Flop subset on the website, for example what does this mean: 2s5dTc:43.36

43.36?

Thx!

As long as bb is 100 then yes.

You can stop it and restart it later as long as the tree is in memory it is however doing so is not recommended as the preflop solver doesn't remember the state of the computations exactly and you may get more fluctuations if you keep stopping/restarting it. It's recommended to do it in one go as of now.

It's under Tools->Configuration->Behavior->Max number of threads

0 means OS default (so max to efficiently use the CPU) you experiment with different values there.

Estimating time of software development is a proven impossible task so we are not attempting to give any ETAs. We also don't make any promises as a rule as that is counterproductive and doesn't help with anything. It's the best to wait and see what's available.

The architecture is this:

Graphical Interace (PioViewer) opens the solver (engine) in the background when you do stuff like building, solving and browsing trees. When you run a script PioViewer opens another instance of an engine and that instance runs the script leaving the main engine free to do other task in the graphical interface. That's why you can run the script and stil be able to do various stuff in the GUI.
In principle you don't need the GUI to run a script. You can open the solver itself (PioSOLVER-pro19.exe if you have a pro version) and type "load_script my_script_name.txt" there. The graphical interface just does it for you.

When you click "generate script" button the viewer looks at the current tree config and generates the script containing that information. Then when you run the script another engine (solver) is opened and the script is sent to it.

You do need to fill in the config but you don't need to build the tree itself.

Can I buy it using neteller account ?

Perfect, this is where I have been messing up, makes total sense on why I never have my full RAM.

Thanks a mil for your time and help.

Hi,
Is there any option to get email when solver stops working. Also, is it possible to use minbet sizing or do "math" on sizings like use 80% of 2e sizing?

Just a thought - have you guys considered implementing Bayesian Nash equilibria? So for example the user would specify two opponent ranges (type A and type B) with corresponding probabilities (of encountering those opponent types) and calculate the resulting Bayesian Nash equilibrium. Although it's 2-way, this would be calculating Nash equilibria for 3 player types, so it would become quite a bit more computationally costly, but it would be a great way to model proper adjustments to population tendencies.

The payment options are listed on the website, that being said please pm me and we may be able to find another option.

Not directly although if you are determined to make it work you might be able to.
There is an option to perform a custom action when the calculation is done under:

Tools->Configuration->Behavior->Notification

You may start a program there or play a sound etc. If you had a program which sends you an email when it starts (that is easy to code but you would need to ask someone) you could request it runs when the calculation is done.

We haven't considered it. It's an interesting idea. I don't think it's as computationally expensive as 3way but I honestly haven't thought about the problem deeply.

It looks like we need a way to model various opponent types first

When building a raked tree, why doesn't the pot size decrease on the next street after betting takes place?

For example, say pot is 100 chips and we have rake set to 10% uncapped. If player 1 bets 50 on flop and player 2 calls, the pot should be 190 on the turn after rake, but it still shows 200.

In my understanding, the ability to assign ranges to multiple opponent types with corresponding probabilities of how likely you are to play that opponent + node locking would be everything you need to model opponent types. So the only addition would be the multiple opponents part.

Yesterday i had 20 aggregation reports running. When i woke up only 8 were saved. Have you ever had any issues like this? i never had something similar

In poker the rake is taken at the end of the hand. I know some poker sites display it differently but it's not the way it works traditionally. It's also very convenient to have pot size independent of the rake as then you can talk about lines like "turn is a bet 150 into a pot of 200" and it shouldn't depend on what rake is.

You can verify that the rake is taken by going to an end node and looking at overall EVs of IP and OOP player there.

It's very difficult to guess what went wrong without seeing many more details sadly.

like?

Is there any way to make a script perform both a full save and a very small save? I want to perform river aggregation and have flop saves, but I don't want to keep the full saves after the river aggregation. But opening each full save and making a very small save is tedious.

In general when reporting that something doesn't or didn't work as expected you need to provide details of what exactly happened, what kind of errors occurred, preferably a log as well.

You are doing something non standard (Pio can only run one report at the time so I guess you opened it 20 times) so that makes things more difficult to diagnose as well although if we know which exact error occurred we could try making a reasonable guess.

Please refer to point 7 here: https://piosolver.myshopify.com/pages/faq

You would need to manually edit it before running it.
When you generate a script you see there is a part like this for every flop:

you want to replicate dump_tree line but without "no_rivers" argument, like this:

Please notice that the names for both saves must be different so one doesn't overwrite the other.
This has to be done for every flop. After that you can save the new script and run it using Tools->load script

I recommend experimenting with very simple trees and short solving time to make sure you get it going before investing a lot of time just to find out something didn't work as expected.

Is there any way to limit the amount of threads used by PioSolver?

#1) im utilizing preflop browser and i'd like to know if there's a way to split the preflop ranges in range explorer (like pairs, suited connectors, etc) to calculate the distributions of values/bluffs preflop (i know its a simplistic view especially preflop but you get the point). if not, is there an easy way to copy pio ranges to flopzilla?

#2) is it possible to configure first color of ev stop at ev 0 when not utilizing colors in gradient? (Ev<0 always red)

Yes, it's under Tools->Configuration->Behavior

You can do that in the range selector if you want. We don't have preflop categories like suited connectors/broadways etc. in the range explorer but you can click that yourself i you want and see what % of the whole range they constitute.

To be honest I don't know. It will have to wait until my friend who programs the GUI is back which will take a few days.

Is it possible to get check/bet/fold %s on aggregated reports or something like that? For example get the check/bet/fold% on preflop analysis for each board, or even for turn runnouts.

Yes, just run the report in the decision node you are interested in (for example first decision on the flop).

Hi there!
What processor is better for Pio? Ryzen 1800x with 64gb of ram? or an i7 (7700k? 6850k?) with 128gb of ram?
Thank you.

What do you think about the possibility of implementing a bet size suggester for 1 and 2 betsizes? Clearly it would have to heavily rely on heuristics, but I think it could end up being pretty useful.

Ryzen is going to be faster than 7700k by about 30% and about 15% faster than not overclocked 6850k. 128GB of RAM might be useful for really big preflop trees. If you are choosing one I would go with 128GB RAM option as that hardware will be future proof.

It's not a priority. It's pretty easy to insert bet sizes you want in our treebuilding form so any value from autosuggestion would need to come from estimating which sizes give the biggest EV. The differences are usually small enough that researching that would take a lot of time to have anything of value. In general fixating on choosing the best bet size is a mistake. It really doesn't matter much as long as you are choosing something reasonable.

Hello,

When i run a runout aggregated freq analysis what does OOP EQR and IP EQR mean and how are they calculated? Is that Equity Realisation? because in my report i got all numbers < 1 but one slightly over 1?

Thanks

It's EV/EQ.
There was a bug when calculating those in the aggregation reports at some point. Please make sure you are running 1.9.2.5 version of the viewer, you can get it here:

https://www.dropbox.com/s/6p13pzi7j3...iewer.exe?dl=0

(download and substitue in your Pio folder).

I switch the viewer but still the same think going on?

https://imgur.com/a/st5s4

It's sorted by OOP EQR