Hello!

I wanted to make some ranges for my short stacked play. I registered for Holdemresources and did some nash calculations. When I tried cross checking a spot in Icmizer I saw a very noteable deviation. I realized that HMR and Icmizer didn't calculate the same numbers.

So my question is, which solver should I use for SnG's? FGS or Nash/Icmizer or HMR? I play 5 max if that's relevant. Also, is the one better than the other when it comes to MTT's aswell?


Thanks a lot!

Give an example of where ranges differ.

Generally speaking, if the setup is identical then the results should largely match between both programs. Some deviations are to be expected for Nash ranges, the Nash ranges are only approximated in both programs. (HRC displays the approximation "error" for Nash ranges as Exploit%, not sure if a similar display is available in Icmizer.)

For FGS calculations you would also need to make sure that both programs use the same depth settings etc.

Thanks for your answers.

Apparently I've done something wrong here or I'm experiencing a bug in ICMizer.
I plot all the info in ICMizer, click "calculate FGS2" and get a pushing range.
I will use this spot as an example.



I entered this hand and ICMizer gave me a 4.7% range, HRC gave me a 16% range. Then, without adjusting anything I clicked "Calculate FGS2 Nash eq.." in ICMizer and got a 19% range. I tried clicking "calculate FGS" again and now the range was 19% there aswell.

This seems to happen with every hand in ICMizer. I fill in a spot, solve for FGS and get a weird answer. Then when I solve for nash I get a better answer and now FGS also adjust the range to about the same.

Maybe this became a bit different problem than what it originally was, but does anyone know what I'm doing wrong here?

Thanks!

You should probably add for those that don't know that the payouts are in a 2-1-0 ratio.
(Given there are 10K chips in total on the table I'm assuming this is a 5-man SNG on Unibet).

I also get around 16% if I put HRC in raw ICM mode. I get 21% with FGS2. IMHO FGS2 is a better reflection of real play and therefore more accurate than ICM.

Poster "Q" will probably come along and talk more about ICMizer, but the ranges you quote for HRC seem correct.

btw are you entering the stacks of the blinds as if at the start of the hand (pre blinds)?

"and now the range was 19% there aswell."

Where is "there" in the above?

Also what do you mean by "Nash"? Everything these solvers calculate is (pseudo)nash equilibria, the difference is whether the payoff table is chipEV, ICM, FGS1, FGS2, FGS3 etc.

-btw are you entering the stacks of the blinds as if at the start of the hand (pre blinds)?
Yes

-Where is "there" in the above?
The range I get when I click "Calculate FGS2". It changes from 4.7% to 19%.

-Also what do you mean by "Nash"?
I thought that what I calculated in HRC (ICM) and the top button (Calculate FGS2 Nash eq..) was "nash" and that the button under (Calculate FGS2) was something different. But I see that I got confused because ICMizer somehow gives me a wrong range the first time I check it. I see now that you get approximately the same answer however you calculate it.

Thanks for you answer!

you have to click the "calculate FGS2 nash equilibrium" button first and then the "calculate FGS2" button.

if you click the "calculate FGS2" button first, it will calculate your range based on the ranges of the opponents, that are set initially.