ICM HoldemResources Calculator [HRC] - Page 9 - Poker Software -- Two Plus Two Poker Forums

I read Mathematics of Poker, Intelligent Poker Player and Winning thought holdem

it would be great if HRC had command line interface
and it would be twice great if with option to make output into text file

Quote

01-26-2019 , 06:17 AM

#202

Vandalar

grinder

Join Date: Jun 2007 Posts: 412

Anyone know a source that has a free thorough manual/tutorial on how to properly use HRC in different formats and different situations and how to set it up properly etc, or are all of the good stuff in training sites? I've been trying to look a video here and video there but it's kind of erratic because it's all over the place.

Quote

01-26-2019 , 01:50 PM

#203

numberonedonk

veteran

Join Date: Oct 2014 Posts: 2,181

Is there no option to save hands?

Quote

01-27-2019 , 04:47 AM

#204

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

@Vandalar: There are some free-to-watch webinars I did on pokerstrategy, but they are slightly outdated by now:
https://www.pokerstrategy.com/video/39510/
https://www.pokerstrategy.com/video/33261/

I'll do a new tutorial series once the current beta is ready for general release.

@numberonedonk: You can upload calculations (Hand: Share Hand) or export them to text / html (Hand: Export Strategies). There is currently no mechanism to load a calculated hand back into the program at a later time though, but that's planned for the current beta cycle.

Quote

01-27-2019 , 05:45 AM

#205

hairyguy

newbie

Join Date: May 2017 Posts: 47

@plexiq can you or someone else help me to understand ICM: Malmuth-Harville Formula by Bill Chen?

I found https://poker.stackexchange.com/ques...a-by-bill-chen

It contains:
"For 3 players, let's set the chances of winning at 50%, 30%, and 20% for players A, B, and C respectively."

How can we get these numbers 0.5, 0.3, 0.2?

Quote

01-27-2019 , 07:32 AM

#206

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

The example you linked looks fine, it's just that sentence you quoted that is formulated a bit confusing.

"For 3 players, let's set the chances of winning at 50%, 30%, and 20% for players A, B, and C respectively."

->

"Let's look at an example with player A holding 50% of remaining chips, player B 30% and player C 20%."

You can also find a definition + example of ICM in here, page 8.

Quote

01-27-2019 , 07:58 AM

#207

hairyguy

newbie

Join Date: May 2017 Posts: 47

big thanx @plexiq!

Quote

01-27-2019 , 08:45 AM

#208

hairyguy

newbie

Join Date: May 2017 Posts: 47

Malmuth-Harville formula looks as the tiniest possible and definitely not precious
but you can't simulate with FGS all ways
for time reason you have to stop emulations and evaluate results with some simple like M-H formula evaluations. am I right?
I wanted to ask: simulating games do you use M-H formula in the end anyway?

Quote

01-27-2019 , 10:04 AM

#209

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

ICM M/H is used in the FGS calculations at maximum depth, that's correct.

Quote

01-27-2019 , 10:25 AM

#210

hairyguy

newbie

Join Date: May 2017 Posts: 47

from your diploma about an improved M/H:

my mind rejects to understand this
can't you please write it in any programming language?

Quote

01-27-2019 , 10:45 AM

#211

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

There's a reference implementation in Java included in the thesis, somewhere in the appendix.

Quote

01-27-2019 , 11:44 AM

#212

hairyguy

newbie

Join Date: May 2017 Posts: 47

you made me happy today
thank you
today was a good day

Quote

01-27-2019 , 12:07 PM

#213

hairyguy

newbie

Join Date: May 2017 Posts: 47

can you write one string of a Java code with good long variable names example of accessing double dimensional array returned by double[][] getHarvilleMatrixFast(double[] x) method?

it looks much more complex than I got a first impression
I don't understand what does means the dimensions and the values of this matrix

Quote

01-27-2019 , 12:15 PM

#214

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

Pretty sure this is explained somewhere in the thesis, but p[i,j] is the probability of player i finishing in place j+1. So p[i,0] is the probability of player i finishing in 1st place, etc.

The variable naming in the Java version mostly corresponds to the algorithm you posted above, the variables are explained there and/or introduced earlier in the section.

Quote

01-27-2019 , 12:43 PM

#215

hairyguy

newbie

Join Date: May 2017 Posts: 47

thank you again

0: 3.862873613335336
1: 1.2232735265676509
2: 0.6742236598050455
3: 3.0705382554412477
4: 8.822266833976414

0 1 2 3 4
0 0.218 0.289 0.282 0.164 0.045
1 0.069 0.105 0.17 0.356 0.298
2 0.038 0.06 0.1 0.222 0.579
3 0.173 0.242 0.3 0.211 0.071
4 0.499 0.302 0.146 0.044 0.006

a wonderful formula
great

is Weitzman's matrix more precise?

Quote

01-27-2019 , 12:49 PM

#216

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

No offense, but I just linked a paper that compares the accuracy of these models. Maybe check the summary?

Quote

01-27-2019 , 01:09 PM

#217

hairyguy

newbie

Join Date: May 2017 Posts: 47

hehe ok, I'll check
I don't like big papers
no offence

there was a time when I wrote a Java tutorial for a total newb
it was written in a chat of 2 guys format
it contained with all code samples and screenshot around 30 pages

Quote

01-27-2019 , 01:20 PM

#218

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

It's weekend and this is fairly far outside of regular product support, I'm leaning towards shorter replies in this case :-)

Quote

01-30-2019 , 08:33 AM

#219

hairyguy

newbie

Join Date: May 2017 Posts: 47

strange result detected

Advanced hand options:

result:

Quote

01-30-2019 , 08:40 AM

#220

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

You have a large number of ~0EV hands in that range and it seems to be heavily using mixed strategies (right click hand grid: Show Strategy). This spot will require additional iterations, as strategies won't converge within he default number of iterations.

Use the "Run Nash Calculation" button in the toolbar to improve the accuracy. Check the Exploit% numbers in the outline, they should be close to zero.

Quote

01-30-2019 , 09:16 AM

#221

numberonedonk

veteran

Join Date: Oct 2014 Posts: 2,181

lol at this guy...

Quote

01-30-2019 , 09:48 AM

#222

hairyguy

newbie

Join Date: May 2017 Posts: 47

yes, more iterations fixed it
I set no 4-bets and allowed all-ins but there are 4-bets in the result tree and all-ins are called by 4-bets
also I wanted to allow all-in instead of 3-bets so the tree contain (somehow it is possible) and 3-bets and all-ins after open raise
also I would like somehow to see a calculation that includes simultaneously and open raise and open all-ins

on my laptop the tree pane on the shared hand page has small height and there is a lot of free space on the page to the right of 3 previous panes: Overview, Stacks, Hands

Quote

01-30-2019 , 09:49 AM

#223

hairyguy

newbie

Join Date: May 2017 Posts: 47

Quote:

Originally Posted by numberonedonk

lol at this guy...

you lie
also I can't say that your avatar is lol but maybe it pretends for it

Quote

01-30-2019 , 09:53 AM

#224

plexiq

veteran

Join Date: Apr 2007 Posts: 2,554

If you leave a line empty then it defaults to all-in only, as explained at the top of the dialog. So you will still see 4bets, even if that line is empty.

You can specify multiple bet sizes for a certain action, e.g. setting the opening size to "2bb, all-in" will allow both 2bbs and all-ins in the calculation.

Most of this stuff is explained in detail in the Advanced Hand section of Help: Help Contents btw.