Two things to check are auto-update settings (make sure your computer won't shut down to update Windows if not directed to) and energy saving settings (sleep/hibernate should be off no matter how long the computer is idle although this one the newest version of Pio takes care of in general).

It's not possible to get add_lines directly from the tree. You can print all lines using a command like "show_all_freqs" (in the future there will be show_all_lines as well) but the best way is to just save the config when you make the tree and use that.

Hello , I have an issue with aggregate reports, my excel report looks kinda awkward like this- https://i.gyazo.com/ed130f186dd3453c...683794d595.png. I tried twice to do different aggregate reports but they look the same .

This is a bug in Excel, see here:
https://forumserver.twoplustwo.com/s...postcount=1666

In general Libre Office or Google Docs should work well out of the box.

Can u guys recommend me a good laptop for piosolver ? sorry about this q but discord is pretty dead

After I've node locked a tree, how do I quickly revert back to the original tree without rebuilding and rerunning the entire tree?

Laptops will always be slow and expensive in comparison to desktop computers so we don't really recommend them. If you want to get one and run the solver on it make sure to get real 4 or 6 core CPU (physical cores) clocked at 2GHz or more. Unfortunately I don't think it's possible to get both a nice looking slim laptop and a decent performing one. Get at least 16GB of RAM as well.

It's not possible but you can make a save before node-locking and then load the tree back.

Can anyone speak to the performance increase from a Threadripper 1950x vs the 2950x?

Somewhere in this thread I read that peformance for pio is basically a linear increase with clock speed and number of logical CPUs. So does that mean that a 1950x overclocked to 4GHz and a 2950x overclocked to 4GHz would produce about the same performance?

IS PIO SOLVER A REAL GTO SOLUTION?

Hi guys. I know Piosolver is the number one commercial "GTO software" out there, but WHAT IS THE REASON FOR TRUSTING IT?

As an old CREV user, I started using GTO+ instead of PioSolver because I had it free and I actually like more the interface and the way it works. I eventually bought Piosolver too. My doubt comes yesterday when I first tried to calculate the exact same tree in GTO+ and Piosolver. At least theoretically, I think solutions should be at least close one to each other, but what a big surprise when I checked they are completely different. I wrote to GTO+ support and I will paste here my question and their replies, but I would like to know what other users think, ands also Piosolver support point of view. Aside from Pio being the "famous" program, I really would like to have some clear evidence that one program is more trustful than the other one, since all my study is going to be based in the program.


1st message to GTO+ support:

Hi there. I'm trying to do some calculations with the program, but I'm a bit scheptical because for same tree, the solution it gives me compared to Pio is totally different. I thought I might have done something wrong, but everything seems to be right. If you can help me to figure out what's going on. I checked preflop ranges and are totally the same.
Just take a look at the flop cbet.
http://prntscr.com/njzyfq
http://prntscr.com/njzyp2
Thank you for helping


1st answer:

Hi,

The advanced tree building wizards for GTO+ and pio are different.
So, although the trees look roughly the same, there will be differences.
And different trees will lead to different results.

GTO+ and pio do however have the exact same basic tree builder.
For us, if you use the tab "Basic" in the tree builder (as opposed to "Advanced") and use "Version 1.0" in pio, then the trees will be identical.
The results will be nearly identical as well (there may be smallish differences in the choice of which hands to bluff with, but all EVs and frequencies will match).

Please let me know if you require further support,

Scylla


2nd message (still no answer):

Well, using basic versions is not an option since many times I need a more specific setup. The problem here, is that solutions are not even close, they are completely different, and that's a huge problem since I'm trying to get conclusions on how should I play my basic game. Some basic decisions like "which size is better for each board" or "which hands should I raise or call". So then, which is the good one? why should I use your program instead of Pio?

There are usually quite small differences between CPUs from generations next to each other. I am not sure what the difference is in AMD case though as the architecture is new so it's reasonable to expect bigger gains than usual (still probably less than 10% though). Unfortunately I don't have benchmarks ready.

We show the full solution, all the EVs, EQs, strategies. The results are quite easy to confirm for a programmer and it's possible to do by hand on some simple trees. We also verified the solutions with different implementations coming from other people.

I won't speak about GTO+ because and as this thread is about discussing Pio. I haven't used GTO+ once, I don't intend to. I will not comment on any solution produced by this software because I just haven't seen even one. The burden is on them to prove they know what they are doing.

Either you can verify it yourself or you have to trust someone. There was a lot of discussion about it in the past and a lot of people went through the solutions verifying their parts. We have more than one implementation of the solver and we compared the results with Jesolver which is done by a programmer who got his results published in Science. Jesolver works with the same interface as Pio so it's possible to have solutions side to side in the same program and go through them.

I won't comment on it because I don't understand the other program interface nor I intend to learn it. What I can do for you is to solve your tree using 3rd party very good solver (Jesolver) or one of many implementation of Pio I have in dev (they are different algorithms and produce a bit different results unless the accuracy is very close to 0 where they all converge to the same thing).
If you would like to see the solution to this tree solved to close to 0.00 accuracy please send us the tree config (you can get it by clicking "copy to clipboard" button in the treebuilding tab).

This sounds reasonable btw. Treebuilding form is a point where there are many decisions to take about how exactly the tree is built so it's likely you are solving a different tree.
We don't use 1.0 treebuilder anymore. It was just too simplistic for anything practical.
It's there as a legacy feature but I discourage you from using it.

Thanks again. Love the level of support.

The trees dont look identical at all. You have the "force OOP check/ IP bet" enabled in Pio, and nothing similiar in GTO+, at least I think so. And you have the "dont bet if previously passive in GTO+", but you have turn bets allowed in Pio.

ty,I was looking for this kind of answer. that's prolly the reason.

I hope this is the right place for this question, sorry if it has happened multiple times before. I've been running a couple of scripts, and I did set 0.3% accuracy for the calculation to stop. What happens is that every solution stops at 2.5%, making the results inconsistent. I did set the timeout to an allegedly large enough number, but it keeps happening. Does it have something to do with memory?

PIO - useless

oh wait - im wrong - of course you won the WSOP

You are likely hitting the timeout. When generating a script set it to something big and you won't have that problem. It's still useful to set it to something which can happen (2-3 hours maybe) so you avoid wasting time if a tree on a specific tree is much harder for a solver.

I mean, I can say that I tried a number of reasonable and big timeout amounts. From what I can see in the prompt window the solves are happening much faster than they are supposed to be, so Idk.

Hello,

Do you know if there are people selling custom preflop solutions made by Edge (not packs)? If so I'd like to contact one of them so I can send them the opening ranges, bet sizings and rake I want to be used so I can get an estimate and possibly buy them, any idea if that's possible and if so how to proceed please (in a safe way)?

Feel free to send us the script so we can take a look (only the beginning and first 1-2 trees are needed).
Possible reasons are that you have used incorrect accuracy settings. For example if you are using an older version and used a comma instead of a dot as decimal separator then "0,3" will be interpreted as 3 (we use a dot a decimal separator and older versions missed the error check). It's really hard to say without seeing an actual script.

I don't have anything to recommend at this point but maybe try asking on our Discord (#marketplace channel).

Thank you for your answer, I read all the market section on discord and I’ll ask for solutions when I know what parameters to ask for but I’d like your advice on that since I don’t know anything about Edge (I only have pro) and I want a knowledgeable unbiased opinion (I have limited financial means so I want to spend only the bare minimum) :

- I want 3-Bet solutions IP and OOP without CCR, with Call 3-Bets, 4-Bets, Call 4-Bets, and 5-Bet shoves, all of it rounded at 1/4. Given the rounding it doesn’t have to be extremely accurate so how many flops would need to be solved and how many bet sizings on each street would need to be put to get just that?

- And assuming one size on each street would be enough what sizings should I put on each street, and do I need to put «add all in» on the river? I don’t have a clue how much postflop sizing matters for preflop solutions.

- Assuming the open raise size is 2,5 bb and effective stacks 100 bb are there «optimal» preflop 3-Bet sizings preflop IP and OOP and if so what are they (in a very high rake environment)?


Sorry for being a pain in the ass, and I also have an unrelated question :
- I see good players everywhere cbetting range for 1/3, but I spent a lot of time on postflop trees and cbetting range for 1/4 actually has almost always more EV than 1/3. Is it because I only use 1 size on turn and river that my results are inaccurate, or are they accurate and the good players use 1/3 instead for a specific reason, and if so do you know what is it?

Thank you for your time and patience.

Hello,

Unfortunately, I don't have Piosolver at the moment.

I have a question :

In the pre-Pio era, there was a general thought saying that "the more you bet on a flop texture, the smaller the bet sizing. And vice versa."

Can someone let me know if PioSolver confirms this principle?

I'd like to read answers from Piosolver users, but an answer from punter11235 would be much appreciated as well.

If this is not the right place for such question, just let me know, I'll post somewhere else.

ty

How do I find out what the updates for PIO are? Is there a record for the updates for each update?

My view is that it's already very precise at 72 flops and close to perfect on 120 so somewhere between that. We run most of our sims on flop subsets between those.

It's the best to simplify postflop play. Use one bet size and eliminate donk bets in at least in most spots. That doesn't influence the preflop ranges much but lets you use more flops.

I don't know, I haven't run enough sims and especially not for high rake environment (solving for that is way less reliable because the game is far away from 0-sum and thus guarantees about equilibrium are gone). In general the solver likes bigger 3bets than people used to use.

I think 33% was natural thing to try and it was better than bigger sizes (on most flops anyway). 25% might be even better but the difference is probably not that big.

Yes, this makes sense and is confirmed on the solver. The solver very often wants to choose very small bet sizing on the flop (25% or 30% for example) and bet almost the whole range.

blog on our website and a #announcements channel on our Discord. The new version also detects the updates available (I think it checks for them every week or so) although it doesn't auto-update.

I checked the blog and there is no record for the latest update.

Is it possible to set up something similar to PT4, where users get a breakdown (in list format) of what has been updated after each software update. This is pretty common amongst other software I use outside of Poker.

Is there any other record of what has been updated beside discord?

There is always an announcement about feature releases, latest ones are just bugfixes. We prefer fixing things quickly and shipping silently as it only affects very small number of users with specific issue. We will think about publicly available changelog as well.

sorry for incoming vague question, but im looking at 3bet simulations in pio solver with static ranges, i.e. SB 3bet vs btn, whats easiest way to see across different flop textures to see what bet sizing is preffered i.e. 1/3, 1/2, 2/3, 3/4, pot.

I've been putting them into solver and sometimes you see that multiple sizings are viable, but usually one is preferable, and then I re-run simulator with just the preferable sizing and analyze that. it would most likely be more +EV to have multiple sizings just can be very difficult to implement as a human.

So regarding 1755 iso morph flop subset, or 184 flop subset, what would you recommend is best way to go about looking at this? manually have to do the above and make my own notes. not exactly sure how scripts work as I am rather new to this software, so any advice here appreciated. Last Question - are there currently any smaller flop subsets where you type in any random flop and it gives you the corresponding flop in that subset that is most similar?