I am not sure what you are decribing tbh. Solver stops when user defined timout is reached or use defined accuracy is reached (whichever comes first). Can you describe in a bit more detail the problem you are having?

Yes, I am putting it for the todo to update the docs with this and several more things.

Is this for cash games or tourneys or both?

I used the 49 flop subsets that I got from Piosolver site. I have stop calculation set to .5% of pot accuracy or a timeout of 600 seconds has passed. After running it overnight I noticed that it was at 10 of 49 for several hours, then resumed and has stopped again at 25 out of 49 done(for many hours).


How can I get it to complete all 49 flops? Or what can I do differently in the future to prevent it from stopping halfway through on future scripts?

Pio is a HU solver. We've added ICM as well so it's possible to solve ICM spots (both preflop and postflop). I think majority of our customers play cash but quite a few people find it useful for tournaments as well.

The solver doesn't stop like that. Are you sure you haven't clicked on the black console window in the meantime causing it to stop? It's a Windows "feature" that you can stop console programs like that. You can look into task manager if it's still running. Another possibility is that computer went into sleep mode. I introduced code to prevent it in 1.10 version though so if you are running the newest one:

https://www.piosolver.com/blogs/news...ersion-1-10-18

it shouldn't happen.

We need to understand what is happening. The solver doesn't just stop by itself. There must be external reason for it happening. First step is to make sure you are running the newest version.

Yeah I def clicked on the black window and stopped it. I clicked script>run this script and it seems to have resumed at the same point. tyvm

Unfortunately there isn't anything we can do about it as it's Windows' feature. A lot of our customers run into it though

Selected node I want to analyze, did Analysis>Runouts aggregated frequencies analysis over multiple files.

Box on the left says:

Perform analysis over 1 files in C:\Users\xxxxxx\Desktop.
Please note that it only makes sense if all trees are identical except for the board.
C:\Users\xxxxxx\Desktop\QdJh10d.cfr
Warning: Found more than one potential script file in the saves directory
Potentially a script file: C:\Users\xxxxxx\Desktop\#.txt
Potentially a script file: C:\Users\xxxxxx\Desktop\SHRB.txt

Box on the right says:

Didn't find information about flop weights

All boxes are unticked and I hit ok

The report I get has only 1 flop(the one thats in the browser). I tried this for two different nodes and same thing. Any help would be greatly appreciated.

Thank you

Hi,

The analysis over multiple files is a feature which loads solutions for different flops and calculates some averages. In your case you have only one file saved on the desktop so the report gives you results for one solution.

The analysis should be run on a group of solution files which have the same betting structure and different board. If you want to use this feature to it's full potential you should combine it with the scripting feature of PioSolver.

It is expected that you first build and run a script, and then run the analysis on the output.

You can see the quick guide on how to do it here https://www.youtube.com/watch?v=VhG_...outu.be&t=1430

Thanks, appreciate

Punter, do you have any inclinations as to how the solver would react if it solved for a 3rd person in the pot?

For example, today I raised a pot UTG and there was a call in the CO and BTN. The flop came: 8T5. PIO actually advised me to check in this spot and if I was going to bet, to bet small. I solved the simulation as if the hand was just against the CO. I wondered how the BTN (who folded on the flop) might have changed the strategy that PIO advocates for. Is there any inclination into this?


Secondly I have made the jump from basic to pro and now am considering edge. I wondered if you could give me the elevator pitch as to why edge is worth the extra $500. I don't find solving preflop spots as interesting as postflop, given that my preflop strategy is already reasonably competent. Why is edge so great that it retails for $500 more?

another thing I wanted to ask was about something you said earlier in the thread.

"we give total exploitability of a solution"

Is this a guarantee? And could you expand on exactly what this is?

ty, got it all squared away now.

From some tests I've seen it seems card removal effect from folds (sometimes called "bunching") wouldn't influence solutions that much but it would make solving way way slower.

Price is a function of (among other things) demand and effort it takes to implement stuff. Implementing and supporting preflop solver is hard and time-consuming. It requires huge hardware to run which makes the market for it small (not many people can get access to hardware powerful enough). Combination of those factors means we have to sell it for bigger price or otherwise it wouldn't make sense to develop it. The edge version is targeted at people who have this kind of hardware and/or run Pio on big servers and need to move it around daily from one to another.
Alternative business model would be to develop it privately and sell the solutions. I think most our customers appreciate we haven't followed that strategy

It's guaranteed that if you follow Pio's solution and play against perfect adversary (who knows your strategy exactly and is using perfect counter-strategy) you will lose exactly what "Exploitable for" value shows per hand. Of course the assumption is that we follow the betting structure in the tree you solved.

Hi,

I have rented a server running Intel Xeon E5 2630v4 (10 x 2.20 GHz) with 256 ram to use for mass solving of scripts with Pio Pro.

However I am experiencing a slower solving time than using my PC which is running an intel i5-8400 2.8ghz with 16gig of RAM.

Am I making a mistake in the setup?
Did I make a mistake renting this exact server?

Hope to hear from you

Greetings,

Roy

It's hard to say without more information.
Usual reasons for this happening are:
1)Energy saving settings (you need to turn them off, "balanced" is not good)
2)Some external antivirus which eats the resources
3)Cooling system not working correctly (when CPU overheats it throttles the performance to save itself)

Try checking those and see if that helps.

Please consider that basic version supports up to 6 threads and pro version up to 16. So if you are using those you may need to run more than one script at the same time get full benefit from the hardware.

Hi, I have a question bothering me. Let's say I have a decent preflop solution with full starting ranges for both OOP and IP. As I understand it guarantees me that my opponent can not increase his EV by changing his own preflop strategy.

But preflop solution is based on numerous postflop solutions. Let's say we are pretty sure that our opponent is not playing optimally. So his starting ranges for postflop spots will be drastically different. It is clear that this fact doesn't scare our preflop strategy, but what about postflop spots?

So, the question is - does preflop solution guarantee an unexploitability for each postflop spot as well? And in other words, does it make sense to study postflop spots using optimal ranges if we are sure our opponent does not play optimally (and so his range in the root of postflop scenario is different every time)?

This is true assuming exploitability is 0. If exploitability is X then your opponent can increase their EV by X per hand at most assuming they play alternating sides (one hand IP, one hand OOP).

They can play better postflop with a different range than the solution but they can make up at most what they have lost altering their EV preflop if it wasn't the case then the optimal solution preflop would be different.

This is relatively obvious when you ask the same question for flop and altering turn ranges. Unfortunately preflop we don't represent the whole game but only a part of it. We took care to provide flop subsets which are possibly close to the whole game so in general the mechanism applies unless the opponent can find a specific weakness of flop subset used and exploit that. The bigger the subset the less it's possible and less EV can be "leaked" that way.

It would guarantee it if run on all 1755 flops. As it is it almost guarantee it where almost is caused by imperfection of flop subsets.

It does make sense as you can't be sure what your opponent is really doing preflop and it makes sense to know what optimal strategy is assuming they are doing well. Now it also makes sense to know what to do if you can pinpoint their exact postflop range but you are guaranteed to do at least decently if you follow one against optimal range.

The main point here is that if it was possible for the opponent to gain more EV on later streets than they lose by altering their strategy on earlier streets then that change would increase their overall EV and it would cause the solution to be different in the first place. Exploitability is calculated across the whole tree so there is no possible way for the opponent to do something clever as we already calculate the best possible exploitive strategy for them.

Yes, but assuming we didn't know the BTN would fold would the flop strategy change much?

I am sorry I don't understand the question. If we are already on the flop we know if BTN folded or not. Can you explain?

The flop is 3 ways, I had to solve the solution against only 1 player. It said to check 80% of the time on a T58 board based on the ranges I assigned against the first player. The last player ended up folding so I just solved the solution pretending he was never there, but I wondered if you have any inclination as to how the flop action is generally altered on PIO solutions if we don't know if the third player will come along or not. For example in this hand I had a set of black 8s, and it said to check 80% of the time. Do you know how it might alter the strategy if it had to account for the third person?

We don't solve those spots.

The solution 3 way won't be applicable for HU spot. You can solve from the moment it's HU so after 3rd player folds but not before.

It would surely alter it but it's impossible to say which way.

Hi,

Can you please help me to understand these two things?

1.) I'm running a sim in a certain spot: on the flop OOP mixes checking and betting (various sizing) from a total of ~500 combos. When I then choose check and IP choses to bet, then I get a strategy (raise, call, fold) from OOP to respond to the bet only for a total of ~150 combos. What's happening with the other 350 combos?

2.) I want to check a hand that I played where I checked flop with QQ. When running the sim Pio then suggests to always bet QQ. But because I checked in game I want to see how Pio would continue on the turn after having played the flop the way I did. So I choose the tree for checking but then suddenly QQ isn't part of the turn range anymore. What am I doing wrong?

Thanks.

We would need some details, preferably the tree config so we can re-run ourselves and look at the numbers. As it is it's hard to guess from the description what is happening.

It's not part of the checking range because Pio always bets. Luckily you can still see the strategy for QQ in the branches it doesn't ever go into because this strategy is needed to determine going to those branches is not the best idea. If you click strategy you will see it. You can also uncheck "squares proportional to weights" checkbox to see it in graphical form.

I can't fit a external hard drive to my computer, is there a way that I can run a script and save it to some cloud service?

I don't know. The solver saver the files to the path provided. You will likely need to move it after they are saved.
Consider making small/micro saves though. Flop + turn saves are around 50x-150x smaller (depending on the tree) so a lot of them should fit:
https://www.youtube.com/watch?v=VhG_YsEvfwU (question number 5)