GTO PioSOLVER - postflop equilibrium solver for Holdem - Page 35 - Poker Software -- Two Plus Two Poker Forums

The idea to support plugins didn't work out but you may get some communication code in C# from us (unsupported). You can also write it yourself as it's a simple standard input/output text interface.

I would highly appreciate that

Quote:

Originally Posted by punter11235

If you describe a specific tree I will tell you how much RAM it takes to build it.

to get a feeling for the sizes I would like to see the estimates for the following two trees

Code:

#TreeBuilding#V2
#Range0#AA,KK,QQ,JJ,TT,99,88,77,66,55,44,33,22,AK,AQ,AJ,AT,A9s,A8s,A7s,A6s,A5s,A4s,A3s,A2s,KQs,KJs,KTs,K9s,K8s,K7s,K6s,K5s,K4s,K3s,K2s,QJs,QTs,Q9s,Q8s,Q7s,Q6s,Q5s,Q4s,Q3s,Q2s
#Range1#55,44,33,22,ATo,A9,A8,A7,A6,A5,A4s,A3s,A2s,KQo,KJo,KTo,K9,K8s,K7s,K6s,K5s,QJo,QTo,Q9s,Q8s,JTo,J9s,J8s,T9o,T8s,T7s,98,97s,96s,87s,86s,76s,75s,65s,54s
#Board#Ac Ks 6h
#Pot#750
#EffectiveStacks#9700
#AllinThreshold#50
#AddAllinOnlyIfLessThanThisTimesThePot#1E+08
#MinimumBetsize#100
#UseUnifiedBetAfterRaise#False
#UnifiedBetAfterRaise#
#ForceIPBet#False
#ForceOOPBet#False
#Cap#0
#UseCap#False
#FlopConfig.BetSize#33, 50, 65, 100, 135, 200
#FlopConfig.RaiseSize#33, 50, 65, 100, 135, 200
#FlopConfig.AddAllin#True
#FlopConfig.IncludeDonk#False
#TurnConfig.BetSize#33, 50, 65, 100, 135, 200
#TurnConfig.RaiseSize#33, 50, 65, 100, 135, 200
#TurnConfig.AddAllin#True
#TurnConfig.IncludeDonk#True
#RiverConfig.BetSize#33, 50, 65, 100, 135, 200
#RiverConfig.RaiseSize#33, 50, 65, 100, 135, 200
#RiverConfig.AddAllin#True
#RiverConfig.IncludeDonk#True
#FlopConfigIP.BetSize#33, 50, 65, 100, 135, 200
#FlopConfigIP.RaiseSize#33, 50, 65, 100, 135, 200
#FlopConfigIP.AddAllin#True
#FlopConfigIP.Dont3bet#False
#TurnConfigIP.BetSize#33, 50, 65, 100, 135, 200
#TurnConfigIP.RaiseSize#33, 50, 65, 100, 135, 200
#TurnConfigIP.AddAllin#True
#TurnConfigIP.Dont3bet#False
#RiverConfigIP.BetSize#33, 50, 65, 100, 135, 200
#RiverConfigIP.RaiseSize#33, 50, 65, 100, 135, 200
#RiverConfigIP.AddAllin#True
#RiverConfigIP.Dont3bet#False

Code:

#TreeBuilding#V2
#Range0#AA,KK,QQ,JJ,TT,99,88,77,66,55,44,33,22,AK,AQ,AJ,AT,A9s,A8s,A7s,A6s,A5s,A4s,A3s,A2s,KQs,KJs,KTs,K9s,K8s,K7s,K6s,K5s,K4s,K3s,K2s,QJs,QTs,Q9s,Q8s,Q7s,Q6s,Q5s,Q4s,Q3s,Q2s
#Range1#55,44,33,22,ATo,A9,A8,A7,A6,A5,A4s,A3s,A2s,KQo,KJo,KTo,K9,K8s,K7s,K6s,K5s,QJo,QTo,Q9s,Q8s,JTo,J9s,J8s,T9o,T8s,T7s,98,97s,96s,87s,86s,76s,75s,65s,54s
#Board#Ac Ks 6h
#Pot#2200
#EffectiveStacks#8950
#AllinThreshold#50
#AddAllinOnlyIfLessThanThisTimesThePot#1E+08
#MinimumBetsize#100
#UseUnifiedBetAfterRaise#False
#UnifiedBetAfterRaise#
#ForceIPBet#False
#ForceOOPBet#False
#Cap#0
#UseCap#False
#FlopConfig.BetSize#33, 50, 65, 100, 135, 200
#FlopConfig.RaiseSize#33, 50, 65, 100, 135, 200
#FlopConfig.AddAllin#True
#FlopConfig.IncludeDonk#False
#TurnConfig.BetSize#33, 50, 65, 100, 135, 200
#TurnConfig.RaiseSize#33, 50, 65, 100, 135, 200
#TurnConfig.AddAllin#True
#TurnConfig.IncludeDonk#True
#RiverConfig.BetSize#33, 50, 65, 100, 135, 200
#RiverConfig.RaiseSize#33, 50, 65, 100, 135, 200
#RiverConfig.AddAllin#True
#RiverConfig.IncludeDonk#True
#FlopConfigIP.BetSize#33, 50, 65, 100, 135, 200
#FlopConfigIP.RaiseSize#33, 50, 65, 100, 135, 200
#FlopConfigIP.AddAllin#True
#FlopConfigIP.Dont3bet#False
#TurnConfigIP.BetSize#33, 50, 65, 100, 135, 200
#TurnConfigIP.RaiseSize#33, 50, 65, 100, 135, 200
#TurnConfigIP.AddAllin#True
#TurnConfigIP.Dont3bet#False
#RiverConfigIP.BetSize#33, 50, 65, 100, 135, 200
#RiverConfigIP.RaiseSize#33, 50, 65, 100, 135, 200
#RiverConfigIP.AddAllin#True
#RiverConfigIP.Dont3bet#False

Thanks ahead

Quote

01-28-2016 , 06:18 AM

#852

punter11235

Carpal \'Tunnel

Join Date: Mar 2005 Posts: 8,210

Quote:

I would highly appreciate that

If you are our customer then email us from registered address (or provide your key) and you are likely to get it in few days.

Quote:

to get a feeling for the sizes I would like to see the estimates for the following two trees

Those trees are humongous (7 bet/raise sizes everywhere).
The first one would be 94GB (introducing cap of 4 reduces it to 88GB).
The second tree is 5.3GB (again it has 7 sizes everywhere but shorter stacks this time).

My view is that building trees like that is pointless and tbh I haven't expected anyone using even 3 sizes at most points (let alone raise sizes) but here you go. They first one will take about forever to solve

Quote

01-28-2016 , 07:13 AM

#853

keerioh

newbie

Join Date: Jan 2016 Posts: 18

Quote:

Originally Posted by punter11235

If you are our customer then email us from registered address (or provide your key) and you are likely to get it in few days.

alright, I will come back to this

Quote:

Originally Posted by punter11235

I am not seriously planning to solve trees like that. it was just to get a feeling for tree sizes

thanks for your efforts

Quote

01-28-2016 , 09:36 AM

#854

livestream

grinder

Join Date: Oct 2007 Posts: 543

Quote:

Originally Posted by punter11235

They first one will take about forever to solve

How long is forever? days weeks month? more or less as an answer works for me

Quote

01-28-2016 , 03:37 PM

#855

punter11235

Carpal \'Tunnel

Join Date: Mar 2005 Posts: 8,210

Quote:

How long is forever?

5 hours on 16 core Xeon (it got to 0.3% of the pot).
It would be about 1.5x faster on a suited or paired flop.

Quote

01-30-2016 , 11:24 AM

#856

poker-hero

journeyman

Join Date: Jan 2010 Posts: 382

Hi,

I have studied 10 spots BTN vs BB.
I have exported several multiple files runout aggregated reports for several nodes of these 10 confrontations.
In each report, I found different values for holecards Equity Realization.
I guess the node chosen before launching the report impacts the Equity Realization calculation.

So my question : how should I proceed to get the average global Equity Realization for each preflop starting hand in these 10 spots?
And this for each position (separately of course) : BB and BTN.

Quote

01-30-2016 , 12:28 PM

#857

Kalupso

journeyman

Join Date: Nov 2012 Posts: 337

Hi,

I would be grateful if anybody would share the difference in EV from SB in HUNL (no rake) with and without limping preflop from simulations. I'm trying to figure if it is worth the time and money to implement it between 40bb and 75bb in my game.

Quote

01-31-2016 , 05:20 AM

#858

punter11235

Carpal \'Tunnel

Join Date: Mar 2005 Posts: 8,210

Quote:

I have exported several multiple files runout aggregated reports for several nodes of these 10 confrontations.
In each report, I found different values for holecards Equity Realization.

Yes because those are different for every point in the tree.

Quote:

I guess the node chosen before launching the report impacts the Equity Realization calculation.

Yes, because equity realization is just EV/EQ and those are different in different parts of the tree.

Quote:

So my question : how should I proceed to get the average global Equity Realization for each preflop starting hand in these 10 spots?

Run the report in the root of the tree.

Quote:

I would be grateful if anybody would share the difference in EV from SB in HUNL (no rake) with and without limping preflop from simulations. I'm trying to figure if it is worth the time and money to implement it between 40bb and 75bb in my game.

It depends how big the sizing for an open is. If it's only 3bb/fold for example then limping will be very valuable (in 100bb it's about 2.5bb-3bb/100 depending on exact structure). Once the open sizings get smaller it less valuable.
With shorter stacks it becomes valuable again even with a minraise so it is really hard to answer in general

Quote

01-31-2016 , 08:55 AM

#859

shahrad

old hand

Join Date: May 2008 Posts: 1,590

Hi,
I guess it is needless to say how good your software is. Thank you for your efforts.

I have a feature request:
1) I would like to be able to see the equity of betting ranges/combos vs Calling/Raising/Folding ranges.
2) I would like to have the possibility that villain (oop) cannot bet turn/river when calls previous street but can bet turn/river when raises

with regards

Quote

01-31-2016 , 05:34 PM

#860

punter11235

Carpal \'Tunnel

Join Date: Mar 2005 Posts: 8,210

Quote:

1) I would like to be able to see the equity of betting ranges/combos vs Calling/Raising/Folding ranges.

You can just go to that branch and then click EQ button for the other player. Like this:

http://i.imgur.com/X8P76zc.gifv

In this GIF OOP is cbetting and then I show how to see what the equity of cbetting range is vs folding range, calling range and raising range respectively.

Quote:

2) I would like to have the possibility that villain (oop) cannot bet turn/river when calls previous street but can bet turn/river when raises

That's exactly what those checkboxes for OOP do. Just try them

(you can build a tree and before solving it go browse it to see if it is what you wanted)

Quote

02-01-2016 , 01:01 PM

#861

shahrad

old hand

Join Date: May 2008 Posts: 1,590

Wow!

Quote

02-01-2016 , 03:04 PM

#862

shahrad

old hand

Join Date: May 2008 Posts: 1,590

Meanwhile I tried a flop range like PLO. But it somehow never comes under 6% exploitability (typical holdem ranges take 5 minutes to come to 0.3, this is now running since 3 hours at 6.2% constantly). I would like to know if it is my system (i3/16gb ram [never uses more than 130mb ram] ) or it is nearly impossible to find unexploitable lines when there are lots of draws and high equity hands.
OOP: AA,KK,QQ,JJ,TT,99,88,7h7c,7h7d,5s5c,3d3c,3s3c,AdKc ,AdQc,AhJc,Ah9d,Ac7d,Ac7h,Ad7c,Ad7h,Ah7c,Ad6h,Ad5h ,Ac4d,KdQc,KhQd,KdJc,KcTd,Kd9h,Kd7c,Kh7c,Kh7d,Kc5d ,Kd5c,QcJd,QdJc,Qh9c,Qd6h,Qc5d,Qh5d,Jd9h,Jh6h,Js6d ,Jh5s,Tc5d,9s8h,9h7c,9h6h,8h7h,8c7d,8d7c,8d7h,8s7d ,8c5d,7c6d,7d5c,7h5c,7h5d,7h5s,7d3c,7d3h,7h3d
IP
AA,KK,QQ,JJ,ThTc,TsTc,ThTd,TsTd,TsTh,9d9c,9h9c,9s9 h,7d7c,7h7c,7h7d,5s5d,3h3d,AdKh,AdQh,AdJh,AdTh,Ad9 h,Ah8d,As7c,Ah6h,Ah6d,As5c,KdQh,KdQs,KdJh,KsJh,KdT h,Kd9h,Kh8d,Kd7h,Kh7c,Kh5d,Ks5c,QcJd,QdJc,QsJd,QcT d,QsTd,Qc9d,Qh8d,Qc7h,Qh6h,Qc6d,Qc5d,Qd5h,JcTd,Jc9 d,Jc8d,Js8h,Jh7c,Jh7d,Js7c,Jc5d,Tc8d,Td8c,Ts8h,Th6 h,Th6d,Tc5d,Td4s,9d8c,9h7d,9h6h,9d5s,9d4s,8c7d,8c7 h,8d7h,8h7c,7c6d,7c5h,7c5s,7d5c,7h5c,7c4d,7h4d,7c3 h,7d3c,7h3c,6h5d,5c3d,5d3h,4d3s
Board 7s 5h 4h
Starting pot 1 effective stacks 13, everything else at 100% and oop cannot bet on the flop.
The ranges on the flop are very close to 15RFI OOP and 30RFI IP from pokerjuice on a T72r board.

Last edited by shahrad; 02-01-2016 at 03:28 PM.

Quote

02-01-2016 , 05:19 PM

#863

punter11235

Carpal \'Tunnel

Join Date: Mar 2005 Posts: 8,210

Yeah, so first when you are reporting a case it's the best to click "copy to clipboard" button in tree building and calculation tab and then paste it (either using code tags here or email it to us or paste to pastebin and copy a link).

As to your specific case: unfortunately in this version the solver isn't able to detect assymetric ranges (ranges which have different weights for strategically the same hands like Jh6h and Js6s). The solver assumes by default that flop ranges are going to be symmetric but doesn't assume that about turn ranges. This means that if you want to use assymetric ranges on the flop you need to disable isomorphism. This is done as follows:

1)Tools->solver->Run arbitrary command
2)set_isomorphism 0 0
3)build tree now
4)run it

It will be taken care of in the next release. I haven't predicted a use for assymetric flop ranges (they don't make sense in Holdem but do make sense in Holdem like toy games).

Quote

02-03-2016 , 07:17 AM

#864

AsUs91

newbie

Join Date: Jun 2012 Posts: 48

can u explain me what the ev number say, in the strategy+ev window in a preflop solve.
for example AA raise 56.773
stacksize is 15bb.
is a chip always 10? so raise aces will give me 5,67big blinds?
Or what this 56.773 means.

Quote

02-03-2016 , 07:29 AM

#865

Decoy

journeyman

Join Date: Sep 2011 Posts: 347

After updating pio on my laptop was required to re-enter my activation code and got this error: http://************/image/s2imuz30d/

What does it mean?

Quote

02-03-2016 , 03:10 PM

#866

punter11235

Carpal \'Tunnel

Join Date: Mar 2005 Posts: 8,210

Quote:

can u explain me what the ev number say, in the strategy+ev window in a preflop solve.
for example AA raise 56.773
stacksize is 15bb.

PioSOLVER doesn't understand the concept of "bb". It's always in chips. You provide starting pot in chips, stacks in chips and the EVs are in chips as well.
If you want to calculate a spot from say 5/10 game and you insert 60 starting pot then 56.773 EV means 5.6773bb.

Quote:

After updating pio on my laptop was required to re-enter my activation code and got this error: http://************/image/s2imuz30d/

What does it mean?

The link doesn't work.
Please send a link to support@piosolver.com following the guidelines here:

http://piosolver.myshopify.com/pages/contact-us (point 2)

Quote

02-04-2016 , 08:55 AM

#867

TheDefiniteArticle

Indecisive

Join Date: Feb 2012 Posts: 14,123

Is there an option to have the solver play a sound when it's hit the requested accuracy?

Quote

02-05-2016 , 10:09 AM

#868

punter11235

Carpal \'Tunnel

Join Date: Mar 2005 Posts: 8,210

Quote:

Is there an option to have the solver play a sound when it's hit the requested accuracy?

Not in this version.
In the next one you will be able to do this:
https://vid.me/LL1L

Quote

02-05-2016 , 07:18 PM

#869

Murdocks

enthusiast

Join Date: Dec 2013 Posts: 84

Hello, I would like to make a question about something that I dont understand.

I have input a BBvsBTN situation of a single raised pot on a turn of 8c7d5d 3c and it looks like BB here should check raise 33 if having the 3d with frequency of 0.931 but he should check raise 33 without the 3d with a frequency of only 0.469 which is half that.
I find it unexpected that BB is more inclined here to checkraise 33 that block some of opponents flushdraws instead of the ones that dont.
I think intuitively that it should be more GTO that we checkraise 33 that dont block the flushdraw because we dont want to check call vs the flush draw as much as we want to check raise.
Can you tell me why Im wrong and we prefer to chkcall vs a range that has more flush draws than chkcall vs a range that doesnt ?
Thanks

Quote

02-06-2016 , 02:37 AM

#870

SiQ

Carpal \'Tunnel

Join Date: Jan 2008 Posts: 7,248

you need to make sure you've run the sim to at least a decently low exploitability to put a lot of stock into some of these blocker plays.
with that said, it's going to depend on the ranges involved, but having 3d can make flushes less likely to fill in. it's also worth looking at which combos in our opponents range we block when we have 3d.

Quote

02-06-2016 , 10:14 AM

#871

Murdocks

enthusiast

Join Date: Dec 2013 Posts: 84

Quote:

Originally Posted by SiQ

I let it run till its 0.034 % * pot exploitable.
I believe this would be enough.

Its definitely worth looking but the only combos that we block when we have the 3d but we dont otherwise are some of his flush draws. Otherwise there isnt much of a difference. I like your argument of its less likely for him to make his flush so our hand has a higher equity vs his range probablybut I would still like to hear about what exactly makes pio give this result from the developers.

It's still a question of theory though. If you had to and it was mandatory to check/call one of your three combos of 33 here would it be one with diamonds or one without?

Quote

02-06-2016 , 07:56 PM

#872

punter11235

Carpal \'Tunnel

Join Date: Mar 2005 Posts: 8,210

Quote:

I find it unexpected that BB is more inclined here to checkraise 33 that block some of opponents flushdraws instead of the ones that dont.

The why question is in general very hard to answer as finding an exact balance for an equilibrium isn't a straightforward nor intuitive process.
If you want people to look into it you need to post the whole config so they can re-run it (copy to clipboard button and use pastebin or code tags here to paste).
Unfortunately I don't that much time for analyzing specific cases, if the case is interesting enough though it's likely that someone else may bite

Quote

02-06-2016 , 10:51 PM

#873

TheDefiniteArticle

Indecisive

Join Date: Feb 2012 Posts: 14,123

Is there generally a relationship between strategies which have similar exploitability? Can I assume that a 0.5% exploitable strat is relatively similar to a 0.25% exploitable strat? I ask because my computer isn't really good enough to process the latter at any reasonable pace (it seems that exploitability comes with diminishing returns) in wide range spots, so I've been scripting a bunch of BTN vs BB spots with 0.5% exploitability despite that being higher than recommended in your intro video.

Quote

02-07-2016 , 01:49 AM

#874

Murdocks

enthusiast

Join Date: Dec 2013 Posts: 84

Quote:

Originally Posted by punter11235

Ok, thanks for the answer.
Interesting that math comes to show that some things which not so long ago were considered basic theory knowledge are actually wrong

Quote

02-07-2016 , 05:54 AM

#875

SiQ

Carpal \'Tunnel

Join Date: Jan 2008 Posts: 7,248

Quote:

Originally Posted by TheDefiniteArticle

Is there generally a relationship between strategies which have similar exploitability?

Generally, .5 is close enough to give you fairly accurate frequencies and a very good look at the overall look of flop play - but some combos here and there might be off. As you go deeper into the tree you're likely to find more and more funky stuff.
So as a quick ballpark of general strategies I think .5 is good, but if you want to analyze on a combo-by-combo basis with confidence I would recommend running it more.

Quote

Page 35 of 299

First

15 25 30 31 32 33 34 35 36 37 38 39 40 45 55 85 135 235

Last

Post Reply Subscribe

...

Page 35 of 299

First

15 25 30 31 32 33 34 35 36 37 38 39 40 45 55 85 135 235

Last