Hi I was playing around with EV-calculations and CREV came with a strange result.

EV of CO if he calls: +0.88




But if I calculate it by hand then I get this:

21,10% of the time the CO will win a $7.13 pot.
78.9% of the time the CO will lose $2.94 if he calls the all-in.

So the EV formular should be:

(0,211*7,13)+(0,789 * (-2,94) )
= 1,50443 + (-2,31966)
= -0,81523

In this case the CO would have a negative EV instead of the +0.88 EV.

Or does CREV include also results from preflop, flop and turn into the river all-in calculation?

This would be the case if Cutoff called with all of his range. However, he actually only calls with the strongest 42.1% of his range, while folding his weaker hands.

In order to carry out this calculation we will need to know what Cutoff's equity is when he calls. Unfortunately I can't tell this information from the screenshot, however, you should be able to get it if you mouse over the "Call" action. The equity is in one of the top lines in the popup. For the moment though, let's just call this equity X.

The EV for calling now becomes:
EVcall = (X*7,13)+((1-X)*(-2,94))

The EV for folding is 0 by definition:
Evfold = 0.

And if we enter the frequencies, the total EV will be:
EV = 42.1%*EVcall + 57.9%*EVfold = 42.1%*EVcall = (X*7,13+(1-X)*-2,94)*42.1%

Unless I made a typo, the equity for calling should turn out to be about 50%.

Is it possible to disable the "storing undo files" function?

It's not possible to disable the undo function at the moment. The storage of undo files should not take a very long time when using an internal hard disk (at least not on any of my systems). It can take a few seconds though when using an external drive though, given that they are slower. So if possible, please use your internal hard disk. That being said, I can look into an option to disable the undo system for future releases.

Thank you for the feedback,

Scylla

Thanks scylla, had another quick question in regards to using max exploit tool

So using max exploit tool is similar to using nash solver tool in that we HAVE TO right click lock/node lock villain's ranges (given our assumptions on his imbalances)?

Yes, both tools respect locked items in the tree.

Hi I have a question about analyzing turn cards:

This is my tree:




Now I want to know how to play the turn optimally. Do I just need to enter a turn card and then click on "calculate" again. Or do I need to run the solver again over this ying-yang symobl?

Btw. after the button flop 4bet (Raise 0.78) their appears this frequencie: 2E-5%?
What does it mean?

You can just enter the turn card and press F7 to see the results for that turn.


2E-1%=0.2%
2E-2%=0.02%
2E-3%=0.002%
2E-4%=0.0002%
2E-5%=0.00002%

So basically it's a very tiny number that's just not yet 0. Given that 0.00002% would take up too much space in the interface, it is instead shown as 2E-5.

I can't find where to download the 32 bit version, anybody have a link? Thank you.

Here is a link to a 32bit version for v333:
https://s3-us-west-2.amazonaws.com/c...33_32bit__.msi

Please do observe though that a 64bit system is highly recommended here.
The problem with 32bit systems is that programs will crash once the memory use exceeds roughly 1.5GB.
This is something beyond our control, given that it's a property of Windows.

So, if possible, please upgrade your system to 64bit.

Cheers,

Scylla

Hey, when I set up an equilibrium solution I find the "run 1 sample" button very useful as it helps me get a feel for how the solution plays out with all sorts of hands on all sorts of runouts. It helps me get a feel for how to play spots way faster than just looking through the tree.

What I would like is a way to say ok show me a sample run where the action goes bet/raise/call or check/bet/bet etc. so I can study specific lines and parts of the game tree.

For that, set a checkpoint at the decision where you want the sample.
To set a checkpoint, press F10 and click on the relevant decision.

Hey, I got this message upon startup today.



I'm assuming its some sort of mistake. Haven't used the program in a while.

Hi,

This is because you requested a new key for license 1 in December 2016, which caused your previous license to be deactivated. I have now reactivated the license for this computer. Can you try again please?

Scylla

Could you add support for progressive-KO in tournament setup?

That would be VERY hard if not impossible to do I would have thought. There is no way to estimate how many people you will knock out, and how many each of them have knocked out before YOU knock them out (so you don't know what bounty you will get).

I can look into it, however, one of the main challenges here is that currently there exists no generally agreed upon method for performing these calculations. This makes any model that I could create open to interpretation. One of the great things about tournament calcs is that we have a general agreed upon method (ICM conversions) for performing these calculations. And even though ICM conversions are not a perfect model, they turn these subjective calculations into an objective question: "what are the outcomes of the EV calculations if we were to apply ICM conversions?". No similar method exists for bounties. So, I can look into it, but any model that I could create will always be subjective.

It is enough if you implement just one of 3 known methods:
- flat compatible with Icmizer
- "Medium" ( kolemoen's suggested)
- Proportional (remaining bounties are distributed proportional to stacks)

http://forumserver.twoplustwo.com/36...ities-1622208/

Seems so that you really care about the program and that you are really motivated to deliver a solid running software. That is the right way !


That would be awesome.


btw. I have also a question to the solver. Is it possible to save the calculated results?

Lets say I solve today a tree and tomorrow I want to take another look in the results but I dont want to calculate everything again.

So is it possible to save the calculated tree?

I think this is one of the first posts in this 231 page thread? If so, then it has to be roughly 9 years old at this point. When going through posts, I would recommend skipping to the end, which will have more recent information.


Yes, for that just save it with Ctrl+S or by going to "File->Save as...".

I can look into it, however, I can not say for certain to which extent this will apply to progressive-KO tournaments.

If I enter a turn card then the flop strategy is also changing. But of course I cant know what turn card will come when I play.

So at first I let the solver calculate the flop strategy. Then I know how I should play the flop and in the next step I enter a turn card and calculate everything with F7 again. Now I can see that I have got also a other flop strategy.

Now comes the part where I´m not sure if I use CardrunnersEV correct:
Is it ok if I use the flop strategy from the first calculation?

I mean I cant know what turn card will come when I play so I guess its correct if I use the flop strategy from the first calculation. Is that right?

btw. is it possible to see what size I should choose with what hand?
In a PIO video I have saw thatit is possible to see the different bet sizes for different hands is there also such a function in CREV?

If you want the flop strategy, then only enter the flop (no turn) and press F7 (or "Calculate") to compute. If you want the strategy for a certain turn, then enter that turn and compute. To return to the flop solution, just clear the turn again and press F7 to compute.

Our engine is perfectly capable of handling trees of any size and with any number of bets. However, our tree building wizard is currently restricted to single bets. This is because we want to keep the wizard basic, giving us time to focus our efforts on other areas like our analysis systems and memory compression (For more on our analysis systems, please watch the second video here: http://www.cardrunnersev.com/download.html). That being said, if you manually enter any tree whatsoever (you can probably save yourself a lot of work by only doing the flop and using the wizard to simulate play with single bets from the turn on), then I think you'll find that it works perfectly well.