speed is almost exclusively decided by the processor speed multiplied by the number of cores. So for example, if you have 4 cores of 4.2gHz then you are essentially running on a 4x4.2=16.8gHz system. As for pre-solved solutions, I'm not aware of anyone offering those.

For this:
1) Load your tree
2) Change the ranges/board/settings as you see fit
3) Go to "Build tree"
4) Select the "Rebuild" tab
5) Click on "Rebuild with current settings"

But this doesn't actually save the tree does it? If I were to close the program and reopen it, I'd have to rebuild it from scratch.

Perhaps I misunderstood your question. What I thought you meant was that you had manually built a tree that you liked, and wanted to use in different scenarios (different ranges, board, settings). The method that I described will allow you to load this tree, change the ranges/board/settings and then rebuilt it with this new configuration. You will of course need to store it as a new savefile, but this method will indeed duplicate your exact tree. Were you looking for something else?

Well, I watched the video in which you (or someone else?) used the advanced tree building settings and added it as "MyProfile," making it the default setting. I would like to do this except make all the edits under the "edit tree" part instead of the advanced settings since it's easier for me to navigate. But, like you said, I guess I can just store it as a new savefile and just open up that savefile each time.

Thanks for your help