I am happy to announce PioSOLVER - very fast turn and river solver for Holdem. It's completely free, no trials, expiration dates and other surprises.

What does it do?

-it solves games from turn or river
-one bet size at any point
-it gets to almost exact Nash equilibrium in seconds
-it exposes the whole solution tree with exact ranges, EV's and strategies

How does it look?



More on our website and blog post announcing it.


What does it mean?

This is very powerful tool and great way to get into game theory on a bit more practical side than books or abstract coaching videos. Not so long ago people thought that turn solution are far away from being computationally feasible on your run-of-the-mill laptop. They were wrong. PioSOLVER solves typical turn spot in seconds.

Why free?

In few days we are going to announce commercial version (solving from flop) on twoplustwo and our website. We only need to sort out payment to twoplustwo and prepare some minor details. Free version is our way to give people great, powerful tool and let them evaluate our claims

Where can I get it?

Here

Anything else?

We spent a lot of hours developing this tool and PioSOLVER run for hundred of hours on our and our testers' machines. That being said it's an early release and there are bugs. We welcome bug reports/suggestions/rants. We will be working extra hard in coming weeks to make the experience of using PioSOLVER as smooth as possible.

Feel free to post here or email us at piosolver@piosolver.com

Played with the software for about 30 minutes. Although I'm not completely sure why you would offer this for free, I must say the software is excellent. Very intuitive and powerful.

Here's to tougher games .

An amazing piece of software.

The GUI is a little ghetto but very streamlined and easy to use once you get used to it.

Looks like a cool program, unfortunately it's crashing every time I open it on my desktop.

Want me to PM/email you some kind of bug info I am receiving?

Yeah this is a bug with linking to our threading library. I have a fix now, will be pushing it in several hours (after making sure it works on more computers). Thanks for reporting, it was a difficult one to catch locally.

We get a lot of feedback in emails/pms but if you have any suggestions to make it easier and more intuitive feel free to post here

It'd be nice to be able to pick the board texture through a graphical interface instead of writing it out. It'd also be nice to be able to select specific suits in players' ranges instead of writing them all out.

I must also thank you for using the exact same syntax for the ranges and how they are represented alphabetically than you did in your ancient software Pokerazor, so I can just copy/paste ranges from there. That's great as I still use it sometimes and was able to do the suit-related things through it.

Well, so that was a bit of a falstart. Thank you to all the people who helped me debuggin this one. I pushed corrected version. It should work now. I've updated website links. Direct download link is here: PioSOLVER-free

If it still crashes for someone be sure to let me know!

Thx punter excited to play w it

Interesting. Thanks for the work. After having read the quick start and blog and having played around with it for 10 minutes, it would be nice to have a tutorial video explaining what the results signify and how to interpret results.

Yeah we will look into this once flop solver is released. It might be tough for me (I am not a native speaker) but once it gets picked up by training sites there will be material available.

command line interface?

Can't run it at the office, says it's not a correct windows 32 application or something like that.

subscribed...

But why didn't use it to crush the games, instead make it available for everyone?

That's just additional feature for advanced users who may want to write scripts (those will be useful for scheduling tasks with flop solver for example). You don't need to care about it as the GUI (PioVIEWER) hides it underneath.

Yeah, this is 64-bit application.
I've already got a request for 32-bit version. I don't have a 32-bit computer available but I will cross-compile and find one to test on. It's just not top priority as of now.
Unfortunately flop solver will not work on 32-bit computers (because they can't use enough RAM).

If you get this error on 64-bit Windows let me know (it shouldn't happen)

Turn solver is not very practical. It's great for playing around/illustrating GTO concepts etc. but you can't really use it to play and even learning practical things from it isn't easy.

As to the flop solver that one is very practical and basically solves something very similar to real NLHE (even though it's for now only one bet size it's pretty close to the real thing in many spots).
Still, the trees are 2-4GB big. The solutions take several minutes to calculate. I am confident I would kill the games if I could execute this strat but it's a lot of work to remember/analyse etc.
At this stage I enjoy programming more than playing and prefer to give it a shot and make an ultimate poker tool. It will be completely new era with flop solutions available as you will now be able to ask all the questions about building range/balance/ev of hands, something no tool or coach could give you before. It will still be a lot of work and effort to build your game around it and find proper adjustments from GTO to your opponents leaks. It won't be auto-playing for you. It will answer a lot of questions though.

Actually if you can just cross compile without testing i'll do it for you and give you any info you need regarding possible bugs. I'm a programmer myself (in Java tho).

Well I will not give you the source code as I am trying to build a small business around it but if you have 32-bit computer available and you would be willing to test it for me then that would be very helpful.
It's not so trivial because the solver uses OpenMP dll's which have bugs in MinGW/GCC distributions for Windows. I managed to make them behave in 64bit version (that was the cause of the crashes some people experienced in the first version) but they may not work in 32bit one. (if you are interested the bug is connected to running OpenMP threads from Windows threads or pthreads).
If you are up for testing 32-bit version pm me.

If I try to run PioViewer, it crashes (before starting), I thought I got the newest version, do you have any suggestions?

Thanks,

MT656

Try to upgrade your Net Framework to 4.5 (http://www.microsoft.com/en-us/downl....aspx?id=30653)
PioVIEWER should check for it but maybe we missed something. If it still doesn't work it would be great if you pm me so we can try to solve that problem (either via email or on Skype).
It's still very early in the release and we weren't able to test on all setups/Windows versions so those things are bound to happen. Let me know!

Tampered PioSOLVER will produce non-sense results in subtle ways (not immediately obvious to tampered binary user)
It will still be easily trackable to original license owner

I like your anti-piracy method ...
looking forward to the pro version

EDIT:
To see your benchmark time start PioSOLVER and type "bench"

Where do we have to type bench?

Free version doesn't implement bench command (because as of now it requires flop tree). I will look into making bench for free version as well.

Sorry, I wasn't clear:

Is there a command line interface available in the free version?

Yes but it's not scriptable and limited. Is there anything specific you would like to call from command line?

this is faster than anything I have seen before, congratulations

is this CFR?

I think it's going to be hard to get people to pay for this because the work and effort required to gain any benefit is so much.

Currently not although it used to be a variant of CFR for a long time (that's the reason for naming of our saves with .cfr extension). CFR turned out to be slower that what we currently do but there are so many parameters to test and so many variants of those algorithms that it may change in the future.

Maybe... it's 100x better than things like Snowie or CREV though and some people pay for that. We will see how it goes. This project is just starting, we have a lot of features in mind to make analyzing HH's easier.
It is true though that we are targeting a small market of analytically minded players (hence the price). We will see how it goes

yeah, i aggree these solutions are > than snowie. But when i talk to others, they say they don't care, snowie is good enough, faster and easier to use. And that that the time investment isn't worth the time lost playing.
So making it easy to use is where it could be made commercial. But that part of it isn't much fun.
For a human I'm not sure there is much benefit to having fast solutions (even though it's impressive!), because already going over a couple a day is exhausting and time consuming.

One benefit from having fast solutions would be that having some aggregate stats would be nice.

But good luck!

I read in one of your posts that you had done some preflop stuff? I didn't think that was possible. Any plans to share taht?