I just tried to save some preflop callingrange by using h@d as syntax. It looks fine for most of the hands like
6cAd7dTs 0.9923
7cAd7dTs 0.9972
…

but then I get some numbers >=1 in between like
8cAd7dTs 6

Do you guys have the same bug and is there a workaround?



@isaac
Thank you!

do you have a rough estimation how much slower this would be? I am hesitant to invest ~4.5k+ and I am not able to put the pieces together myself. So, something ready and cheaper would be nice.
If it was a bit slower that would be fine with me.

A rough guess would be that it would be at least 2-3 times slower. It's hard to say, because I have no reference for the precise hardware you are trying to use, but I have spoken to people who've attempted to use ddr3 hardware, and it was either unusable or significantly slower. I would only really use a machine like that if I wanted to just view solutions, but then you are probably just better off running a spot instance, such as is described here:

https://solveoptimized.com/2018/02/2...r-mokersolver/

I think if you don't want to front a lot of cash, your best bet is to rent a server with Contabo. It also depends a bit on your use case. You can probably do a lot already with a Ryzen 2700 and 64GB RAM (1-1.5k USD), unless your goal is to solve 5+ handed preflop play.

Can monkersolver handle straddles and antes starting from preflop?

Yes.

Been trying to contact the devs... anyone know if they actually respond to email? no response in 8 days? other contact methods?

Your best bet is to use an email address that is registered to a license. Other than that I don't know of a better way to contact the devs.

why is "abstraction" in settings capped to 6,875 gb to me? I have a 128 gb rig.

Locate 'MonkerSolver.l4j.ini' in your MonkerSolver folder. You will need to manually increase the Xmx value to accomdate for your needs. Don't forget to leave RAM for system resources.

-Xms3g
-Xmx7g

what do you suggest i put there got a 64gb comp and a 128gb comp that im running sims on. sorry for being a idiot

nvm, tried myself.

put 21/49 on the figs, now its running @ 44gb on abstraction, leaves me with ~13gb free on the 64gb rig, seems good enough.

I would recommend adding the following two lines to your config file:

-XX:+UseStringDeduplication
-XX:+UseG1GC

This will improve RAM efficiency, and make it such that the estimate you see from within the solver is the same as what you set it to in the config file. I would leave the Xms value as is. It simply specifies how much RAM the solver uses upon startup. Setting it to 21 is just wasting system resources and can lead to instability in some cases. You don't want to risk crashing the solver unnecessary ( you can lose unsaved data that way).

Further more, the config file is applied to each instance individually, so I would recommend adjusting it according to your requirements for each individual solve and not necessarily just using one global setting and leaving it like that.

okay, ill try this when this solve is done, ty.

Want to trade sims?

What I have: 100,50,25bb PLO preflop solved from 15-24 I/N respectively, rake at 250mchips made by a high-profile, mid/high stakes player and coach. I'd prefer not to post his name here, but we can talk. My sims are of good quality.

Why I want to trade? 5%, max 250mchips is way too low for my low stakes games and I need something that more closely resembles my stake, so I'm looking (optimally) for something at 5%, max 4000mchips, but anything significantly higher than 250 would be of interest. Any stack depth is of interest to me also. Thanks!

When I build a tree and look under settings and abstraction i'm getting that I only have 6.2gb for abstraction and I cannot make a tree. The computer i am using has much more ram than that. Do you know how i make more ram available for this?

https://gyazo.com/fc1eccf96b172460b17f04ec5cea465a

I have a machine with 256gb ram.

The following is written in the monkersolver.l4j file:

-Xms3g
-Xmx240g
-XX:+UseStringDeduplication
-XX:+UseG1GC

In settings it says: Abstraction (119.6GB/240GB).

However, when I solve, in task manager, it indicates that java platform is using 206gb ram. I tried to solve some larger trees with approx Abstraction (200/240GB) and it crashed.

Any idea what's causing this inaccurate ram estimate?

It's common for the solver to underestimate, esp. for pre-flop sims. There's nothing you can do except reduce branches or increase abstraction. I doubt anyone but the developer knows why the estimates are inaccurate.

So basically the best method to run these sims would be to just run the solver several times, end task when it crashes, and sort of trial and error out the best abstraction settings that utilize the most memory without crashing?

It seems problematic to estimate because the amount of memory being used was 164gb a couple hours ago and now it's down to 138gb. Once the iterations get going though, I'm guessing that's sort of a guarantee that it won't require more memory than available and won't crash?

You need to adjust the settings according to the sim you are trying to run. I would recommend using a tolerance of 4g. So if you receive an estimate of 119.6, set Xmx to 124g. If you observer and see how much RAM the solver is using, you will see that it's quite a bit less than 206g.

A mistake I see people make sometimes is that they will over-assign RAM in their machine. Even if you are not using an instance, assigning RAM to it may cause some or all of you Monkersolver instances to crash. So for example, if you are trying to solve a game tree with an estimate of 200g, first close all instances of Monkersolver, change your configuration file to have Xmx be 204g, and then open a single instance of Monker, and start the solve. Also make sure that you do not have any other resources running in the background that could potentially use up a lot of RAM (like a browser with a lot of tabs for example).

how did monker make the game small enough to solve?

http://monkerware.com/compare.html

If I want to run 2 instances of monker solver, of let's say a 90GB ram sim and a 60GB ram sam, how do I configure that monkersolver.l4j file?

Do I need to pause one sim and configure the file or can I configure the file while one sim is running and just add the other sim?

is the reason why monkerviewer change the ranges from each time you open it because of it randomising mixed strat options into one of the weights (100/0) ? Like it can say open 100% A3o one day, and fold the other day.

Im not sure why viewer only show 100/0 weights to begin with, its pretty wierd/bad imo.

The configuration of monkersolver.l4j is valid prior to each initiation of Monkersolver. So if you want to open a second instance of Monkersolver with a different RAM configuration, all you need to do is change the Xmx setting in the monkersolver.l4j file to whatever you want it to be (and save it) before launching the instance of Monkersolver. Once you open Monkersolver with specific monkersolver.l4j settings, those will be locked in for that specific instance, and not affected by changes you make to the monkersolver.l4j file.

Hi,

Would anyone be able to help me with the following please.

Looking at the HU 25bb tree in monkersolver, I see that when SB min opens at depths between 22-25bb's, the BB 3bet response is either all in, or Pot(6bb's).

Why Pot?

I thought this would be too big of a sizing at these stacks.
The standard sizing's i see usually being used is 4.5 to 5bb's.

Is this really the sizing the solver vastly preferred over all others, and therefore the other sizing's were pruned away from the tree?

Also, from 21 down to 15bb's, the BB 3bet sizing to min open (besides all in), is 8sb's.
Does this mean 4bb's? so t80 at 10/20 level.

I'm really interested in buying some ranges, but don't know if it's a waste of money, if these sizing's are just sort of wrong.
Or am i totally getting it wrong

Cheers in advance for any help with this.

Press enter to change it to the mixed view if you want.

I can´t read the nGTO Omaha ranges in MonkerViewer.

I saved in results tab to view in Monker and dragged it into the MonkerViewer.exe folder. A tab "HU" (HU results) appear, but no results. Please advice