One thing which takes a while to calculate in preflop trees are overall frequencies for lines (the ones displayed in the right upper corner). You can disable them by going to:

Tools->Configuration->Data presentation - other ->Don't show average line frequency info when browsing the tree

Other than that if you are using small saves then recalculating turns just has to take a while. You can increase re-solving frequency in:

Tools->Configuration->Behavior from default 0.1% (on the turn) to something like 0.25%. You may also consider saving the tree with flops and turns (it will be bigger but recalculating rivers is very fast).

Thanks, it's much faster now.

hi again punter ,

I am facing a problem here , you probably can help me out

the script interface through pioviewer seems to be limited to a certain number of characters
you probably wonder how I reached the limit but I use ranges from preflop calculations and I just copy pasted it to do postflop calculations
these ranges are not rounded so each number is like 10 digits or so.

I tried to open directly the solver command and input the ranges there through the command set_range IP/OOP but I get an "invalid argument" message although the command is fine.

so Im wondering is there a way to bypass that or is there an easy way to round ranges through the pioviewer interface somewhere ?

thank you

It might be the case, are you using 1.9.2.5 version of it?

You can create a script in notepad (or any other text editor), save it as for example script1.txt and then go to Tools->Load Script in the top menu to run it.
You can even generate script with simpler ranges, then copy-paste the whole thing to notepad, change the ranges, save and then run.

There isn't. I am putting it on the to-do though as that's a certainly a nice feature to have.

i did not upgrade to the latest version on my personal laptop , that might be the reason

anyway thanks a lot for your help , I used notepad as your adviced and it works perfectly

When I click on an action to view the range that takes that action, is it possible to view the frequencies as a gradient (like when you input ranges), rather than a rectangle that gets smaller for less frequent combos

So first those are not frequencies, those are ranges. If you want to see frequencies of actions then you need to look at strategy (click the strategy button).

Not right now although there are two ways of making rectangles smaller ("glass view" and smaller rectangles in the center. You may find the other one more visible, you can find it the option to switch here:

https://gyazo.com/e2ef967218b0b62610b6b5af5cd51d1d

(3rd one from the bottom)

Hello,

Is this fixed?

But in this training video it appears you can. https://gyazo.com/1c42acf3fad116f52550e5d92069b133

Notice how the T8s hands which appears in the range at a low frequency, is a more pale yellow. Whereas the K7s which is appears at almost 100% of the initial 4 combos is a solid yellow

Hello, I recently started using a remote server for larger PIO solver and I'm getting an error message about the PIO activations being used even after I 'Disconect'.

Can you please help me so I can use the programm inside the remote desktop.
Thank you
Christian

Yes, please pm and I will send you the fix after the weekend.

Right, you can do that by unchecking "square size proportional to weight" checkbox on the right (there are 4 checkboxes there). I am sorry, I forgot about this option

You need to deactivate it if you want to move it another computer like here:
https://www.youtube.com/watch?v=5Ozjfnatwms (question number 3)

Hello,
just recently purchased PIO and I'm having trouble understanding how the total EV of IP+OOP can be less than starting pot (minus exploit-ability and rake).

In simulating a standard BB-defense (3x BTN open, BB call, 65 chips starting pot, 970 effectivestacks). 5% rake, capped at 6bb (= 60 chips?). I've solved for a long-time to achieve 2.32 bb/100 exploit-ability. This is what I get:

SOLVER:
running time: 669.775
EV OOP: 23.899
EV IP: 32.836
OOP's MES: 24.132
IP's MES: 33.068
Exploitable for: 0.232
SOLVER: stopped (required accuracy reached)

How can EV of IP+OOP not be 65*0.95 - 0.232 = 61.518?

It should be at least total pot - max rake.

Do you really pay 6bb rake? Such a game would be unplayable.
It's likely 0.6bb rake so 6chips. If it's 5$/10$ game then chips = dollars.

Standard rake in online poker is 3chips at 5/10.

I am not sure what you are doing here and why you are subtracting exploitability from EV. It makes sense to add EVs and those will add up if you correct the rake.

Exploitability is the difference (per hand) of what maximal exploitive strategy can win against current equilibrium approximation (so how much an opponent who knows exactly what we are doing and adjust perfectly would win).

Yes, this makes sense. I'm however not quite clear what the mathematical relationship would then be. Would this be correct: Starting Pot = OOP's MES + IP's MES + average rake?

The example was played online at 50NL. So ~5% rake ($0.01 per $0.2 in the pot), capped at $3 or 6bb.

This still begs the question is the "missing" EV here really rake?

Starting Pot - (OOP's MES + IP's MES) = 7.557 = average rake?

If so, that's just and exemplary demonstration of how horrendously high the rake is. It's taking ~12% of the EV in this spot! So BB defense here has to be even tighter pre so that EV OOP = 25+. So some of worst hands in this range (like J9o, A7o, 64s) are actually -EV here because of rake.
I think I'm going to go throw up now.

It isn't. It's OOP's EV + IP's EV - average rake = pot.
You shouldn't add MES'es. The sum can be arbitrarly big if both strategies are terrible and it's getting closer to the some of EVs when the strategies are getting closer to the equilibrium.

Yes, then it's correct. This game is unplayable by any reasonable standards.

Yes.

Again, you should add EVs not MESes. Adding MESes doesn't mean anything.

Yes, rake at low stakes makes the game completely unplayable unless you have big donators in the game.

It would be -EV against competent players but you know, the reason you play in that game in the first place is that there are very bad players at the table. Otherwise only the site wins.

Well, playing small stakes without hefty rakeback is likely not the best idea unless the games are really good. You are not going to win with this rake against anyone competent.

Hi, Is there an option for piosolver to not use 100% cpu (maybe like 90%)?

Yes. Go to Tools->Configuration->Behavior and change the number of threads there (0 is max available given by OS). It's the best to experiment with that but probably something like 3 threads on quad i5 or 5 on quad i7 is what you are looking for.

Hello, I am wondering about LHE. What does PioSOLVER do for LHE players? Can you use a 5-bet cap? Can you add a rake?

add: What about Linux? Do you have to use Wine or something?

Ran the free version on Linux Wine. Does not work. It appears to safe-guard against programs like Wine. Looks like I have to spin a VM.

I believe you (or your team?) created the kuba flops.
Have these been tested for 4 card, 5 card, and 6 card (for the respective PLOs) and for 4 and 5 card hi-lo 8b?

Nevermind, got the VM to work.

You can solve LHE trees although they are quite big in a full form.
A full postflop LHE tree with HU ranges and 4 bet cap on every street takes around 12GB of RAM.

Yes although such a tree will be even bigger so you need a really serious hardware to solve it.

Of course.

We don't support Wine but some people managed to run it in Wine HQ I have heard.

We certainly haven't implemented such a safe-guard intentionally. We just don't support Linux or Windows emulators. It's a Windows software. If it works on something which can run Windows then fine but we don't think about it when developing the software.

No. There will be separate subsets for PLO at some point.

VM should work as long as it runs modern 64bit Windows.

Hallo, are there any plans for a GTO-Trainer where you can load your solutions and train them?

Let me copy-paste my answer to one email about it today:

This is often mentioned feature and we even had that in our old decision tree software. Sadly it's not so easy to make it interesting as in GTO solutions it's mainly about the balance not about clear mistakes (which are easier to quiz for).
The problem here is that even if you always choose 10% actions out of 90-10 mixes you are technically still "correct" (not losing against GTO) and it's very hard to quiz if your strategy would be exploitable.
We have some ideas there but decided to work on other things first so this likely has to wait unless we come up with implementation we are happy with and we think is helpful.

I am currently having issues when I try to analyse runout over multiple files. I have the files saved as full (so including river). When I run the reports for flop or turn situations its all good. But when I try to run the river I run into an out of memory exception.

Is there any way to fix this? Do I not have enough RAM and would getting more fix this? Or is this another issue entirely?

Regards,

Akroma