Also, another quick question. If I am making a tree for 30/60 limit holdem, should I make the turn and river cap "4" or should I put "240"?

Limit trees are just very big. It's possible to build them below 16GB now but they are still slow to solve. It's a good idea to look for some simplifications (removing very rare lines) to get the solutions faster.

It's the easiest to use the limit tab:
https://gyazo.com/1caebecb46cb5d304c65648895153225

cap is number of bets so put 4 there (or less if you want to simplify the tree as limit trees are quite big).

It's possible to do limit in the main tab as well by using "c" postfix for bets, like here:
http://pastebin.com/Gz7bd9MS

(to see the configuration in your viewer copy the text and then to go Tools->Paste treebuilding configuration)

ok, thanks. I bet that's part of the reason why my tree was so big. I was putting the actual pot total for "cap" (eg 240 instead of 4).

Sorry if this question got answered already, but does the full version come with more aprox preflop ranges or am I able to get them anywhere else? Like SBvsCO openraise etc

Hi, I'm new to piosolver. Just a simple question I hope. I have Basic. When I put in the figures for postflop tree building & calculations, I entered Effective Stacks as 480. Then once I've built the tree & hit Go, Calculate & then gone to Browser, it says Starting Stacks are 120.
I'm hoping there's something simple I'm doing wrong?

i have this problem, simulated the hand and the results were confusing so i compared same hand with a friend and he has different results







what i dont get is what OOP doesnt have a calling range is just raising everything

and this are the results my friend gets (hes using 1.8.2.3 version and i have 1.9.2.4)

I followed the video, however I was trying to apply it to the lhe tab and not for no limit...i am unsure how to lock nodes there, thanks.

We don't ship any optimal ranges at the moment (just some ballpark ones for demonstrations).

Sounds like you are doing something wrong but it's very hard to say what without seeing the config (which you can get by clicking "copy to clipboard" button) or at least a screenshot from the config.

Pio produces the same results for the same tree. The most common reason we get reports like yours ("friend run the same tree and got different result than me") is that the tree is actually not the same. What you can do is to use "copy to clipboard" button in treebuilding and calculation tab and send the config to your friend. They can then copy it and go to Tools->paste treebuilding config to get exactly the same tree.

From looking at the screenshots you omitted river betting which is not a good idea as the results won't be realistic (we would play differently if we knew the opponent is unable to bet on the river). In general:

-avoid many betsizes (at least when starting, add them gradually as you are getting more familiar with the software)
-fill in at least the most important betting options (maybe not donkbets but surely at least one size for bets for both IP and OOP)

Again, this is most likely not the same tree. The only way to ensure the tree is the same is to copy the config. If you want us to re-run it and take a look we need the config as well.

In that case you need to:
-first construct the limit tree in the main tab, as shown here:

Then remove the lines you don't want by using remove line interface (or just removing some sizings altoghether for example turn donkbet if you don't want that).

This is a rare usecase, maybe pm me what kind of tree you want to build in some detail and I will send you a working config so you can start from there.

I am getting some strange results coming up when using the round strategies option.

So I run a Q64ss flop. I 3bet pre, cb flop and opponent raises and it gives me the following OOP EV/Strategy:

As Kh 105.86 and CALL 0.996
Kh Qh 527.79 and CALL 0.740

Now when I round strategies to 1/5 it changes them as follows:

As Kh 76.005 and CALL 1.000
Kh Qh 436.93 and CALL 0.800

This changes rather strangely when I round to 1/1:

As Kh -330.07 and CALL 1.000
Kh Qh -795.87 and CALL 1.000

How can the ev reduce so drastically in both cases when all I have done is shifted the strategy from betting 96% of the time to 100% for AK or from 74% of the time to 100% in the KQ case.

It's very hard to tell without seeing the tree config and the node path (the place you are looking at in the tree). It's not clear what you are doing either, the EV if actions don't change if you are only rounding one player's actions. If you are rounding the other one's as well then the EV may shift drastically.

Anyway, the best idea is to share (or email) the tree config and information which line it is.

What does EV represent in pio? i.e BB, % of pot? Thanks

Create a preflop sim where OOP folds all hands and solve it.
Or some other game that you know the answer.

Chips per hand. In range explorer we normalize it to % of the pot to make it easily comparable with equity. In all other places it's chips per hand.

I just started using PioSOLVER and want to thank you for making a great tool!

I have two questions, I apologize if these have been answered before:

1. In the text file (.cfr.txt) for a tree, I noticed that there are 'add_line's missing from the set of all possible lines. Is this done intentionally when the tree is generated to reduce complexity by not including some of the rarer lines? If so, do we just need to edit that crf.txt file and add more add_lines as needed, then hit 'go' again in the gui? Or is there another procedure to do this?

2. After doing many runs, I noticed that in all of them, the direction of convergence is always the same: the EV of the OOP player gradually rises (and the EV of the IP player falls) until the accuracy gets very good (< ~.01% of the pot). Then I start to see it oscillate a bit. I found this interesting. Shouldn't the convergence be more random or is this an expected artifact of the algorithm?

We don't remove any lines undercover
The reason there is less add_lines than lines is that the way it works is that checks/calls are filled automatically (you can remove them using remove_lines in both GUI and scripts but they are there by default). You can verify that by building a tree and browsing it (preferably before solving so you don't waste time). You will see that all the lines are there.

This is a very good observation. The reason for that is that we do iterations in interleaving manner: first for OOP then for IP. One full iteration comprise two for individual players and the information is always printed after 25 full iterations (well, by default on the flop at least). This means that the information is always printed after IP has the last chance to adjust. This is of course worth more at the beginning when the solution is not very good yet. That's why IP gets less and less from the effect during solving and OOP "catches up". I hope that makes sense, tbh I haven't thought about it myself but it makes perfect sense the way the solver is implemented.

if pro can do 16 cores is it 8 cores + 8 virtual cores, or 16 cores + 0 virtual cores?

or 16+16

Hi guys! Maybe someone has video from this post?
Unfortunatly, this video was removed from youtube.
Or someone write message to 6seven8, please. I have new account and cant use private messaging.

It means 16 threads. 16 cores with hyperthreading won't be fully utilized (although you will only lose hyperthreading bonus). The current pro version supports 12 threads, it will be 16 in the next release.

I am going to ask the author to re-upload it.

for filename pattern, is it possible to save other stuff? right now it shows {board} but what about other information like ip/oop ev, bet frequencies for a certain spot, etc?

Hello,

Sorry if answered, but can't find a solution to my very simple problem. How do I run scripts if I don't want to give a text file to the solver, but instead the raw line of commands?

Thanks a lot!

Sure, {board} just means that the board will be inserted there. Try something like IPvsOOP6max{board}

the way to experiment with those things is to run very small tree on 2-3 trees to bad accuracy to get the saves fast and see if it worked. Once you have that you can go back to running serious things.

I am not sure what you mean here, the script is a text file by definition.
You can execute solver commands using either the console (by clicking on PioSOLVER executable and typing stuff there) or using a programming language to open the solver process and communicate with it (opening console applications and using standard input/output to communicate is a basic operation and every programming language out there has a way to do that).

but there is no way to get {Board} IPvsOOP6max {OOPEV} {IPEV}?

There isn't but if your intention is to quickly see EVs for both players on many boards it looks like you need a multifile aggregation report:

https://www.youtube.com/watch?v=5Ozjfnatwms (question number 5)

hey
is there any way how to open .cfr as a chart/picture ?
thanks

Well, that contains the whole solution (or at least things necessary to re-create it) so that wouldn't fit in a chart/picture. What do you need exactly? There is a tool to create preflop charts from a preflop solution if that what you want (Tools->generate preflop charts from current solution). You can then browse them in preflop charts tool (ctrl+p) or even generate pictures with the ranges.