Hi, I just bought the Pio Edge.

Building some preflop trees, I see I am unable to remove the call option vs 4b, I am interested in seeing what a '5b jam or fold' strategy looks like. Is there any way of doing this?

Hi punter,

Having a problem on complicated solves while running a script (many boards) after it completes 1 board instead of moving on to the next it says "error, not enough memory..." even though the stated memory needed is less than the available memory it says is available on my server.

I am able to close PIO, reopen it, restart "Jobs", "run all" where it will do 1-2 more boards before the same thing happens over and over again.

Any solutions for this issue?

Thank you!

Probably not.
The reason is that there are a lot of judgement calls involved in doing that. You need to construct a tree based on HH and to do that you need to be pretty confident what good enough tree to give reliable EV is (sizes, options etc.)
Another thing is that I don't really want to deal with importing HHs from poker sites as something may break at any time they change the format.
I understand it's a tempting feature but we have so many on our plate so that's unlikely to happen unless we partner with someone who already does the import part.

Yeah, you can dectivate. It's not a big deal as it's always possible to deactivate remotely later if it doesn't work. See here:
https://piofiles.com/faq/#Licensing__moving_activations

To remove the call click on that call in the tree builder so it's yellow and then click "Remove selected action". See the screenshot:
https://gyazo.com/9a9d41b68e73fe40164112c725f0fbc6

We are aware of that problem and it's a priority to fix it. Hopefully there is something on this front in a few days. There should at least be workaround like "ignore memory check for subsequent trees".

Pio 2.0 looks amazing! Is it now possible to switch suits on a board without having to resolve? Say 7s6s2h to 7d6d2c?

Hey, I was playing around with this sim today (BB 3bet vs SB) and I noticed something strange...

After reducing BB flop cbet size to a single size, EV went up significantly. This doesn't seem right, given that both sims have this size as an option. I tried running the sims to higher accuracy, but it's still giving the same result.

Has anyone encountered something like this before? Why would this happening? The single flop size sim is the one outputting 998.6 IP EV, multiple flop size sim gives 929.7. I'm using Pio 2.0, flop size is the only thing different about the sims.

https://imgur.com/a/8cORQGw

I'm trying to do pre flop sims on a 16 gig ram computer. The only thing I want at the moment is pre flop ranges and don't care about post flop play. If I remove the option to bet on the turn/river and don't allow 3B etc. to increase the subsets that my computer can run will this corrupt the preflop calling/3B ranges?

Edit: early returns on the sim suggest yes, very much.

It's a raked sim (EVs don't add up to the pot), and because raked game is non zero-sum there can exist equilibrium solutions that have significantly different EVs.

Oh interesting.. Can you elaborate on why that is? Would you recommend running sims without rake for this reason then?

Also - just ran the sim without rake, still getting different EVs... Multiple size sim is returning 963.4, single size sim is returning 1036.7.

Is card bunching used in the preflop solutions? Is it an option or is it always on/off?

We think it will be very unreliable.

It's not taken into account.

First, as to your example, we can't verify anything from screenshot. We will need the tree config which you can get by clicking "copy to clipboard" button and pasting it here or to email.
It's likely you have made a mistake somewhere but of course we will verify it.

As to zero-sum thing. Non zero-sum games have the following properties:

-there is usually more than one solution
-those solutions can have different EVs

This means, that unlike in zero-sum games, the one true solution doesn't exist. There is absolutely nothing the solver can do about it as it's immovable mathematical property of non-zero sum games. Sure, we can find one equilibrium but:

-it might a different one every time you run the solver
-it might be a different one depending on where you start (so having more betting options for example means it starts from a different point and might converge to a different solution).

That being said the differences in EV in your solutions are quite big which suggests there is a mistake somewhere (the trees are not the same). To verify that we will need the tree config as on the screenshot not everything is visible in enough detail.

I get this message when I try to open PIO even after I have closed all sites, I have to restart my computer for it to work again?
https://gyazo.com/4c598678d5dc655f78b90cdeaf3dac61

Hand2Note was causing that by creating a process with PokerStars.exe in the name.
I fixed it in the newest version but haven't backported to 1.10 yet.

My antivirus software is preventing me from downloading the link after purchasing the license. Anyone know how to get around this?

I downloaded the update that was put out recently, but is there somewhere I can see what's new in the update?

Isn't it a bigger picture thing? In the multiple flop sizings sim IP has different ranges for each of the bet sizes than in the single bet size sim. It's that simple I think.

Hi Punter,

In the GTO trainer (which is pretty slick btw) how can we choose a freq. when we have a hand that takes mixed action?
Is there anything similar to what GTO trainer offers where they have a rounding feature for mixed actions?

Thanks!

Uninstall it.
Complaining to their support is a good idea as well.
I mean 3rd party antiviruses are basically malware, there is 0 reason to run them. We took care to submit the updater to Microsoft so they confirm it's ok but we don't think the burden of fixing flawed heuristic for 10+ 3rd party antivirus providers should be on us. Still, we would try it if there was an easy way to do that but my experience contacting them is not very encouraging (unlike Microsoft)

Check changelog.txt file in your Pio folder.

This is about always caused by a mistake of some kind. Every time it happens and I press for exact configs it turns out there was a mistake.
I mean, if you are sure it doesn't work correctly it's a good idea to report but please be very specific which means:

1)this exact config A and this exact configs B are the same despite bet sizes at point X. After running them B (which has fewer options) have bigger EV in root (comparing anywhere else is incorrect)

2)Attach the configs (which you can get with copy to clipboard button

There will be an option to play with RNG in the very near future which I think covers what you want (and if it doesn't let us know once it available!)

[QUOTE


There will be an option to play with RNG in the very near future which I think covers what you want (and if it doesn't let us know once it available!)[/QUOTE]


Great, looking forward to it!

Great, looking forward to it![/QUOTE]

Done in dev.
If we don't find any issues during testing it should be available sometimes next week.

Is it normal that with the pio_cfr algorithm the saved files are about 70% larger in size than the original algorithm?

Yes.
Making small/micro saves smaller is on a short list of things I would like to work on in the near future though so the situation might improve.

Mine still has this problem, just with a different error message:

If I close Hand2Note, pio runs no problem.

Does anyone know what a "step" means to pio? I maybe want to use "go xx steps" instead of setting a target accuracy, but I can't figure out what a step really is and what sort of numbers would be appropriate here. I thought of deducing this by playing around with PioViewer, but the UI doesn't appear to support the specification of steps - that only seems to be available through scripting.

Any insights would be greatly appreciated.

It looks like the same problem (the solver dies in the background because it detects what it thinks is a poker site).
If you run the updater and then PioSOLVER-pro.exe (assuming the pro version) what date is the compile from? It should April 1st 2021.

For flop trees the info string in the treebuilding tab (on the right side when you are solving) appears every 25 steps. You can solve a sample tree, see how many steps it takes and then use that.
The value is 50 for the turn trees and on the river it's eithe 100 steps or 500 miliseconds (usually the latter).