Doing pre flop calculations at the moment and making small saves (no rivers). Two questions about restarting them from saves (say, if I want to calculate it to a higher accuracy, or do node locking, then calculate a new equilibrium):

1) All I need to continue a simulation do is load the small savefile and click "Go"? Or do I need to run some solver command before I run?

2) A "metadata" file is always made with a save. Does this file need to be present in the folder where I load the save from in order for the tree to load correctly?

You need at least a basic version for that.

Most of the solving is on the rivers so once those are discarded the ability to efficiently resume solving is gone. You can theoretically do that, like this:

1)make sure you are on 1.9.2 version
2)load a small save
3)ctrl+b and type: "rebuild_forgotten_streets" there (without quotes) and hit enter
4)wait
5)you can now resume solving

Notice that step 3) rebuild the tree but the rivers are still not solved. It won't be faster than just building the tree from scratch and solving.
Step 3) is mainly useful if you don't have the config anymore but you still want lock nodes, round strategies etc.

The .metadata file contains information you can feel in Tree information tab (so general comments about the tree). You don't need it to load the tree.

I'm having trouble nodelocking turns. It seems like whenever I try to nodelock turns the models get screwed up with regard to the actual turn card. Like this model I'm looking at, its a Qc6s6d9d board and I was nodelocking IP to fold 77 and 88 sometimes...now its locked to fold everything with the 9s including 9s9x.

I've tried fixing it manually but it doesn't work. I've tried editing the turn fold node as fixed and proportional. I also went through a bunch or turns...it seems to work fine on heart turns but the error shows up on clubs, spades and diamonds (so on a rainbow flop the error is happening on cards that bring backdoor flush draws). I have the Aug 31 update (PioSOLVER-edge 1.9.2)

ok thanks

Please provide the tree config so I can try reproducing it. To get the tree config go to treebuilding and calculation tab and click "copy to clipboard", then paste it to pastebin (or similar service) and link here).
I won't be able to reproduce anything if I don't have the tree config you are using.

When I load a small save (with rivers excluded from save) and browse the tree, will river play for a specific board get recalculated automatically on the fly if I browse my way to a river spot?

Yes. It's usually very fast as well (should be less than 200ms for most trees on a decent CPU). You can even choose the accuracy the rivers are recalculated to by going to:
Tools->Configuration->Behavior->Recalc accuracy as a fraction of the pot

The default is 0.0005 which is 0.05% of the pot.
I talk a bit more about saves in an updater Quick Start Guide video, here:
https://youtu.be/QnzFpjSr3Lk?t=1852

I re-ran the trees and the problem seems to be solved but I had run into the issue a few times before, so if it comes up again, I'll e-mail you the tree config and everything. Thanks for the reply.

First I've got you wrong about this: "1.9 features and ability to solve one flop (Qs Jh 2h)"
Than you got me wrong.
Sorry.

Hi,I was loading trees comfortably by drag and dropping files into PIO but I can't load a file in specific.When I try to open that file by drag and dropping,it shows this:

https://gyazo.com/5fda7d9ebcc54fc273540faf3c23dd67

Then PIO disconnects to the solver,so apparently I have to connect again to it if I want to load another tree.I tried to load it by going on CTRL+O then selecting the file I want to open but the same exact problem happens.All other trees are loading fine,only this one that isn't.How can I fix that?

Thanks

You probably can't as it's likely to file is corrupted. You can try going to File->Load tree to verify it (if it crashes after loading it means it's corrupted).

There was a short time where the solver sometimes produced corrupted files if you clicked save too fast after stopping the calculations. It was fixed in 1.8.2 I think but if they were older save it could be the reason.
Make sure you update to the 1.9.2 version if you are still using an older one.

hi, sorry guys, I have a question about EV:

https://gyazo.com/81f71d93743c899b39ab609db7cf83d0
OOP EV = 2.9689 = weighted mean of hands set's EV
In this SB vs BB (6max) spot, in which SB can fold, limp or open to 3bb, SB has an EV of 2.97-5 = -2.03 meaning -20.3 bb/100

In the same spot, removing the openraise option and leaving costant the rest of the tree:
https://gyazo.com/a6aa34e0e3e2cb59907d93a072475fdb
OOP EV becomes 8.667 meaning +36.67 bb/100

Question is: why is there a big EV increase, reducing OOP's options? I'd have expected the EV to be decreasing or at least costant.
I am aware of the rake effect explained in the FAQ [Q: EV drops after adding an extra strategy option to the player. Is this a bug?], but there they talk about steep rake (while ours is a 5-10 simulation, rake is 3.3% cap 3.5), and the difference in EV is small; here we see a huge difference.
what could be the reason?

the two trees configuration files, inca se you need, are here:
https://www.dropbox.com/s/q309h17pgb...Blimp.txt?dl=0
https://www.dropbox.com/s/dvjtu0ricb...bbIso.txt?dl=0


thanks

Thanks for reporting. It will take a while to re-run as those are huge trees.
How many flops did you use? What accuracy did you reach?

** edit: sorry, I realize now that I probalby ****ed up my tree configuration (it looks like OOP and IP are inverted in the starting pot)

as usually, a double check was not enough, should have checked again . sorry for bothering you and thanks for your time

It's always a good idea to first use "Build pure preflop tree" and browse the resulting preflop tree to see if it is what you want it to be. It's very fast and the easiest way to verify the config.
Anyway, I've re-run your trees without rake for now (always first step when checking the results) and the value of that initial raise is surprisingly small (around 1bb/100 but I only used 45 flops and didn't run to very good accuracy).

I am trying to add .cfr files to PioSolver, and I get this message

https://gyazo.com/0375d687da6da483a81e21ce0932b9d9

I have the newest version, not sure what I'm doing wrong.

Did you follow the instructions in that message?
Please try doing exactly what is described and if it doesn't work send us the save (assuming it's a small save and you can upload it somewhere or it's small enough to fit as attachment) to support@piosolver.com

Hey everybody, i'm hesitating to buy the basic version of piosolver, to help me get my game to the next level

However, my laptop was quite cheap and not the strongest ever, my CPU is
Intel pentium CPU 3825U 1.9Ghz (dual core if i remember correctly)

Is it ok or is it gonna take me 10 minutes everytime i want to simulate a BvB flop ?

This is a typical laptop CPU and it's in fact not the fastest one (likely 4x slower than average desktop from 2-3 years ago).
You can test the performance by downloading a free version:
http://piosolver.myshopify.com/colle...piosolver-free

or directly from here:
https://cdn.shopify.com/s/files/1/07...90791015049404

and running some trees on Qs Jh 2h flop (the free version now supports solving on this particular flop; it also supports most fatures from the commercial version). The updated quick start video is here:
https://www.youtube.com/watch?v=QnzFpjSr3Lk

I'm doing pre flop calculations at the moment and have a question (for all here) about how (if possible) to extract from Pio's HU calculations decent pre flop calling/3B ranges when there are players behind us left to act.

For example, UTG opens some range (that we'll assume is fixed and reasonable, chosen by us), and we can calculate an optimal strategy for 3B/call/fold on the BU. In the calculations it is assumed that the blinds are 100% dead money that UTG and BU are dividing between them, thus BU will be playing more hands than would be optimal in a real game (where the blinds will sometimes get involved).

I'm looking for an intelligent way to approximate BU's optimal ranges in a real game. For example, we could reduce the amount of dead money in the pot and use that as a proxy for the reduced EV of our weakest hands when we flat or 3B them, and sometimes have to play a multiway pot. Now if we rerun the equilibrium calculation in Pio with less dead money, BU will play less hands. For CO it would be the same, just that now we have three players behind, and the effective amount of money we can get from the 1.5bb in the pot is even smaller. So if we went with such a model, we should should make the pre flop pot gradually smaller (in some intelligent and consistent way), as we move further away from the BU.

1) Is this a reasonable big-picture idea for manipulating Pio to produce more realistic responses to a raise in front when there are players behind?

2) How could we implement such an approach?

I've been thinking about running a simulation in HoldemResources Calculator to approximate how often the players behind us are getting involved. HRC would overvalue flatting for everyone behind us (because of the check down model it uses), so we would get a (for us) pessimistic estimate, which would lead to conservative ranges for us. Not necessarily a bad thing.

Then the question becomes, once we can approximate how often one or more players behind us get involved, how can we translate that into an effective reduced pot size for feeding into Pio's HU calculations.

Thoughts?

Hi,I want to nodelock a combo,let's say I want every Axs to bet at least 75% on A64r(I'm sure villain bets Axs at least 75% of the time),and I want PIO to figure out what to do with the other 25% of Axs given villain is forced to bet Ax at least 75% of the time.Is that possible?

What I'm getting atm is that PIO is betting Axs fixed at 75% of the time and checking the remaining 25% with every single Axs combo(I would expect PIO to bet some Axs combos more than 75% at least sometimes,but for every Axs combo is locked exactly to bet 75% of the time(and also to be checked 25% of the time consequently).

Thanks!

We get this question a lot. At the moment I don't have anything to recommend. In 3 way pots (BTN/SB/BB) it probably makes sense to just made dead money amount smaller and ignore SB (to solve for BTN range) but this is still eye-balling it more than being a rigorous process. Maybe it's a good question for a Skype group, you may get some ideas there if people are willing to share what they did.

This is not possible at the moment.

I am trying to open turn aggregation reports with open office, but all the numbers show up in just one cell.

https://gyazo.com/5a8863dc92523c9ee28eb9d37d5631cd

How do I fix? Can I make these reports or do I need to get excel?

thanks

This is because of your delimiter settings. The reports are produced using American notation which means "." is a decimal separator (and "," is a separator between fields but that should go without saying as those are .csv files).
What you need to do is to find those settings in Open Office and change them.

From what I've seen Libre Office as well as Google Docs open those correctly by default. Excel has the same problems as the ones in the screenshot (depending where you live and what are default settings for your country). Every program designed to open .cvs files will have delimiter settings somewhere in the options.

Thanks. I got it to work in open office by importing from open office(not having PIO auto open), and clicking settings to "comma"