Another thing: there is "Timeout in seconds" in the configuration.

Can we please have an option to let the solver run to a certain accuracy ? For example, since I don't really know how long solving will take, especially on very big trees, I'd like to just configure it so that it stops when 0.3% of the pot is reached.

I know this could be done with a script as well, but I'd rather have that option in the configuration if possible.

Small feature request:

the weighted strategy for all combos in a hand only shows if the screen size is big enough and actually only shows sometimes in the 13x13 grid like here: https://gyazo.com/80001f1b70f3a36f31e655730d231f34

and when it shows the text is usually too small, so there is some empthy space below the suit symbols, is it possible to have that overall weighted total for the selcted combo showing up in that space, something like this:

https://gyazo.com/9eccc7fdb5556266dc767d3b5bbf1e1c

Thanks in advance

Yeah, it's done in dev and will be there in the next release:
http://i.imgur.com/5ACsnXj.png

Yeah, we will see (we're a bit reluctant to add more strings to already crowded view) but this sounds like a good suggestion.
Btw, you can try moving the vertical bar between 13x13 view and the buttons to make it a bit wider - the numbers appear then as well.

Hi, first thx for your soft,

I would like to know the average(in all boards) capture factor (% of materialisation of preflop equity) of each hands vs same range,

The goal would be make preflop strategy with these results in crev,

How could i get these % in pio? (Gto)

And lets say i want to look witch better limp or or with each hands, vs given range of vilain, i cant choose all range for hero right? Because it will change vilain strategy postflop right? So how can i look the diference between ;(equity preflop %of each of my hands vs his range ) and (equity realized posflop vs his range)

Sorry for my ****ty english, i hope i make you understand my point

No preflop strategy planning to add in pio?

Thank you very much

You can achieve it now by calculating all the trees (1755), then making multi-flop aggregation report then weighting them manually (rainbow unpaired flop by 24, suited/paired by 12 and same rank/suit by 4). It is a lot of work though so it's the best to wait till next release where an option to insert weighted flops to our automatic script generation form will be added. List of all the flops with a corresponding weight is here:

http://pastebin.com/YwUWkD72

(notice that this list has one representative from every group of strategically different flops)
For task like this I would recommend very short solving times as the EVs should be more or less correct even at 2%-3% accuracy. It's also good to avoid huge trees as adding bet sizing options rarely makes big difference in EV but makes solving time longer.

I am not sure I am following you here so maybe elaborate a bit if I am not answering your question. You can see preflop equity in PioViewer by going to Tools->RangeExplorer->open (not in current spot) then inserting ranges and removing the board (preflop equityu works in range explorer in the newest version).

Next release will have simple pure preflop games added. We will see how it goes from there

Im sorry if this has been discussed earlier but is it possible to "lock" strategies and run the GTO versus that non GTO strategy?

For example UTG cbets 60% even though GTO freq is only 25%

search the thread for node locking and that should help

Yes, it's in the Tools->Node locking.
At the moment it's only possible to lock strategies for the whole decision (so you would need to specify which hands bet, which check in your example). We are planning to add locking for specific combos only as well.

PioSOLVER free now has some additional features:

Fully functional range explorer (works for flops as well)


yet i cant browse my solutions computed on another computer, ver 1.15

To browse postflop tree you need a full version of the solver. PioSOLVER free doesn't have functionality to understand flop trees.
A lot of data has to be calculated on the fly when browsing (as the saves only contain strategies) so we would need to put basically the full solver out there to make it possible.
The range explorer works for both flop/preflop in a free version but only if you input the ranges yourself.

hey punter.

i have i7 4770 and 16gb RAM, if i add extra 16gb RAM will it increase calculation time?

No. It will only allow you to build even bigger (and slower to solve) trees. You can use "Estimate tree size" button to see how big a tree constructed from given configuration would be. If you are not above say 13-14GB there wouldn't be any benefits from having more RAM.

Is it possible to rounding combos? As example, preflop we have 2 combos QTs, flop decision bet or check, pio suggest 1.987 combos bet, other check.
Would be easily that was displayed 2 combo bet and 0 check, anyway 0.013 closer to 0 and therefore it will never be in bet.
Need option to choose between rounding to the nearest tenth, hundredths or thousandths, or even to whole number combos.

Not during solving but once you solve a tree you can try Tools->Round Strategies, like this:
http://i.imgur.com/CkyjD1J.gifv

big thanks!

I'm trying to figure out setting up ranges - is there an easy way to create a turn range similar to crev where I give it a starting flop range and then tell it to get to the turn with something like any pair/FD/gutshot+/backdoor?

It seems really tedious and takes forever to try to create a turn range from scratch for a given flop unless I'm missing a simple way of doing it.

You can use Range Explorer for that (Tools->Range Explorer).
There are checkboxes with hand categories which you can check/uncheck and then copy-paste the range. You can edit it manually before copy-pasting as well.

Why would you want that though? After solving the flop you can just go to the turn and copy-paste from there.
It's even possible to choose % of the category by equity like in this gif:
http://i.imgur.com/BOeGHc8.gif

Hi

I am considering purchasing this program after trialing the basic version

I have one question with regards to multiple bet sizing for a given spot.

I understand pio solver allows you to compare multiple bet sizings versus other programs that allow 2.

How many can it compare on any given street and also how easy/presentable is the information?
I have done it in crEV before and it was a bit clunky to be honest (like the whole program actually)

Thanks in advance

You can solve trees with up to 10 different decision at any point (so 9 bet sizings and a check for example). The limit is completely arbitrary and I can make it higher if there is demand (it was very slightly slower when I made it 20 so I sticked with 10 but the difference is negligible). The presentation then looks like this:

http://i.imgur.com/7UFW2wn.png

The trees grows fast though and it's not very practical to use many bet sizes at many points for full flop trees. Even if you have enough RAM the solving process would be quite slow. Some of our customers solve things like 3-4 bet sizes everywhere but you need powerful hardware for that (64GB of RAM for big ranges) and it's still not very fast.

I am not sure if it's the best way to go about it either. Maybe it's more efficient to solve more trees with various combinations of 1-2 sizes to see what produces higher EV. Anyway, the functionality is there and you are just limited by your hardware.
PioSOLVER is very memory efficient as it is now but there is still a long way to go in this department so maybe in the future it will be more convenient to solve those huge trees.

Thank you for the fast response

Does pio solver have minimal exploit capabilities like simple postflop?

Thx again

Yeah, it has node-locking. You can lock any node you want, resume solving. Lock other/unlock, resume solving etc. In the near future we plan to add locking for specific combos as well (so it will be possible to lock only parts of the range instead of the whole decision node).
You may want to read this:
http://piosolver.com/pages/feature-overview

which is a quick overview of major functionality.

my 1st time using this.

3 bet pot i gave myself bb oop range of QQ-AA. vs cutoff. Jd 9d 3s Js 8h.....cbet flop. check call turn....on river the solver tels me to call more QQ 72% of the time. KK and AA around 53% of the time

also it prefers to call the Qd combos 99% of the time and the non Qd 42% of the time. doesnt this seem weird that it wants to call the combo that blocks the bluffs?
also for AA/KK it doesnt descriminate between suits. all 53%

Im not GTO versed at all but this solution seems weird to me.

any care to explain?

If you would like others to look at the hand you really need to provide exact configuration so we can re-run it. The easiest way to do that is to use "copy to clipboard" button and copy paste the output. It can then be loaded by using Tools->Paste tree building configuration

Again, it's impossible to tell without seeing the ranges. We don't know hat the best bluffs would be as they depends on what the ranges are in the first place.

I will take a look once you give me the full configuration but let me make a general remark. The "why" question is difficult. Most attempts to answer the "why" question out there are just exercise in rationalizing and confirmation bias. You can't hope to have every question like that answered, especially once you start solving bigger trees.
On the river (or maybe on the turn) it's usually not very hard to find out the answer (and it will probably easy in your case but we will see) but you can't build an expectation that it will always be that way.

The optimal is difficult and computers are way way better than humans as getting it right. We can hope to understand a little pieces of it here and there but not really a whole thing.

That being said, I am waiting for the config and we will see

#TreeBuilding#V2
#Range0#AA,KK,QQ,JJ,TT,99,AK,AQ,AJ,ATs,A9s:0.5,A8s :0.5,A7s:0.5,A6s:0.5,A5s:0.5,A4s:0.5,A3s:0.5,A2s:0 .5,KQ,KJ:0.5,KTs:0.5,K9s:0.5,QJs,QTs:0.5,Q9s:0.5,J Ts,J9s:0.5,T9s,T8s:0.5,98s,97s:0.5,87s,76s
#Range1#JJ,TT,99,88,77,66,55,44,33,22,AK,AQ,AJ,ATs ,A9s,A8s,A7s,A6s,A5s,A4s,A3s,A2s,KQ,KJ,KTs,K9s:0.5 ,QJs,QTs,Q9s:0.5,JTs,J9s:0.5,T9s:0.5,98s:0.5,87s:0 .5,76s:0.5
#Board#Jd 3d 9s Js
#Pot#164
#EffectiveStacks#318
#AllinThreshold#67
#AddAllinOnlyIfLessThanThisTimesThePot#500
#MinimumBetsize#0
#UseUnifiedBetAfterRaise#False
#UnifiedBetAfterRaise#
#ForceIPBet#False
#ForceOOPBet#False
#Cap#0
#UseCap#False
#FlopConfig.BetSize#
#FlopConfig.RaiseSize#
#FlopConfig.AddAllin#False
#FlopConfig.IncludeDonk#False
#TurnConfig.BetSize#50
#TurnConfig.RaiseSize#
#TurnConfig.AddAllin#False
#TurnConfig.IncludeDonk#False
#RiverConfig.BetSize#
#RiverConfig.RaiseSize#
#RiverConfig.AddAllin#False
#RiverConfig.IncludeDonk#False
#FlopConfigIP.BetSize#
#FlopConfigIP.RaiseSize#
#FlopConfigIP.AddAllin#False
#FlopConfigIP.Dont3bet#False
#TurnConfigIP.BetSize#50
#TurnConfigIP.RaiseSize#
#TurnConfigIP.AddAllin#True
#TurnConfigIP.Dont3bet#False
#RiverConfigIP.BetSize#100
#RiverConfigIP.RaiseSize#
#RiverConfigIP.AddAllin#True
#RiverConfigIP.Dont3bet#False

the config i used before had oop with a range of only QQ-AA which i guess was too narrow?

anyway. in this result we see that Qd Qs is the only combo it wants to call alot. Yet for Ad As its the only AA combo it wants to fold. folding the aces that block his nfds make sense . but why is the opposite true with Qd Qs?

First, it's useful to look at strategy+EV view:



Here you can see that it is really close for QsQh QSQd and QsQc. It wants to call with QsQd more but the reason for that is miniscule EV difference.
Interestingly though QhQd/QhQc/QdQc are quite clear folds and in fact when we look at pure equity:



We see that it's right on the edge comparing to odds needed to call for 3 combos but lower for 3 different ones.
We can try to investigate a bit using Range Explorer (Tools->Range Explorer->Open in current spot):



The situation with AA is a bit more complicated because we need to consider two things:

1)we don't want to block bluffs which are made with Asxs and Adxd (that's why AdAs is the worst bluff catcher)

2)We also need to consider which value combos we block (for example AJ) You can see that in the range explorer as well.

The differences in all the cases here are quite small as you can see (with the exception of calling with AsAd which is terrible or folding AhAc which is equall bad).
It takes some time to dig into this stuff and then another question arises: "why would IP bluff those combos and not other ones?".
As a general rule you should avoid bluffing missed flush draws in optimal play but sometimes you need to include some if you don't have enough bluffs. It's a complicated thing and really it's better to accept the fact we are not going to get every "why" question answered instead of making up some feel good theories which would only work in a case at hand.

I hope that helps a bit