Hey,
i tried to create a small script, was already watching the small Youtube intro regarding those scripts.
I think i created one but when i click on ->Script->Run this script
a window pops up like in the youtube vid, but nothing happens.

[IMG][/IMG]



Did i missed a step, are my settings wrong?

Thanks for the help

Hey...there is nothing explained how to share a config.
already got it and will share.

That's the tree:
http://pastebin.com/0M4dcQXA

Dunno how I can add the completed board to the config. Rundown is 4h 7s and my info was that he does not bet top pairs except AJ.

Big thx for your help.
Think I already got the answer, that the OOP range just does not change that much even if IP does not value bet that wide which still amazes me.

Sup guys, I made a short video resuming my wall of text last page, my three question in the video are:

1) Why the solver dismiss the 1/3 and 1/4 cbet (btn vs bb on dry pair board).
2) Why the solver prefer 2/3 cbet to 1/2 cbet?
3) Why the solver is checking IP 40% of the time on the pair flop?

https://www.youtube.com/watch?v=3Yyj...ature=youtu.be

done, ty

maybe u should put in turn and river actions

I recommend spending some time watching the instruction videos. Possibly a few times each.

Hello, I've been using PioSolver Basic for a few weeks now and have one question. I've watched most of the faq videos and have tried searching but haven't found the answer to this. Sorry if it's been asked already or if I'm missing something simple.

I'd like to know if there is any easy way to force IP or OOP to do a certain action on specific next unknown cards.

Say I've constructed two ranges on a flop of Kh 5h 2d. On the turn, I want OOP to always check (I just use Lock Node), but then I want IP to bet their entire range on every turn card that is a Heart.

So far I've been manually selecting every turn heart card (3h,4h,6h, etc.. to Ah), then using Lock Node and forcing IP to bet all their hands on the card.

Is there an easier way to do this?

I'm having some trouble rebuilding the full tree from my small saves.

In particular I want to load a saved (small save) tree that solved from the flop. When I load that tree I'd like it's configuration to automatically go into the Tree Building tab so I can start solving it again and then do some node locking and resolving on various turn cards. (currently the tree building tab is just blank when I load a tree from small save)

How can I get the tree that is displayed in the browser to have its configuration loaded back into the tree building tab?

I don't want a turn and river action.

Yes, you do! You may only want to browse the flop action but you must specify a complete game tree including turn and river otherwise the flop can't be sensibly solved.

You must specify Turn and river action and also you have not specified your opponent action. By default you have forced checkdown everywhere or call/fold with no raises. I don't think that is what you want. You want a complete game tree so the flop part can be optimised as part of sensible play on later streets.

As you analyse you will see that your optimal play on the flop can vary quite a bit depending on what your opponent does and what you do on other streets. The flop cannot be analysed apart because it is all part of one big game tree. Push the game tree in one part and it pops out somewhere else.

For example, in your spot, OOP BB might donk lead somewhat but you have forced an OOP 100% check. You have also prevented a check raise in a spot where a high check/raise to your small bet might be expected.

Good luck. Respectfully, I suggest you watch a few Pio videos to get the basics down before posting.

Ok, I have watch a video, but now a new bug occur:

Server Error: ERROR: build_tree not enough memory; needed: 16847 MB; availble 5842 MB.

It seem like went I try to make a complete game tree (I have follow everything in the video I saw), I don't have enough memory.

For your stated 442r and D vs BB, start with a simple betting tree because it will make it easier for you to see what is going on conceptually. But also because the more betsize options the bigger the tree and the bigger the memory requirements: say one bet size on each street or maybe 2 on flop but one on each subsequent street. For OOP, force flop check by leaving blank. One betsize on OOP Turn and river. Put something in raise for OOP flop for your 443r, D vs BB as BB is expected to raise significantly.

Once you get comfortable with that you can start to make changes one step at a time in a structured way. Say, try just one preferred betsize on flop and compare the resultant EV to see if much different from the multi-bet option. Then try just one betsize on flop but make two sizes on turn. Again compare overall EV. And so on.

Or try modelling player types: what happens to optimal play if OOP player tends not to XR flop at all? What happens if OOP pre-flop play is Nitty? Laggy?

So first I am sorry for all the delays. I am travelling this weekend and have access to very spotty internet. This means the delays are bigger than usual and I am not able to run any serious analysis - just answer simple questions. With that disclaimer let me answer some of the things mentioned in recent posts:


Please refer to some recent posts about posting the config. The description - no matter how careful you are when writing it - will never be enough to replicate the tree. The way to share trees is to provide the full tree config which can then be pasted into PioViewer by going to Tools->paste treebuilding config. This allows for exact replication of the tree, solving it locally and looking at the results.

That being said, the why question is always the hardest one and we are not really better equipped to answer it than poker coaches and experienced players. Our job is to make the solver correct and to allow for displaying various kind of information to make analysis easier. Looking for the answer to the why question is hard and it's a job of poker coaches/players/theoreticians. Sometimes it's just too hard and some underlying pattern is still waiting to be discovered.

Basic version doesn't support scripts which is one of its few limitations compared to the pro one.

Yes, I am sorry for the wrong link.

Well, there are a lot of things out of those the solver shows which are hard to believe at first. That's the value of the tool: it shows you what is optimal even if it's so counterintuitive no human would come up with that. The problem is the solver doesn't have the concept of "why". I understand people always want explanations - sadly we just don't have them. We can only guarantee the results are correct.
Yes, you didn't fill in turn/river bet sizes. The results with forced check-down on turn/river are going to be way off from really is optimal. It's a good idea to follow the quick start video example and look at sample tree configs provided with the solver.

It was our design decision to make the betting lines the same regardless of what turn/river cards are. It has so many advantages when it comes to possible features (aggregation along the same line, overall EVs, switching between trees with the same structure, jumping from line to line changing turn/river cards, optimizations when it comes to implementation etc.) that it is worth sacrificing not being able to use different lines on different cards (you can always provide more bet sizes if you really need it).
There is not an easier way to do it right now.

For that you need a .txt config file which is created along with the save. Make sure to move them along with .cfr files if you change folders etc.

Thanks for answering in comprehensive and clear way, I wouldn't have explained it in a better way

Again, post the config. In general 16GB postflop tree is so huge it's almost a sure thing you made a mistake or provided like 3 or 4 betsizes everywhere. Postflop trees with one bet size are smaller than 2GB these days and even if you use more sizes you should comfortably be below 5GB if you don't go wild with all the options.

"When I load that tree I'd like it's configuration to automatically go into the Tree Building tab"
---> For that you need a .txt config file which is created along with the save. Make sure to move them along with .cfr files if you change folders etc.
---------------------------------------------------------------------------------------

I don't have any .txt config files. I do change folders of the files after I create them (always with a script.) I don't see any txt files except for the script txt file in the saves folder.

Are improvements/updates to the software free? Any big features or a PIO2 rolling out in 2017? Basically, should I hold off on a purchase haha.

Thanks for answering my questions punter. So far I like the program.

In your next update, or in an upcoming update, could you please create an option where you can set a Default Rounding Strategy? I imagine other users may have brought this up or have wanted this feature.

I always round to 1/1, because seeing whole combos is easier to work with, and put into practice. But I find myself having to hit Control+D, then check the box that says "flop, turn, and river" and then change the Rounding to "1/1" after each time I've pressed "Go" on the Postflop Tree Building and Calculations.

I'll change certain actions and move around hands in the tree all the time, so having to change the rounding strategy each time after I press "Go" (because the rounding strategy defaults back to its original 3 decimal setting) is a bit tedious.

If the script is still in the folder with other saves the config will be read from that. This only works if there are no other .txt files in that folder.

I've answered a similar question here:
http://forumserver.twoplustwo.com/sh...postcount=2110 (in the middle)

and here:
http://forumserver.twoplustwo.com/sh...postcount=1920

Rounding to 1/1 will really cripple the solution (in terms of exploitability at least). I am mentioning it to make sure you are aware of it
For now you can try using ctrl+b and type the command directly and then it will be in your history all the time although it is a bit messy solution. I will add your suggestion to the to-do although no promises as the list of things to improve is already long enough for now.

Hello! Bought PioSOLVER, but never got an email from you guys. I think it may have been sent to a now-defunct email address that's associated with my ancient paypal acct. Does that seem possible? If so, how can I best provide the correct information?

You should get automatic order confirmation email and then within 24 hours an email with license key, personal link, instrucitons etc.
If you haven't got them then please drop us an email to [email protected] providing some details to allow us to identify your order. The best would be full name or PayPal email. Another thing we will need is an email address you want your license registered to/sent to.

Hello Piotrek

Is it possible to see the Equity Realisation somewhere ? the only place i find this number is in the aggregated report.

Go to the spot you want equity realization calculated in (it might be beginning of the tree for the most common case but it can be calculated at any place) then open range explorer and choose this option:
https://gyazo.com/0952e4730258994e0becec067fb8adc2

You can then see equity realization for the whole range as well as individual combos, like that:

https://gyazo.com/846cbf309818b5ac0b8d67138dff25bb

hi, is it possible to have some kind of benchmarks for an hypothetical PC setup?
Let's say for example that i would buy a pc with an i7 6850K (6cores 12threads) and 32gb RAM, is it possible to know how fast would be a calculation of a tree like BTN 55% open vs BB 45% call? Or for example how much time it takes to calc a river stat like river cbet in a report aggregation over multiple files?
In other words is there a way to know how much faster these calculation would be in a PC more powerful than mine which i don't have for trying? (in order to understand if it be worth to buy a new one). thanks

As long as those are modern Intel CPUs speed is more or less proportional to number of cores multiplied by clock frequency. There is around 15% bonus for hyperthreading (i7 vs i5) and slight penalty once you get to a lot of cores. Newer generations are faster than older ones but starting from Ivy Bridge the differences are quite small.

The reason it takes a lot of time is likely because you are using small saves so the rivers need to be recalculated. It's worth considering changing recalc accuracy for rivers to something less precise (it shouldn't matter much for aggregated reports).
You can access recalc accuracy settings in Tools->Configuration->Behavior.

Something like 0.002 should be good enough for aggregated stats and re-solving will be much faster. The default settings are good for browsing because then you only need to solve one river at the time which is fast enough even to very good accuracy.

so in few words you are saying that having a i7 3770 (4cores 8 thread 3.9ghz) or a i7 6850K (6cores 12threads) it's more or less the same?

nope, i tried with full saves and 0.005 for river accuracy and it takes about 1 hour for a single river stat aggregation of a 3bet pot ranges..even 2 hours for a btn vs bb raised spot range (both on 74 boards aggregation).
So i think there's something strange that i don't understand :/

3770 is clocked at 3.4Ghz, it's the CPU on my main machine
You are likely talking about turbo frequency - that doesn't matter as CPUs can only sustain that for short period of time and probably for even shorter (or 0) when all cores are in use.

So 3770 is 3.4 x 4 = 13.6
6850k has 6 cores at 3.6Ghz each so 6 x 3.6 = 21.6

From that alone 6850k will be ~1.6x faster. There should be a slight bonus for 2 generations as well. One CPU worth considering if you can get it cheaper is 5820k as it's 6cores at 3.3Ghz and overclocks easily to 4Ghz if you are into that.
As you can see it's not easy to get big improvements without spending tons of money on dualbox setup.

well yeah, with full saves disk loading time is the bottleneck

Try changing the accuracy on small saves, it should provide a significant improvement.