HoldemResources Calculator [HRC]

Nah, there's no way to do that in HRC. (Or in general really, as Nash Equilibria are only guaranteed to exist when you allow mixing.)

How come the sim can end up choosing an inferior strategy when given more options to choose from? Help me understand this, please.

Example: FT 4-handed. Payouts 14.75% / 11,5% / 9% / 7%
CO 20bb
BTN 8bb
SB 4bb
BB 10bb

I run two sims (Monte Carlo, Malmuth, CI 1.00, reset regret).

Sim 1: Everyone plays allin/fold.
Sim 2: Only the CO gets to play minraise/allin/fold, while everyone else is still allin/fold.

Sim 1: CO's EQDiff% is +0.181
Sim 2: CO's EQDiff% is +0.176

(And just to double check, Sim 3 where I edit and lock the minraise range back to 0% gives +0.181.)

I've previously seen the difference come out three times bigger than this, but this is the spot I was working on now. The difference isn't just noise - no matter how many times I run it, CO's equity is lower if it's given more options to choose from.

Why doesn't CO just play allin/fold if it's the better strategy here, regardless of the existence of the 2bb option?

In ICM scenarios (or more generally, non-zero-sum games ), adding additional actions for a player can actually decrease that player's EV in equilibrium.

Tombos wrote a good article about this:
https://blog.gtowizard.com/the-limitatio...

A very similar topic was recently discussed in Poker Theory:
https://forumserver.twoplustwo.com/15/po...

Interesting article and very swift reply, thank you!

Right, so even just the threat of a minraise changes the equilibrium strategy and increases leakage. I understand that part now.

What's still unclear to me is why can't CO remove that threat herself by simply never minraising? As per Sim 3, where we edit the minraise range to 0% and the EQDiff returns higher to where it was. How come that line has to be forced, rather than chosen by CO as optimal?

Or from the other perspective, does this also mean that whenever we custom edit a certain range down to zero, it fundamentally affects equities beyond simply "this player chooses not to do X" into the territory of "line X isn't available at all", therefore reducing aggressiveness and leakage in the equilibrium responses?

Once you introduce the minraises there, a pure push/fold is no longer a Nash Equilibrium. You can verify that by locking minraises to 0%, calculating, and then checking if any hand has their highest EV as minraise (select grid display mode: EV vs best other action). If even a single hand has the highest EV there then 0% minraising can't be the final point of a Nash Equilibrium calculation for that spot.

Remember, a Nash Equilibrium only guarantees that no player can unilaterally change the strategy to increase their EV if all other players remain static. Clearly above spot would violate that condition if any hand can switch to minraising to improve the EV against the current strategies. The re-adjustments that follow introducing using a minraise may eventually lead to a worse EV for the minraising player in the equilibrium, but that's a different issue entirely. Nash Equilibria make no guarantees wrt what happens if you allow re-adjustments.

Got it, I see where my thinking was flawed. Great replies, thanks again!

Feature request: When generating MTT stacks for other tables in the multi-table ICM model, I'd love to have an option for setting the average stack size in BBs instead of chips. When I save that information from my client along with the hand history, it's in BBs, which means now I need to do math to convert for every sim, rather than just enter "X players left, Y bb average".

It's a minor issue, but becomes annoying fast when you have 50 hands to go through.

I don't want a separate input field there, but we can probably support something like entering "10bb" which is then automatically converted to the chip amount.

Oh, I was assuming a toggle option somewhere, but that's an even cleaner way to implement it. Would be happy to see the update some day!

Feature request: When entering stacks and blinds, it would be great to have a "Player sitting out" checkbox, to avoid having to zero out BB's range to 0 in all the various or relevant subtrees every time.

I'll look into adding a "sitting out" option. For now you could use scripting to have certain players auto-fold in every node, as a work around.

It's possible to use HRC for FL, but not via the standard UI tree configuration. You need to use a scripted tree config.

Check the pinned messages in the #scriping channel on our discord, there are a few FL scripts there that you can customize:

https://www.holdemresources.net/discord

Discord

Join the HoldemResources Discord Server!

Feature request for Mac version: More keyboard shortcuts, please. Some exist currently, but a lot are missing:

* Cmd-W to close out the current hand tab.
* Any kind of shortcut for "Don't Save" when closing the tab.
* Esc to cancel out of Nash Calculation and Hand Setup popups (maybe others as well).
* Enter to OK the Nash Calculation popup after Option-R.
* Cmd-V to paste the hand history in Hand Setup without clicking the little button.
* Cmd-N or Cmd-Enter or something to hit Next/Finish in Hand Setup.
* Something better than "Cmd-W M" to start a new Monte Carlo hand.
* Something much better than "Cmd-H B" to show Bubble Factors, since it doesn't even work - Cmd-H is already an overriding system shortcut for hiding the active window.

None of these are too bothersome when doing just a few hands, but if there's a thousand sims to go through then having to use the mouse starts to slow you down quite a bit. Otherwise the software is amazing.

Hm, a lot of these do already work on Windows/Linux, I'm not sure why OSX is behaving differently for you. (e.g. closing dialogs with ESC or pressing Enter to start the Nash Calc) I think re missing hotkeys (or bad defaults), we could include a configuration option where you can customize the settings. But this wouldn't affect standard stuff like closing a dialog via ESC though, that should just work out of the box.

I've to note that we do little testing specifically on the OSX version. The main development is done on Linux and we test with Windows since that's the main OS for users. But I'll check up on it.

I'm having trouble understanding EQDiff% again.

The way I've assumed this works is that for the given situation, Nash equilibrium means finding the highest EQDiff%, which equals the highest EV range. But I just tried editing my open shoving range and got higher EQDiff, so now I don't know what any of this actually means.

Example spot 5-handed FT, only allin/fold possible.

HJ: 19.4bb
CO: 10.6bb
BTN: 12.6bb
SB: 10.7bb
BB: 8.7bb

Remaining prize pool: 21.63%, 17.73%, 14.54%, 11.92%, 9.78%.

Looking at HJ, Nash calculation (no FGS) gives a shoving range of 50.3% and EQDiff% +0.124.

As you'd expect, EQDiff goes down if I edit HJ's range wider.

It also goes down if I make HJ a super nit shoving only 10%.

But if I try any range between 36-43% for example, EQDiff is higher than at Nash, up to +0.140 or even more. Which is quite a significant improvement.

So what's happening here? Why aren't we shoving the tighter range as Nash if it's higher EQDiff%? What even is this value?

I think the misunderstandin here is about Nash Equilibria, rather than EQDiff%.

This isn't generally true:

A Nash Equilibrium is a state where no single player can unilaterally change their strategy to increase their EV if other players remain static. That's not equivalent to what you posted (unless we are talking 2-player zero sum), but especially not in ICM spots.

The important part here: HJ can't increase their own EV by deviating if other players remain static! It's perfectly possible that HJ might be able to increase their EV if we allow others re-adjust though.

In high bubble factor spots, it's often better to play a bit tighter if that means that opponents will also call tighter (chance of all-in confrontations goes way down; which means less EV leakage to uninvolved player). BUT: That new state is not a Nash Equilibrium - against those tighter calling ranges, HJ would now be able to unilaterally go much wider - assuming others don't re-adjust.

That makes sense. Thank you for another good reply.

I’m assuming HRC won’t allow for us to create an unbalanced sizing strategy, but I’m trying to find a way to do it with nodelocks. I’m trying to nodelock different preflop sizes to create more EV disparity between hands. If I run out a 3 size preflop strategy, let’s say 2bb, 3bb, 4bb and then nodelock every opponent response - if I were to then nodelock one hand, let’s say QTs for one of our opening sizes and rerun it out, would the EV change? Whenever I run multiple sizes, hands all end up gravitating to close to the same EV regardless of size, even though the strategy will almost pure 2bb raise some and almost pure 4bb some others. Trying to better understand how it is functioning, thanks.

Hm, I'm not entirely sure what you are trying to achieve. You could use "frequency locking" to force the use of multiple sizings without explicitly locking any combos.

This could be what I’m looking for. How does it work? What I’m trying to achieve is to find which size is the highest EV for each combo in a vacuum. I’m trying to create that vacuum. So is A2o highest EV as a 3x on the button but KQs is highest as a Minraise? I’m assuming a lot of the necessity of HRC mixing is to not be face up, as HRC calculates as knowing its opponent’s strategy. Is this possible?

I'm not sure what you mean by "in a vacuum"? The EV of any single hand depends on your opponent's strategies, and these do in turn depend on your range.

If you mean "in a vacuum" to mean against opponents playing GTO, then that's what HRC is displaying by default (assuming no node locks etc).

If you mean "in a vacuum" to mean just analyzing a single hand, then i don't think that's useful at all. That would mean opponents play against a range containing just a single hand, ie playing with the knowledge of your exact hand class.

I mean against opponents that are playing a static strategy. Their strategy does not depend on our range, only on our size. They will respond to a 3.5x open with X% of calls and 3bets regardless of what our range is, because they do not know our range. Same for facing a minraise. I am trying to find a way to max exploit this tendency, is this possible? After node locks, many of these are still mixed frequencies, so perhaps the EVs regardless of size are just very close. Or does post flop readjust for pre flop imbalances? The frequency lock is a great tool. I am setting a threshold for each action and having good results, thank you.

I might be misunderstanding you. That's what you see as standard EVs after calculating. When you run the calculation, the result will be an approximate Nash Equilibrium and the displayed EVs are against opponents playing a static strategy (their Nash Equilibrium strategy).

That's very very common, yes. Especially if sizes are relatively close.

Hi, is it possible to study Mystey KO tournaments with HRC?