As I've said it'd have a negative effect in other places that I do not want. So there's basically no possibility to select a draw that is drawing to the bottom end?

Hey,
Maybe we test that on different versions ? I use latest public - 2.9.2.

there is nothing to send, just:

a) create or use any existing multicondition (picture)
b) Create script checkpoint on that CONDITION (NOT ACTION). It is impossible to do so.

Checked again and there was some kind of bug. I dont know what exactly kind of bug, but it didn't allow me to add checkpoint again, hwoever after addiding another checkpoint it allowed me to do. So everything is fine there, however there is possibility of some kind of minor bug there.

EDIT:

Ok, bug still exists. It can't put checpoint on dynamic multicondition, however it can do that on static one.

Any axplanation or just bug?

If the board is known then there's plenty of ways of doing this.
If the board is unknown and you're looking at 2 card gutshots then I suppose you could just specify that the hand is at most a highcard with a kicker above 4th card (which basically means below the flop).

I've tried both kinds of multiconditions and they work fine over here. Once again, please send a savefile. The fact that it seems trivial doesn't matter. Maybe I'm looking postflop and you're looking preflop. Or some other difference that leads to me not being able to replicate the issue.

sent

Suggestion for postflop range visualizer:

In the graph tool


It would be nice, that you would add right click button, on checkbox, that would behave in that way:

a) Click on CHECKED checkbox, when all conditions are graphed: unchecks all others conditions and just this condition is graphed
b) Click on CHECKED checkbox, when not all conditions are graphed (maybe just this one): turns all conditions to be checked and graphed
c) Click on UNCHECKED checkbox, this conditions becomes checked now and others unchecked, so only this condition is displayed now.

Also it could remember what values for X and Y axis I've used last time (NR, Equity, EV, Tie)

Ok, I'll look into it.
Thank you for the feedback.

somehow the new beta is pretty instable and disapears very often.

plus, this max exploitiv calculation seems to be wrong:
BB restealing 48%, if thats the case SB should be opening 26% ( and not 100%) and call off 4% in this case.

http://www.file-upload.net/download-...EX-SB.stx.html


that is what nash calculated:
http://gyazo.com/c088b41535be0d27a605e5337589afcd

As far as I know the beta is stable. Your issue is probably the result of some external program interfering with CardRunnersEV; most likely your virus scanner.


SB is opening "all hands" in this case because open-raising is the only option that it has to choose from. If you want to give SB the alternative of folding then you need to add a "fold all hands" action as well.

The software is simply doing what you told it to do, namely assign all hands to the optimal action. In the case of SB's first decision, there's only one action to choose from.

Other than that, the solution that the software arrives at is the optimal one. If nash (or some other program) gives you some other solution, then just enter that solution into CardRunnersEV and recompute. You'll find that that solution will result in a lower EV for SB.

A request.

Can you expand the math engine for following situation?
2 players (heads-up)
you can start preflop with 1 player with all his hands
you can only start preflop with the second player with 1 hand (ie T7s, AQo)
you can set-up wathever preflop & postflop tree.

When you press EVrun, CREV computes everything exactly.

Strictly seen:
player 1 = 52*51/2=1326 combo's possible
postflop = 50*49*48*47*46/(5*4*3*2*1)= 2 118 760 combo's possible
player 2 = 1 hand is maximally 12 combo's possible

This results into = 1326 * 2118760*12/6=5 618 951 520 combo's to be required to run trough.
Essentially a factor 112 bigger then the current biggest monte carlo run.
Or a factor 56 bigger then the old 100M monte carlo run.

This isn't possible.
Enumerating that many matchups would take days or even longer.

I see. Next best 2 options.

Is it possible to increase the current max Monte Carlo amount of runs with 10, to 500M?

Or is it possible to add a delete action to where you insert flops. This so I can ie insert 200 different flops and all other flops will be considered to be non existing (like the delete action for hands) ?

I hope it's okey to ask in this thread. I'm following along what sthief09 is doing in There Will Be Homework episode 2, I have the script set up exactly like him but it doesn't work, I've tried playing around with different things but I'm not really sure what I'm doing. He gets a curved graph while I just get two downward going lines.

Tree


Script


Graph


Thanks in advance!

Changed the X-value in "Add data point" to "#(2)", think that fixed the problem, guess he made a misstake in the video? Weird that he got it to work though.

Sure, not a problem.

Well, you cóuld just add a delete action on the flop and delete boards that don't meet your requirements.

Well, I've made plenty of videos myself.
The correction was probably edited out by accident.

Thanks for the increase to 500M.

Hey,
When flop it is not entered, then you can't pop up Postflop hand visualization tool (Alt + A),
However you can get it by going to condition editor, Postflop range editor and see very same tree and graph there. Is it accurate? is there any reason, why it is impossible to call it with Alt + A, but it is accessible there?

Thnx.

Hey,

There is big appeal:

I am pretty sure that it was asked in this or some other forms before:
Ability to write some directives how to generate some random boards.

That could be like general strategy for entire project - you describe what kind of board textures should be dealt, for example just monotone board textures.

Or additional montecarlo directive, lets say that you put breakpoint at some very rare spot, that you just don't get enough situations occuring for particular filters and u run only those boards that will correspond to that situation.

Where could be that applayable? well- everywhere. For example I want to check situations where only monotones flops occurs. So if I will use montecarlo then with delete action I will waste 95% of itterations!!! very same for other interesting board textures, like High High Low, Ace and what is difference between two lines when flop is Ac2c5d and Ad2c5c. While last one could be replicated with two different board textures, still I can't compare and simialr parts of trees using action scripts.

Thnx.
I believe that it should be pretty easy to implement that. (at least directives for montecarlo for entire project).

So maybe that could be achieved, for example in that dialog:
http://i.imgur.com/cVnOlR7.png
By having some syntax similar to board structure synax (http://www.cardrunnersev.com/manual/boardsyntax.html)
Except That I would like to see ability to select particular suits as well.

Ok, clearly an oversight.
I'll look into it.

Ok, I've just had a look and there's a workaround for this.
On the flop, before beginning the actual tree, just make SB make a tiny bet (for example to 1 cent) and BB make a tiny reraise (to 2 cents).
Let SB's raise be under the required board conditions.
Otherwise let him delete.
Now, in a script, just draw a random board.

In the example below, only monotone boards will be drawn.



And for your convenience, here's a link to the savefile (all the script needs to do is draw a random board): www.cardrunnersev.com/savefiles/randomboard.stx

Hey, thnx scylla for that tip, but I know that and there is still few issues with that:

1) I still believe that it waste 95 of montecarlo itterations, correct? It doesn't generate additional trials for runs, that was deleted?

2) While it allows to generate monotone board textures, but it doesn't differencate between A2cc 5s, 25cc Ad and similar style. And that could be dramatical difference when blocking effects comes to arena.

3) While it works well while showing freaquencies, however Postflop range visualizer Graph doesn't work well:

There is two conditions:
Pocketpairs + Flush draw
Rest of pocket pairs (without flush draw on monotone board texture).

Graph shows all pocketpairs only as HC pocketpairs (flushdraw- supreme condiotion- is ignored).
That could be program bug tho.

Illiustration:

No, if you've set 100k runs, then you'll get 100k runs.
The program may have to go through a much larger number of runs in order to get tó those 100k succesful runs though.
So the simulation will take longer.

No, you can still do that.
For example, if you want a 2flush AXX board, where XX are cards of 2 through 5 and the A is non-flush, then you can enter:
- A[2,5]c[2,5]c + 2flush + no pairs
- A[2,5]s[2,5]s + 2flush + no pairs
- A[2,5]d[2,5]d + 2flush + no pairs
- A[2,5]h[2,5]h + 2flush + no pairs

Not sure.
Could you please send a savefile to support?

That is what I mean that 95% will be deleted (non successful), so wasted

AWESOME!!! THanks!!!
Didn't know that it is possible to tell condition what suit to generate - On the help page such thing is not listed!!

This thing solved(actually just allowed to use that as workarround) my issue with graph!


Unfortunately that decreases my chance to hit all clubs to 1 in 100 and accuracy of montecarlo
Sent.