Hey. This is my first post here although I've been using reading these forums for years. Thanks to everyone who posts here for being such a great help to me. Now I'm involved with the community I'll try my best to help others out as well.

Anyway, my question is regarding future play in general but more specifically in regards to 6 max hyper-turbos super-KOs.

My understanding of ICM is that it works on the basis that if everyone were to split the money at that moment what ICM concludes is what everyone should recieve. Now I understand that ICM has limitations and this (especially in this format) I believe constitutes a HUGE misjudgement of ICM (potentially game changing misjudgement) but I have no idea how to mathamatically work it out.

Let's just consider that everyone is hypothetically playing perfect NASH push fold for sake of ease.

If you were to have a big stack (or particularly short stack) in 6 max hyper-turbo SKO it means you can profitable shove more hands (sometimes any two from any position). So using this logic we can conclude that in many cases playing the next hand (rather than hypothetically take your ICM $EV) is actually going to be +$EV for a big stack because the accumalation of your $EV over your entire range is positive. This could then surely carry on to the hand after and so on and so on. If this is the case then aren't the $EV according to ICM predictions for big stacks massively underated due to what they can do with that big stack. If my logic is correct a big stack should add the average +$EV of his next hand and the one after and so on. Perhapes this effect stops when it goes heads up due to ICM reverting back to CEV (but I'm not even sure about that).

Am I thinking about this right or not? If I am does anybody have a clue how to accuratley calculate your actual $EV including future play.

p.s. I am aware that ICMizer 2 has got FGS but from what I've read that is only considering blinds and position not actually the $EV of playing your next hand(s).

Thanks in advance for any help. This is boggling my mind.

Hey again. I've studied ICM theory some more and I've concluded that perhapes I've misunderstood ICM calculation. If so it'd rectify this whole thing.

To formulate what your chip stack is worth according to ICM does is effectively simulate a game where people are playing perfect NASH push/fold with their available stacks and see what money they win on average?

If that's the case future play is no longer an issue because it's already considered what $EV you would win/lose in the future. Your stack is only worth the value of what ICM says BECAUSE it's simulated the future of the game according to perfect push/fold.

I'm sorry this is so confusing but I'm seriously confused myself as you can probably tell and figuring this out would be incredibly helpful. Thanks.

Yes, FGS calculates future hands based on everybody using NASH ranges.

FGS does simulate the next hand or hands too, but modern computers can't really simulate far enough ahead. You'd really want to be able to simulate it to the end of the game (or at least to heads-up).

But what little can be done with FGS is enough to show ICM is pretty flawed.

You think right about a lot of things, there is just this mistake:

Blinds and position are actually a function of future hands, they are accounted for in the FGS calculations.

e.g:
If u're UTG and using FGS 2 incorporating the next 2 hands in the calc, it will give u a bigger opt. jamming range than icm cause it knows u're gonna post two blinds and others won't have to. If u include the next 4 hands, so having 2 w/o posting blinds, the former opt. calc jamming range gets smaller again cause others are gonna post the blinds.

Here's info on how FGS works:
http://www.icmpoker.com/en/blog/how-...culator-works/

Q has lots of more infos on it, he's good

However always remember one important thing: FGS is assumed perfect nash equilibrium ranges in future hands. This is not reality in poker, ppl differ from it. This can have HUGE effects on what ranges are actually optimal in spots, esp. on/close to bubbles.

e.g.:
You're on the bubble of a 6m and calculate your openjamming range with FGS5 (which'd be a pretty cool decision on the model cause it has every player posting equal blinds ). This assumes both villains are making nash equilibrium decisions within the next hands. But one or both might actually be very bad at poker and call 60% instead of the optimal 12% or the big stack sb walking the bb where jamming 32o is +2% or w/e. So this can largely influence your opt jamming range, cause opp might be willing to make these mistakes FGS can not include in the calcs. Rarely is a model "perfect" in a situation where ppl aren't exactly regs who use the identical model and make the exact decisions.
We have to find one that suits reality the closest and then use our brains to adapt it to what we estimate the real villains to do. FGS does pretty well just imho and it really helps to practice with it for everyone taking sng poker serious

GL at the tables!

Thank you all for your help so far. I am aware that FGS does consider future play because blinds and position are a function of the future and therefore does affect your ranges. I'm also aware of how if somebody strays from NASH equilibrium it effects things a lot (thank God I discovered this early because I initially thought you would always be +EV no matter). But let me give you an example to better try and explain what I mean.

If ICM calculates that if you shove this hand (the hand doesn't matter really it's just any hand that'd be +$EV in the situation) you on average increase your $EV from $10 to $15 (which in this example puts you significantly ahead of everyone else) is it also simulating the stack advantage for the rest of the game? So to try and put it simply if for the rest of the game computers took over and played perfect push/fold based on NASH equilibrium would you end up winning $15 on average or would it be more because of the number of $EV plays you can make because you dominate the table. Hopefully I explained myself better.

This is not possible within a single hand. Ok we're not talking about some super short stack spot in a 5k $ SnG here, are we?


No, just for the x in FGSx.

This is way too complicated to solve for computers as of today (thank god, solved games aren't beatable longterm if botting can't be prevented and the word spreads). U'll prob run into trouble with FGS3 on a 9h table already, let alone FGS100 or sth

Well, if you went all in vs another stack your size or slightly under holding AA I think you would probably increase your $EV on average from $10 to $15 on a 6max ko table. In the first hand if you double up you not only double your chips but you also win a bounty. Especially if you're not risking your full stack it's close to CEV actually if my calculations are right. But this is besides the point it was just arbitrary numbers really :P

So... I think most players can agree there is a substantial advantage to being the dominating stack ESPECIALLY in a gametype such as 6max hyper kos therefore, in your opinion, does ICM considerably underrate the actual $EV for the big stack in my given example? If so then that's HUGE. That's literally game changing.

Yeah it underrated big stack value but not so much that u need to play a totally different style. Just use FGS(players-1) and u're good to go

Those 6m hyper ko are cool cause the bounty is so big that a single chip more or less than villain can literally switch ranges. I love them, w/e complicates decisions creates edges.

Exactly why I play them as well. I was primarily a cash game player but I switched to SnGs when I realised the skill egde you could have on low level players. Like you said the more complicated the ICM decision the more edge there is. What better than 6max hyper KO Well thank you for your knowledge and help. Good luck on the tables bud.

No worries, gl yourself