Yes, that's what this is all about. I'm fairly certain those monker's calculations have nothing to do with gto. That's not even close.

Question to punter11235:
Given the current way pio calculates preflop, how much memory would it take to build a very simplistic 6-max tree, say 1 size/1 raise allowed pre, only 2 players postflop etc. ?

what is analysis>runouts aggregated frequencies analysis over multiple files?
im assuming its aggregate data pulled from the scripts , like the aggregate flop cbet stat etc?

however, when i click GO it says that the weight wasnt assigned to one file, when i uncheck "use weights from the script file" it actually runs

-do i need excel to view this report?
-how do i resolve the root hand not having a weight assigned?
-what other tips/tricks/uses and suggestions, for this feature?

thanks

Well, it is a judgement call. One point of reference is that at 5/10 game in a postflop pot rake is 30bb/100 (3$ per hand, 300$ per 100 hands. That means that gain smaller than 15bb/100 for one player is smaller than rake paid. It's still a judgement call but maybe you find it useful.
Unfortunately right onw our fastest compile works well with up to 64 threads. This means 64 cores without hyperthreading. We tried making a compile working with more threads but it was very inefficient in our tests. Hyperthreading is such a small gain though and the work required to make it efficient on 64cores+ht is significant so it's very unlikely to happen. If there is 64core ThreadRipper it will be the best to disable hyperthreading on it to get max performance with Pio.
One good thing is that Pio will be much faster in foreseeable future and that coupled with improving hardware will make huge difference.

I answered it several posts back. There is software on the market that produces 6max GTO approximations. There are many concessions to make it possible (abstractions to lump hands together, algorithms with no guarantee of convergence etc.). If those solutions are good or not reamins to be seen. I don't have any opinion on it other than I tend to believe claims Monker author made (no guarantee of convergence, abstractions but some nice properties like pure strategies being right).

I am not sure what you are asking. We are of course taking card removal into account when solving, calculating equity or EV. I can't imagine doing it without taking card removal into account as it's such a fundamental part of poker and influences solutions in great deal.

You can kinda estimate but looking at number of preflop exits in such a tree and multipliying by how much currently one preflop exit takes. I think it won't be feasible without introducing abstractions to reduce memory usage.


Yes, please look at FAQ question, here;
https://www.youtube.com/watch?v=5Ozjfnatwms (question number 5, links in the video descrption).

You can use anything capable of reading .csv files. Libre Office is one good and free option. I am sorry I don't understand what root hand is. In general you need to run a script first and then run a report over the produces solutions.

Where can i find syntax guide for PLOcalc, normal PPT syntax expressions doesnt seem to be working?

It's in the program.

Trying to determine if upgrading to edge makes sense and how much that investment will truly cost. What do you think the minimum hardware requirements for preflop solves? Generally how much time/money does this take on rented hardware? Do you recommend any of the "pre-solved" packs that may be floating around?

Hi, has anyone had any experiences with frying their computer over time due to running too many sims on it? Weird question I know. My computer sounds like it's giving birth sometimes when I run big trees, and I'm wondering if over time I'll just kill the computer from having run so many.

I’ve had computer running 24/7 for a year with no problems. If you think something is wrong the first thing I’d check for is Dust buildup inside the desktop.

Thanks for the reply. No, nothing at all has happened that makes me think something is wrong. I just hear the fans going crazy as I run these trees and wonder if--over the long term--I'm doing the computer damage by running them and if it's worth buying a computer (or renting computing power over the cloud) purely to run sims (and maintain the health of my personal computer).

I would also like to point out that we don't use "normal PPT syntax" as you call it be cause that syntax is very limited and not suitable for describing postflop ranges (where suits matter).

You should definitely monitor temperaturs to see if it doesn't get too hot. If it reaches >70C on regular basis (I mean for desktop computers, you often can't do much about laptops other than buying a cooling pad) it means cooling doesn't work well and/or is dirty and needs cleaning (both the heatsink and the fan). Modern CPUs shouldn't burn as even 95C is a safe temperature for them and they start throttling the performance if they get too hot.
If the fan gives up that's only like 25$ to get a very good one for replacement. CPU really shouldn't burn.

I told you I could hear your fan crying for help!

Punter I can't figure out the right syntax for equity (eq) and equity percentiles, can you help me out? What's the syntax to node lock for "Equity is between 40% and 50%"?

"eq > 40 & eq < 50"



Please keep in mind that these expressions are not conditions for solver, but rather range edition sugar.
The locked range will be "equity between 40 and 50" at the moment of editing.

After the range is saved it's not being automatically adjusted when solution and equities change later.

Hello Punter,

I tried searching for this but didn't see any questions in the thread.

I have a 32 core 64 thread CPU. If I run two instances of pio pro and run solves simulataneously, do you think it will use up all 32 cores?

Hi
I am attempting to run and aggregated report and receive this message

The given board is specficied with two different weights in the scripts (2.39!=2.03)
appreciate the help

Any advice for building a PC with solvers as the biggest goal/task? Wouldn't do much gaming. Was looking at a threadripper 1920x with 32gb ram but that would be pushing my budget of say ~$1k-~2k CAD. Advice on best chips?

Yes, it should use 32 cores without hyperthreading. You can also try running 3 or 4 to get max efficiency (if you have enough RAM for that).

It's important that a script and the saves produced by running it are in one folder and that this folder contains only those saves and this one script. The error is likely caused by more scripts being present and the solver not being able to determine which saves came from which script.
You should get pretty good deals for older Thread Rippers now, just look around, I've seen some available for 200$ already. If you are willing to wait a bit more then it's likely those will be even cheaper once the new generation comes out.
In general Threadrippers/Ryzens are by far the best when it comes to value/cost ratio. It will alwys be cheaper to get parts separately as well (not prebuilt systems) as this way you can shop around for good deals.

With that said, what if I wanted to run a flop subset let's say the 184 flops, how do I accurately split them into two scripts?

Should I just:
-Copy first half (92 flops) and run script on first Pio
-Copy second half (last 92) and run script on second Pio

OR

-Copy all flops into both scripts
-Reverse the order of second script
-Stop when both finish 92 flops

Will these two methods produce the same results?

I've noticed that sometimes when I give PIO two flop bet sizes, and then run it again but this time only giving it one bet size, the more frequently chosen one, the EV actually goes up a tiny bit.

If this is the case, why doesn't PIO just ignore the option of the other bet size?

what do the strategies/ev mean for hands that have 0 weight. For example, if you are at the river, you are able to see strategies for hands that have totally folded. Aren't the strategies for individual hands dependent upon what other combos are in the range, so if you node lock to not fold that particular hand by the river, shouldn't the overall strategies change?

That should not happen, you must be doing something wrong.

If your opponent´s range is fixed, they are not. Maybe Pio displays how would these hands do against villain's river strategy?

I am not talking about fixed ranges. I understand it is showing how the hands do against villain's river strategy, but shouldn't that be dependent on what hands you have remaining in your range?

Sorry for posting a duplicate but looks like my question got lost above.

Trying to determine if upgrading to edge makes sense and how much that investment will truly cost. What do you think the minimum hardware requirements for preflop solves? Generally how much time/money does this take on rented hardware? Do you recommend any of the "pre-solved" packs that may be floating around?

Well, running more Pios at the same time is not a usecase we really want to spend time on supporting as we have an edge version for that.
That being said you can just copy the script, call it script2 and remove half the flop from script1 and the other half from script 2. That won't be that much editing.
If there is rake or ICM it may be the case that the solver finds another equilibrium (starting from a different point it's likely). If there is no rake nor ICM then the only explanation is accuracy (it didn't yet converge enough to have the same or lower EV).

EV is independent on weight. You can always calculate it even if you never have a hand in a given spot by answering a question "how much this hand would make if I had it here". To calculate that you only need opponent's range.
As to strategy it answer the question: "how should you play if you had a given hand at this point". This evolves to max exploit over time as there is no need for the opponent ot adjust (as we never have that hand anyway).
They are (becaus the opponent adjusts) but that's not a problem. You still have other combos in range even if you don't have one particular hand there.
Well if you resolve they will change. It doesn't mean the strategies for hands not in range can't be calculated though.
I realize it's all not obvious at first so feel free to ask follow up questions if you have problems visualizing it.

See hardware subsection in FAQ on our website:
https://www.piosolver.com/pages/faq#hardware
(PM if 2p2 removes the link and you can't find it)
It's really hard to tell because it depends what exactly you want to calculate.
As to packs: I don't recommend anything specific. You can download some we offer for free though, from here:
https://www.piosolver.com/blogs/news/piosolver-1-10-13

EV depends only on opponents strategy. You can take his river strategy and ask "how much EV every possible hand has versus this strategy" - this surely doesnt depend on the rest of your range, since we are considering a fixed river strategy for villain and calculating our EVs against it.

I recently bought this new PC and was wondering if the requirements are acceptable to run Pro.

4.1 GHz AMD Ryzen 7
RAM 16 GB DDR4
Hard Drive 1 TB 1TB (64MB Cache) 7200 RPM SATA 6Gb/s
Graphics Coprocessor AMD Radeon RX 580 4GB GDDR5
Chipset Brand AMD
Card Description Dedicated
Graphics Card Ram Size 4 GB
Number of USB 2.0 Ports 4