Open Side Menu Go to the Top
Register
MonkerSolver MonkerSolver

05-06-2017 , 05:46 AM
Quote:
Originally Posted by Oskaris
Hi,
I’ve some question for you.

1) How much time do you think I approximatively need to run a complete 6Max tree with 1 000Go abbreviation?

2) How the buckets works exactly, compare to texture? For example, if I want to reduce the size of my tree, is it better to reduce the texture, or the bucket? Which one is the most important for precision?

3) Is there a way, when my run is finished, to save all my range (and only my Preflop range, not my tree) without doing it one by one? Because I can’t save the final files, it’s just to big. And save the range one by one is not practical… I will miss some and it not a option for me.

4) Is there a way to have a global point of view of the evolution of my range when I’m running? Because, if I understood right, the run is finished when the percentage are not evolving anymore. But if I’ve to check every range of every spot all the time, it’s not practical too…
Thanks forward for your answer.

Regards.
1) Will vary alot on multiple factors. A week, perhaps.

2) You should probably reduce buckets to ~15 before reducing texture. Try to keep the size of the streets somewhat even.

3) Right-click result table and choose "Save preflop tree". This will save the preflop ranges in monkerviewer format.

4) You could perhaps navigate to a flop situation, and choose "Board overview" to show the frequencies across all boards there. Only works in hold'em though.

Quote:
Originally Posted by Oskaris
Hi Sorry i cant edit my message

i Cant run more than 10Go ram tree (my computer have 32go)

Dont understand... i run it as administrator before i used PiosolverEdge Version and have no problemwith it but with your solver, iam out of memory when is starting, any idea where this problem come from?

thanks

If you have set the -Xmx value higher, then either the estimate is incorrect (unlikely to be that incorrect) or you have other RAM consuming programs open such as pio.
Quote:
Originally Posted by de0
would like if you could add the ability to paste ranges in the ranges window for each player when setting up trees that start postflop. would be nice if you would support pasting piosolver format as well. just want to say thank you for continuing to update this software daily, software is amazing and will only be better with time
Sure, will add.
Quote:
Originally Posted by IsaacAsimov
Using the full version, I still get inconsistent results regarding the syntax for this range:

'((*!RRR:8%,AA:8%,KK:6%,QQ:7%,JJ:6%,AK!RR:7%,Axxy! RR:13%,*!RR![6-]:8%,OORR20%!JJ+),$0G:50%,$1G:28%,(AQT9-, AKJ9-, AQJ9-):21%,(AKJ8-, AQ98-, AQJ8-):10%,(KQJ-,KQT-,KJT-)!RR:10%,(KQ9-,KJ9-)!RR:8%,(4556+,4456+,6654+):25%,(3556+,4457+,4476+ ,4463+,6643+,6653+):16%)$nt)*)!((J7,QJ),(98,T8,T 9,88,99,TT,76,J7,QJ)ass,kss),(T9,88,99,TT,76,J7, QJ):ss,jj)(98,T4,T8,T9,44,88,99,TT,76,J7,QJ),ass ,kq:ss)'

This shows up as 3.9% in Odds Oracle and 3% in Monker Solver. What are the restrictions for the full version here?
Probably differences in the hand rankings. To use PPTs hand rankings you have to:
1. Download the handranking file (I believe it's the 10-handed one) from PPT here: http://www.propokertools.com/oracle_help/range_of_hands
2. Rename it to %.txt and place it in the Omaha subfolder.
3. Restart MonkerSolver.
Quote:
Originally Posted by IsaacAsimov
I see now that 'Call' is used here instead of check, correct?
Yes, checks are labeled call, and will always be available unless you remove them. You should generally not remove the calls/folds. For example, in your image the button does not have the option to fold, which means he will open 100%. Is this what you intended?
MonkerSolver Quote
05-06-2017 , 06:27 AM
Quote:
Originally Posted by MonkerWare
Yes, checks are labeled call, and will always be available unless you remove them. You should generally not remove the calls/folds. For example, in your image the button does not have the option to fold, which means he will open 100%. Is this what you intended?
No, not at all. The screenshot was from before I figured out that 'Call' = 'Check'.
MonkerSolver Quote
05-06-2017 , 08:42 AM
any chance you will implement icm?
MonkerSolver Quote
05-06-2017 , 09:51 PM
Is there a way to change what 'auto' does in the tree building tab? For example right now it seems to add something like 3/4 pot bet and pot bet etc. What if I just want it to add pot bet and then specific bet sizings for later streets etc.
MonkerSolver Quote
05-07-2017 , 12:27 AM
I'm not sure about these RAM estimates for PLO trees. I have 16 GB RAM installed and I've set -Xmx15g. If I try to run a tree that is estimated to use 6 GB RAM, the solver get's stuck on 'starting...' and after a while my RAM usage in Task Manager goes up to 96%, which I presume means it's out of RAM.
MonkerSolver Quote
05-07-2017 , 05:47 AM
I tried again with setting -Xmx8g, since the estimate is at 6 GB, and this produces an 'Out of RAM' error.
MonkerSolver Quote
05-07-2017 , 06:13 PM
imo, if you are planning on using this extensively for plo, you should just get a new machine with 64/128GB so you don't continually encounter memory constraints
MonkerSolver Quote
05-07-2017 , 09:43 PM
Quote:
Originally Posted by Hoopster81
imo, if you are planning on using this extensively for plo, you should just get a new machine with 64/128GB so you don't continually encounter memory constraints
I don't disagree, I'm just saying that the estimate is way off. I'm still pretty new to the software, so would like to run some tests before investing in new hardware etc. I will try the software on a server with more RAM, just thought that I would perhaps report what might be a bug.
MonkerSolver Quote
05-08-2017 , 11:41 AM
Will 96gb RAM be enough for solving any kind of single raised HU-pot for PLO? About how much does one need for solving pre-flop trees?
MonkerSolver Quote
05-08-2017 , 03:45 PM
Quote:
Originally Posted by gtlol
any chance you will implement icm?
Yes, but no guarantee.

Quote:
Originally Posted by IsaacAsimov
Is there a way to change what 'auto' does in the tree building tab? For example right now it seems to add something like 3/4 pot bet and pot bet etc. What if I just want it to add pot bet and then specific bet sizings for later streets etc.
No, but you can achieve that by using the custom filters. Right-click->add to all->custom filter, then apply 4 filters, one for each street.

Quote:
Originally Posted by IsaacAsimov
I'm not sure about these RAM estimates for PLO trees. I have 16 GB RAM installed and I've set -Xmx15g. If I try to run a tree that is estimated to use 6 GB RAM, the solver get's stuck on 'starting...' and after a while my RAM usage in Task Manager goes up to 96%, which I presume means it's out of RAM.
If this is a preflop tree, the estimate does not take into account the creation of tables which requires 15g and takes a few hours (this only has to be done once).
If it's a postflop tree, and you have waited for a long time, then the tree is most likely not valid. There is no validation so it will simply crash.
Please send the tree to support@monkerware.com

Quote:
Originally Posted by TriggerHappyTray
Will 96gb RAM be enough for solving any kind of single raised HU-pot for PLO? About how much does one need for solving pre-flop trees?
It's always possible to build trees which don't fit in memory. Virtually any tree limited to one betsize will fit. In preflop trees, you choose the size of the abstraction. It is not obvious how big the size has to be to produce good results however.
MonkerSolver Quote
05-09-2017 , 10:22 AM
Quote:
It's always possible to build trees which don't fit in memory. Virtually any tree limited to one betsize will fit. In preflop trees, you choose the size of the abstraction. It is not obvious how big the size has to be to produce good results however.
If you select multiple bet size options, will monkersolver select the more game theory optimal to the three, or will it just fill the three with the multiple bet sizes?
MonkerSolver Quote
05-09-2017 , 01:41 PM
Okey, so I got two problems:

1) The new updates doesn't allow me to import my PLO-ranges. I have my ranges written down, combo for combo. This always worked with the older versions

https://gyazo.com/d68ca14767470a75a254a3afdfe44469

2) I ran a big flop tree over night. Now when I try to save/run the tree, it says "ERROR: Out of memory". My machine has 96gb RAM, and 500gb free storage space, shouldn't be a problem?

Note that this is with an old version (and thus there's no estimation for tree size available), since the newer updates doesn't allow me to import my preflop-ranges.
MonkerSolver Quote
05-09-2017 , 01:50 PM
Quote:
Originally Posted by TriggerHappyTray
Okey, so I got two problems:

1) The new updates doesn't allow me to import my PLO-ranges. I have my ranges written down, combo for combo. This always worked with the older versions
+1 on improving the syntax.

It´s very time consuming when monkey doesnt fully support PPT syntax.

Syntax like [88+], [97+], double negatives like *!(2345!xxyy) and so on.
MonkerSolver Quote
05-09-2017 , 03:32 PM
Quote:
Originally Posted by Maroel
If you select multiple bet size options, will monkersolver select the more game theory optimal to the three, or will it just fill the three with the multiple bet sizes?
The latter.

Quote:
Originally Posted by TriggerHappyTray
Okey, so I got two problems:

1) The new updates doesn't allow me to import my PLO-ranges. I have my ranges written down, combo for combo. This always worked with the older versions

https://gyazo.com/d68ca14767470a75a254a3afdfe44469

2) I ran a big flop tree over night. Now when I try to save/run the tree, it says "ERROR: Out of memory". My machine has 96gb RAM, and 500gb free storage space, shouldn't be a problem?

Note that this is with an old version (and thus there's no estimation for tree size available), since the newer updates doesn't allow me to import my preflop-ranges.
1) Will be fixed next update. In the meantime, placing the range in a textfile, and drag-and-dropping the file works.

2) It's possible there was some issue with this in previous versions. Please use the latest version and use the above method for ranges.

Quote:
Originally Posted by sooraw
Syntax like [88+], [97+]
I was not aware this is legal syntax. There is no mention of having two ranks within brackets in the ppt documentation (correct me if I'm wrong), and while [88+] works in their simulator, [88,99] for example does not, as one would expect if it was legal.

Quote:
Originally Posted by sooraw
, double negatives like *!(2345!xxyy)
Seems to work fine.

Quote:
Originally Posted by sooraw
and so on.
Please provide examples of expressions which are incorrectly parsed in the full version and it will be fixed. The free version has known issues.
MonkerSolver Quote
05-09-2017 , 09:48 PM
Quote:
Originally Posted by MonkerWare
T
1) Will be fixed next update. In the meantime, placing the range in a textfile, and drag-and-dropping the file works.

2) It's possible there was some issue with this in previous versions. Please use the latest version and use the above method for ranges.
With the latest version I receive the same error message.

The abstraction estimates to 75gb, I have 96gb RAM, and 500gb free storage. I've changed to -Xmx92g.

edit: The Task Manager also shows that Monker do actually appears to run out of RAM

Last edited by TriggerHappyTray; 05-09-2017 at 09:55 PM.
MonkerSolver Quote
05-10-2017 , 06:04 AM
i would be grateful if you did
MonkerSolver Quote
05-10-2017 , 01:22 PM
Quote:
Originally Posted by TriggerHappyTray
With the latest version I receive the same error message.

The abstraction estimates to 75gb, I have 96gb RAM, and 500gb free storage. I've changed to -Xmx92g.

edit: The Task Manager also shows that Monker do actually appears to run out of RAM
Just to clarify, which of these is it?
1) You can start the tree but can't save it.
2) You can start and save but not load the save.
3) You can't start it.
MonkerSolver Quote
05-11-2017 , 08:10 AM
Regarding the Linux version, I was able to extract the .dmg file using 7-zip in Ubuntu 16.04. After installing JRE 8, I am able to execute 'start2.sh' from the terminal, however to me it seems like this launches Monkerviewer and not Monkersolver, since it just opens the login window of Monkerviewer, showing the patch note changes to Monkerviewer.

You can see the login window that appears here: https://imgur.com/a/Ripyf

Here is a screenshot of all the files in the folder: https://imgur.com/a/ZPMcb

I've also tried running 'Monkersolver.jar' through JRE, but the result is the same. I can't compare to the native Mac OS version, because I don't own a Mac and Apple doesn't allow virtualization.
Not sure what to try next, maybe you can give me a hint.
MonkerSolver Quote
05-11-2017 , 09:17 AM
Quote:
Originally Posted by IsaacAsimov
Regarding the Linux version, I was able to extract the .dmg file using 7-zip in Ubuntu 16.04. After installing JRE 8, I am able to execute 'start2.sh' from the terminal, however to me it seems like this launches Monkerviewer and not Monkersolver, since it just opens the login window of Monkerviewer, showing the patch note changes to Monkerviewer.

You can see the login window that appears here: https://imgur.com/a/Ripyf

Here is a screenshot of all the files in the folder: https://imgur.com/a/ZPMcb

I've also tried running 'Monkersolver.jar' through JRE, but the result is the same. I can't compare to the native Mac OS version, because I don't own a Mac and Apple doesn't allow virtualization.
Not sure what to try next, maybe you can give me a hint.
Well, you can log in. This is the same exact window that opens in the Windows version.
MonkerSolver Quote
05-11-2017 , 11:26 AM
Quote:
Originally Posted by MonkerWare
Well, you can log in. This is the same exact window that opens in the Windows version.
I tried logging in, but I get a 'Log in failed.' error. I can log-in to my account fine from Windows. I think it's possible that the communication to your server is being blocked somehow. I'm not sure how to test that. Do you have any idea? Like a ping to the Monkerware servers or something?
MonkerSolver Quote
05-11-2017 , 11:53 AM
Quote:
Originally Posted by IsaacAsimov
I tried logging in, but I get a 'Log in failed.' error. I can log-in to my account fine from Windows. I think it's possible that the communication to your server is being blocked somehow. I'm not sure how to test that. Do you have any idea? Like a ping to the Monkerware servers or something?
If you have a firewall, you could try fiddling with those settings. Or, send an email to support@monkerware.com from your account email and I can give you more specific help.
MonkerSolver Quote
05-11-2017 , 11:59 AM
Quote:
Originally Posted by MonkerWare
Just to clarify, which of these is it?
1) You can start the tree but can't save it.
2) You can start and save but not load the save.
3) You can't start it.
Right now I can't start the tree. Status goes from "Starting..." to "ERROR: Out of memory".

I've run the same tree before, but then I couldn't save it, after letting it run over night. Same error message I think, "Starting..." to "ERROR: Out of memory".
MonkerSolver Quote
05-11-2017 , 01:10 PM
Quote:
Originally Posted by TriggerHappyTray
Right now I can't start the tree. Status goes from "Starting..." to "ERROR: Out of memory".

I've run the same tree before, but then I couldn't save it, after letting it run over night. Same error message I think, "ERROR: Out of memory".
edit
MonkerSolver Quote
05-12-2017 , 03:12 PM
Filtering has changed since the latest update, and no longer works correctly.

For example, if I have a PLO pre-flop tree and filter for:

Quote:
(*!RRR:8%6h,AA:8%6h,KK:6%6h,QQ:7%6h,JJ:6%6h,AK!RR: 7%6h,Axxy!RR:13%6h,*!RR![6-]:8%6h,OORR20%6h!JJ+),$0G:50%6h,$1G:28%6h,(AQT9-, AKJ9-, AQJ9-):21%6h,(AKJ8-, AQ98-, AQJ8-):10%6h,(KQJ-,KQT-,KJT-)!RR:10%6h,(KQ9-,KJ9-)!RR:8%6h,(4556+,4456+,6654+):25%6h,(3556+,4457+,4 476+,4463+,6643+,6653+):16%6h)$nt


It returns 61.1% of all hands, whereas the range above is only supposed to represent a 12% range.
MonkerSolver Quote
05-13-2017 , 06:30 AM
Hi

Is it possible to run Monkerware using the Mac Installer on Linux through command line?
MonkerSolver Quote

      
m