Hi Plexiq,

I'm having the following error message when trying to edit a range with the latest HRC Beta version:
"Could not initialize class javax.swing.UIManager"

When I right-clic --> edit range, nothing happens.
When I double_clic, I have the above error message.

Everything worked fine with previous Beta version, including range bookmarks.
On the same PC, I can still edit ranges with the Official version.

I'm on windows 10 with java up-to-date.
I just installed the latest HRC Beta version, so it has never worked for me with this version.

Do you have any clue what the problem could be ?
If not, is there a link to previous Beta versions ?

Thanks !

I just reinstalled the "net.holdemresources.calculator.beta.installer.win 32.x86_64.exe" I had from the 08/20/2021, and everything works fine now. And HRC Beta tells me I am up-to-date...

So I don't understand what was the problem but it seems fixed.

It's not up-to-date, I don't have the script betting and other new features, but I can edit ranges...

Try downloading the latest installer from the beta blog post, lmk if you still have troubles.

I just did it, same issue as in first post:
"An error has occurred. See error log for more details.
Could not initialize class javax.swing.UIManager"

Will look into it tomorrow, I think we'll do a small update to fix just that.

Try updating, this should be fixed now.

Yes, problem fixed, I can edit the ranges now.
Thanks for your reactivity !

Plexiq,

I think you already know that, but when closing a large tree the RAM doesn't free totally.
To reopen a large tree I need to close and relaunch HRC Beta.
I put Xmx20G in hrc.ini in case it's linked.
Maybe you want to fix that in a future update.

Try enabling "Window: Preferences: Show Heap Status", restart HRC and you should see the memory display with a trash can button.

If you notice the same problem again then please close all hands and click the trash can icon, memory usage should go back below 1G.

Let us know if that's not the case.

I enabled the "Show Heap Status", restarted HRC, opened a large tree: 15 909 M of 20480 M
I closed the large tree: 15909 M of 20480 M
I clicked the trash can: 15391 M of 20480 M

I have 2 x 8 Go in dual channel + 1 x 8 Go RAM in case it matters.

If this happens again, please try opening a small hand, selecting a range, then closing the small hand and using the trash can. Does it clear then?

System is 24G total, right?

Open small tree: 914 M of 2624 M
Select a range
Close small tree: 943 M of 2624 M
Click trash can: 840 M of 2624 M

Yes 24 Go total, temporary setup.

I meant if a large tree is stuck in memory after closing it: Then try opening a small tree & selecting a range, closing and trash can.

I'll look into it after the weekend. Any steps to reproduce the issue are appreciated. Does this always happens for you?

As far as I know the trees have always been stuck in memory after closing them for me.
The "Show Heap Status" confirms this.
However the trash can seems to work fine now, it clears memory < 1 Go, so I cannot try your suggestion right now.
I'll let you know if the trash can doesn't work again.

It's normal for the trees to remain in memory after closing, but they should be cleared with the next garbage collection. This either happens automatically if HRC needs the memory or if you click the trash can icon.

If memory of closed hands is not cleared by the garbage collection then that's definitely not normal. If you can reproduce the steps to get hands stuck in memory that would be a great help.

hey i was wondering why hrc doesn't calculate the ev of folding in icm situations as it should be higher than 0

The $EV when folding is of course calculated behind the scenes.

For display purposes all EVs are displayed in comparison to folding though, so folding will be 0EV by definition here.

It happened again, I did the following:
- Open a large tree
- Run a solve
- Save the tree
- Close the tree
- Open a second large tree --> Memory error

What was the memory usage after closing all hands and clicking the trash can?

I believe I found the issue, an update will be online within a few minutes.

Hi, I am currently working with the beta version of HRC with the monte carlo/postflop

To what extent is HRC taking into account certain future game aspects (e.g. likeliness to take down a tournament if we double up, ICM pressure we can now exert on others).

My gut tells me that mid stacks should often be calling off wider than HRC outputs on FTs, because if they double up they now have a great chance of taking down a top-heavy tournament - or is this already being considered?

Really appreciate any insight

It depends on the equity model:

ICM and MTT ICM both use the vanilla "Malmuth Harville" ICM algorithm to estimate the value of the stacks after you e.g. double up. While vanilla ICM does tend to slightly undervalue large stacks, you should be careful not to overdo it with manual compensations.

FGS on the other hand actually calculates the tournament tree of the following hands up to the selected depth. This will provide much more accurate equity estimates, but to use this model you need stacks that are suitable for push-or-fold play.

Maybe run some deep FGS calculations on short-stacked hands and compare the results to the regular ICM output to get a sense for the adjustments.

Thank you for your reply

If I tick "Skip SB", is the small blind still posted to the pot (but treated as if he always folds)? If not, is this possible to achieve?