ICM - Independent Chip Model

  • Sit and Go
  • SNG
(24 Votes) 16762


Free membership

Join now


pzhon shares with us his thoughts on ICM, the Independent Chip Model, and how in his opinion, it should be redesigned to make it a more accurate reflection of current tournament situation


ICM Theory Video

Comments (32)

newest first
  • FishermansFriend


    Enjoy the new theory video by pzhon!

    Please keep the comments in English!
  • Laccyby


    Error Code: NetStream.Play.StreamNotFound

    Die fehlermeldung bekommt man regelmäßig wenn das video grade hochgeladen wurde. könnt ihr da nicht was fixen damit es von anfang an klappt?
  • Ribbo


    laccyby, clear your internet cache and try again
  • furminator


    thanks a lot for this video!
  • MikeAK47


    Please keep the comments in English!

    Hahahaha next reply is in German.
  • ElNunho


  • johncec914


    gracias x el video!
  • RealRauch


    Very good job pzhon.
  • alex91sry


    #4 gutes Video!
  • fish27o


    köszönjük :D
  • Maniac81


    i have a question: At the end of the video you say that we should take a -EV Edge when we have a small Stack. Do you think it is better to distribute the Edge over several Positions?
    For Example:We have to take a -0.3 Edge and we take a -0.1 EV Edge from every Position. Then we can take a -0.1 EV Edge from UTG and dont have to shove so much -EV.
    I hope you understand me because my english is not so good.
  • elhh82


    at minutes 22, you have the looks ahead values. How did you get the steal 0.897% and so on values? I only get the fold = 0.937 value.
  • Rabotorg


    Очень хочется перевод на русский
  • pzhon


    @11 It's not that we are required to make some plays the ICM says are -EV. The ICM is not perfect, and so some of the pushes which are slightly -EV according to the ICM are actually profitable. The calculation was for UTG. You can't divide that -0.3% among other positions. The other positions have their own values which may even be positive, so that you should not push hands which appear to gain equity according to the ICM.

    Free hands outside the blinds generally give you equity. Hitting the blinds costs you equity. So, you should lean toward plays which let you get more free hands by pushing slightly less often from the SB, and lean toward pushing slightly more often UTG so that you can bust out and skip posting the blinds some of the time.
  • pzhon


    @12 To get the other values, you set up different stacks and then advance the blinds, then use the EQDiff reported by the Nash equilibrium calculator. You have to make some assumptions about who won or lost after you fold, and who called you if you push and get called. In some hands it would make more of a difference, but here I assumed that there is no collision if you fold, and that the big blind is the player who might call you.

    The 59% came from the ranges entered into SNG Wizard. You can see it at 21:00. You may need to toggle the analysis details on to see it in SNG Wizard. The other percentages used to weight the average were estimates based on the equity of A7s against some calling ranges.
  • iNspiRe


    Hello there, Douglas!
    Here is one question from russian community for you:
    Is there any fast method to calculate changes of our push range when blinds are about to hit us, but without using Nash calculator? For example: We are five-handed. And if out stack is about 6-8 bb => edge = -0.31%, but if our stack is about 3-5 bb => edge = -0.45% and so on. Maybe there is some chart that can illustrate these differences? Thank you
  • pzhon


    @16 That's a good question. I have only studied a few isolated examples, but I believe the pattern depends on the other stacks (number of players and distribution), not just your stack size, and you also have to factor in the absolute size of your stack. If you have a 6 bb short stack, you have less equity to lose than if you have a 6 bb average stack. I may work on this and present some conclusions as a part of a video on -EV shoves later. Please let me know if someone else makes a table of some edges to use UTG.
  • iNspiRe


    #17 Thanks for answer! If we will find some table like that - we'll let you know
  • lessthanthreee


    great job pzhon. loved the last 5 minutes especially.
  • lessthanthreee


    hi pzhon,
    im struggling to get the correct values for steal as 0.897%. must be doing something wrong with setting up the stacks and advancing the blinds. can you post the correct nash inputs to get this result so i can work backwards?

  • pzhon


    @20 Stealing meant pushing and taking the blinds, so your stack increases from 1200 to 1500. So, at 100/200 we get stacks of 2400, 1500, 4000, 1500, 2600, and your 1500.

    I hope this link works: http://www.holdemresources.net/hr/sngs/icmcalculator.html?action=calculate&bb=200&sb=100&ante=0&structure=0.5,0.3,0.2&s1=2400&s2=1500&s3=4000&s4=1500&s5=2600&s6=1500&s7=&s8=&s9=
  • lessthanthreee


    oh i see what i did... i thought the stack under the BB/SB in sng wiz was pre-blinds, when it is post-blinds.

  • lessthanthreee


  • lessthanthreee


    #23. nvm i figured it out.. looks like the comment bugged out anyway


    OMG this is incredible hard to understand this has completely gone over my head its even made me forget what i thought i knew about icm :(. I'll keep being a lil fish in a big pond
  • pzhon


    @25 My plan wasn't to make it hard to understand, but there is a lot of material in this video, more than you can get in one viewing. If you feel this contradicts other things about the ICM you have heard, please make a thread about it in the forum, and I can either clarify something from the video or confirm that the other idea was a misconception.
  • Tim64


    Awesome! Thanks very much.
  • DeBats


    You can learn the ICM way by understanding the model. It`s inpossible to use it while playing, right? It takes way to long to setup etc.
  • gp00053


    Just finished grerat video
  • chaserr


    Very nice vid, great explaining of ICM.
  • Th0m4sBC


    great video
  • ElPokerin0


    well done, thx for that!