Why are the numbers different in each tab? Table view says BTN should push 21%, Details says 32%, Equilibrium Tree says 45%. Which one is the "solved" percentage?

ignore the Details View for now... It will make more sense on the next update.

Everything that is more then 20 sec is too much. Thing is when I'm reviewing my hands I have a bunch of them, and waiting for every one for a half minute is irritating.
It's different when you want to review one hand and want to get best results, then even 2-3 minutes are tolerable.

When you decide to implement hand histories, you could make sng solver go through all the hands automatically, in the background, while we are reviewing hand histories one by one. This way players wouldn't have to wait at all, and processing time wouldn't be that important.


Thanks for the info, I'm an amatuer programmer and always love to hear these kind of things.


I didn't notice before that range expanding is different in wiz then in sngsolver. This is the main thing why results are different. Also probably selective memory and isolated case where sngwiz showed better results (closer to advanced icm). So, forget about that.


For 3 or 4 players, would going deeper in analysis show visably better results? I thought that computing time is really fast for 3,4 players and you could make more complex calculations there..


About that script I made, it should work on any windows, if somebody wants the code I'll send it, it's in c#.

any problem with Ps if i run it same time the client?

i just dled this and it says beta version expired.

i sent an email, they fixed it... download again : )

What's this about all other sng ICm calculators being wrong. I can assure that position makes a difference in any calculator, it would be way to obvious if it didn't.

how do i factor for villains caling range?

also how do i use the tree thingy jiggy?

None of the standard ICM equations (Malmuth-Harville, Malmuth-Weitzman, etc...) make any provision for position. The only input variables are the stack sizes and payout structure. So, any tool or program based on standard ICM is going to suffer its weaknesses.

I discussed some examples of this in an earlier post in this thread here: http://forumserver.twoplustwo.com/sh...2&postcount=38

Given the thousands of posts in the strategy forums here on 2+2 that talk about having to "take a -EV edge in early position" based on results from some "other" programs, I'd say their problems dealing with position are obvious indeed. I humbly suggest that you've had to live with those shortcomings for so long that you've simply learned to overlook them.

If you click on any opponents's range in the Table View, it will bring up a range edit dialog.

The Equilibrium Tree view displays the *default* ranges for each possible situation at each position but is not interactive.

The Details View will soon contain lots of jiggy thingys to play with (I'm working on that now).

You basically make no sense at all, asigning a numeric value (pulled out of your arse) which you call equity to positions is no advance from any ICM calculation.
I love it how you say
When it in fact makes no sense. What are you trying to say when you say SB has equity?
Open up any calculator and you'll see how it tells you to shove any two if you have 2BB in small blind but it tells you to shove a lower range on UTG. Come on, check it up yourself.

Edit: Or maybe I'm just a jerk and I'm wrong, I'll give it a try.
And it weighs 470MB you can work on that too.

?

LoveThee,

To be honest, I don't think I completely understand your post... so I apologize if my response seems generic...

Converting chip count into prize equity is going to be at the heart of any tournament strategy calculator. In fact, this is exactly what ICM does. Additionally, the limitations and problems with the standard ICM equations are well known. There are probably hundreds (if not thousands) of posts on these forums confirming this fact.

An alternative to using ICM is simulation. This is what SnG Solver does. It simulates literally millions of possible outcomes for a given hand to come up with equity estimates for each stack in play. And, I assure you, my arse is not capable of that kind of math.

Yes, a 470MB download is not small and, if anything, it will probably get bigger in the future. But its that big for good reason. There's nothing else on the market that does what SnG Solver does and, if there were, it would need to be just as big.

When you say "Ps", you're referring to PokerStars?

If so, I'm not really familiar with their T&C so I cant say if there's any problem. You'd be better off asking them.

Maybe somebody else has an idea about this?

I notice what you mean now, the equity a player has on the prize pool on CO is bigger than the one on BB.
But when taking decisions all ICM calculators take into account position because they also take into account other's villains ranges.
I'll take a look.
I'm trying to compare this with pokerstrategies ICM trainer, take a look. It's 10 times lighter, is the program stupid?

For 2 players, ranges are different if payout is 1st=1000 and if payout is for example 500/300/200.

Bug is still here. Let me raise the significance of this...

If I play 9man, 3 players are left, equilibrium ranges change if payouts is 50/30/20 or 62.5/37.5 (which is normalized 50/30)

If I play 18man, 4 players are left, equilbirum ranges change if payouts is
40/30/20/10 or 44,4/33,3/22,2

Which should be more accurate?
It seems like an easy fix cause all you need to do is to ignore payouts for which applies payout place >= number of player.

Yeah, I know I said I had it fixed for the next version... theres sort of a long story as to why its not, which i wont bore you with here...

But suffice to say that it is a bug of the oops-the-decimal-place-is-in-the-wrong-place variety...and not an error of process. It'll be correct in an update real soon... I appreciate your patience

A new version (0.9.13) is up at the usual spot: http://www.sngsolver.com/beta.html

This version has a couple important things... 1) a fix for the problem ginandbread mentioned above and, 2) a much improved relationship between the "equilibrium" ranges and the recommended "Hero" range. There was a rounding issue that could cause them to diverge by quite a bit and this was causing a lot of confusion.

There's no "Detail View" yet... its not quite, but and I wanted to get this update out with its math fixes ASAP.


Enjoy!

... and of course an hour later I notice that I broke one of the graphs. I guess #13 was unlucky for me.

0.9.14 is up.

Once you click on a hand to see the numbers (Q2o for example), how do you turn it off? I can't get rid of it without clicking on another hand. I want the little box to go away.

If you click on the card bitmaps that represent the Hero's hand, you'll bring up the Hero hand picker. On that dialog is a button marked "??". Clicking on that will clear the Hero's hand selection.

You sure?

I see it works good for 2 players, but for >2 players still isn't right.

Your example doesnt have equivalent payouts for 3 players... 50-30-20 vs 50-30 are supposed to be different. 50-30-20 should be the same as say, 30-10 or 75-25.

So, I think I'm sure... although, I've been workin on this stuff all night so maybe my brain has finally turned to pudding.

Oops, mine is already puding obviously. I was looking at ratio between first and second which is 5/3 and is wrong.
If there is three players left we already paid 60% of reward, not 20% and so ratio between first and second is 3/1. Ok, nice work!