Can someone please tell me a situation where what would normally be a profitable shoving hand on the button in a 10 seater table with 10bb no longer is a correct shove because of icm and how and why icm changed the situation. Would love to wrap my head around this.

If there were a few players with short stacks, they are more likely to blind out or bust out before you do. Your share of the prize pool would increase more by folding and letting other people bust out than by shoving to try and add a couple of blinds to your stacks.
The underlying concept is that the chips you win are less valuable than the chips you could lose. The most valuable chip in your stack is the last one, since that represents your tournament life. To put it another way, doubling your stack is nice, since it increases your chance of winning the tournament, but it doesn't guarantee you'll win the tournament, and even then you don't win the entire prize pool, just a fraction of it. Losing your stack guarantees you lose your entire buyin. Losing a 10bb stack and busting the event, and then seeing some others cash with 2bb stacks is absolutely sickening. You can minimize the risk of that happening by not putting your tournament life on the line unnecessarily.
In most cases, the ICM factors mean you can shove wide, but have to call tight, but often when you have a middling stack (and the players to act are either much deeper or much shorter) you have to play even tighter than that, because the risk is too great for the potential reward. #ICMcoffin

This article by Dara O' Kearney explains how ICM is particularly important in satties, but the explanation is relevant to any SNG or MTT with more than one player getting paid: http://www.bluffeurope.com/PokerMaga...attle_238.aspx

This, basically. It's all down to the variable payouts, and it's most recognisable around bubbles as people have stated, but has applications everywhere. For example, we're 3 handed in an MTT, prizes being 40k, 20k and nothing, and we can quantify our EV right now to be 25k. Button folds, SB who covers us jams. Now if we double up, our chips double, but does our EV?

Thanks for the post Arty

Ok arty I think I get it so let’s say I was playing a $10 10 seater sit n go and 1st got $80 and 2nd gets $30 and the firsthand I knock out 7 players. So I have $70 and the other 2 guys have $10 each and in the next hand one player jams with aks and accidentally flips his cards and the other short stack has 22 then he is but isn’t really risking $10 on a coin flip he’s really risking $30 on the coin flip. Is that right?

"the chips that you win are less valuable than the chips that you lose"-arty

sums it all up beautifully right there i think.

p.s theorangeone I like the question but i've had red wine right now so i can't answer it

Assuming there is no rake (big assumption!), the total prize pool is $100, and a starting stack is worth $10. As soon as the stacks are uneven, however, 10% of the chips is no longer worth exactly 10% of the prize pool, because the prizes are also uneven. Notice that even if someone wins 100% of the chips, he only wins $80, or 80% of the prize pool. (2nd place will get $20, not $30, unless an additional $10 was added in overlay).
You can type the stacks and prizes into an ICM calculator to find the ICM value of each stack.
e.g. If starting stacks are 1k, and you knock out 7 players on the first hand, you now have 8000 chips (80% of the chips in play), and you're looking very good to cash, but you still haven't guaranteed you'll win $80. If you stole the blinds on the next hand and got your chip stack to 8150, you still can't win more than $80 at best. Those additional chips are almost worthless. In fact, you haven't even guaranteed that you'll win $20, because there's still a chance that the other players will go on heaters and bust you on the bubble. Nevertheless, having 80% of the chips in play in that spot would mean your EV is high. It's $67.56:


http://www.icmpoker.com/icmcalculator/#PRXI

Notice that the other two players only have 10% of the chips, but their stacks are worth $16.22.

Thus, if one of the shorties goes all in on the next hand, the other needs to be a significant favourite to call, because his 10% of the chips has an ICM value of 16.22% of the prize pool. He might double up in a flip, and then be guaranteed at least a $20 payout, but he'll still be way behind the chip leader.
Running a new calc, a double up means his chip stack will be worth $32 heads up:

http://www.icmpoker.com/icmcalculator/#yiGr

So you see, doubling up in chips doesn't double the ICM value of those chips. ($32 is less than double $16.22). (As an aside, notice that even when it goes heads up and the big stack has a 4:1 chip lead, his ICM value is not much more than double the shorty's).

Now let's use the concepts in game simulation. Hero has 80bb on the BTN. SB and BB had 10bb each before they paid the blinds. In such a scenario, hero can jam very wide (79%), because both of the shorties are risk averse. They effectively have "too much to lose" by calling, so the SB only calls with 9.4% and the BB calls with 15% (getting better pot odds). The big stack applies maximum pressure when the other players are in ICM handcuffs. He shoves super-wide and they have to call very tight. This typically means the big stack keeps adding chips by stealing blinds from the "risk-averse" opponents, such that he's a huge leader when it finally gets to heads up.



If hero open folds, then the SB should push with 56% of hands. Even though he started with the same stack size, and only needs to call 9bb, the BB can only call with 35%. He literally can't take "flips", because he won't break even by doing so. He has to fold most of his range, and hope that the other shorty makes a mistake before he does, or he wakes up with a hand in the SB and gets a double up vs the big stack.

P.S. A more extreme version of "ICM handcuffs" would exist if the chip leader had 80% of the chips, the middle stack had 15% and the shortest only had 5%. The middle stack would be in a coffin in that spot. He can't risk going broke against the big stack, but he also doesn't want to risk doubling up the shorty. A very short stack is more likely to "want to gamble", because he can't sit around and wait for the middle stack to blind out. The big stack would often jam any two when then middle stack is in the BB, because the middle stack has to play so tight in such a spot.

im sorry about my nonsensical sit n go example because i was in a rush and didn't proof read my comment. What i had meant was 1st gets $80 2nd gets $20 and two guys were coin flipping for 2nd place.

But thanks for the detailed reply. This is why you are so well respected on this forum. I'm gonna be using that icm calculator a lot from now on. You seem pretty on top of your icm game so i just wanna ask you another quick question and that is: at the very start of this $10 sit n go will icm interfere with any of your pot odds decisions in any significant way or does icm really only come into play closer to the end of game where players are on the bubble and/or in the money?

also in a 10 seater 50/50 style game is icm still a factor in these games?

I'm far from an ICM expert. I make mistakes in my satties all the time, because multiple stack sizes can lead to counter-intuitive strategies. Fortunately, most of my opponents make even bigger mistakes.

ICM pressure exists throughout a tourney (a double up on hand one doesn't precisely double your expected share of the prize pool), but it greatly increases the closer you get to a pay jump. The bubble is effectively the biggest pay-jump of all, so that's where the ICM factors are usually greatest.

50/fifties and DONs have extreme ICM pressure on the bubble, a bit like satties do. Typically, it means that medium stacks should be the most risk-averse (i.e. nitty), so those are the players you should be jamming on if you have a stack that can heavily damage or bust them. Probably the most stressful position to be in in a 50/50 or satty is second to last on the bubble, as you can rarely jam, let alone call a shove (your hands are tied) and you have to hope/pray that the shortest of all gets all in and loses, or two other players commit ICM suicide. In those games, your position on the leaderboard as the bubble approaches is kind of more important than your actual stack depth. i.e. You might only have 3bb (which might make you think you're desperate to chip up by jamming/calling very light), but if someone only has 1bb, your stack is way too valuable to risk, since the microstack could double up and still be in last. You just need to do whatever you can to not be the player that bubbles. If 1bb is enough to outlast someone with <1bb (everyone has a story about binking a satty or DON with only one ante left) then it's enough.
A big chip leader in a 50/50 or DON can often jam ATC every hand if he's got everyone well covered, and it's just a case of waiting for someone to wake up with aces (and hold) or get forced all in blind before anything changes. In those games, it's important to see when the blinds are going up and whether someone else will be forced all in before you. (Stalling is sometimes appropriate, as is turbo-folding if you want to get through the blinds before they increase and completely cripple you). The strategy is surprisingly complex (and very different to cash or mainstream MTTs), and fish tend to be very bad in those kind of games. In my experience, they get quite boring after a while though, and are only fun if you're the chip leader. :/

Okay but one thing I’m now struggling with is why do people bother to learn push charts for tourneys when icm skews ev so much? Like is my beloved ace x rule not as relevant in a tourney vs in a cash game?

In Harrington on Holdem, IIRC Dan Harrington did five pages of math to show what the considerations were for each player in an example similar to this.

In my opinion, it all comes down to one thing. Put a percentage on how likely you would be to win a ticket (or whatever it is you win in the satellite) if you shove. Keep in mind that even against just one player aces lose about 20% of the time. You're playing aces against the field.

Now, consider how likely you are to win the prize if you fold. Keep in mind that only one player has to be knocked out for that to happen.

You don't really need math skills to figure this one out.

Push charts are useful for generic MTT situations, where bubble factor is only relevant intermittently, since those charts are based on ChipEV. They are less useful for SNGs, where ICM is a major consideration. That said, I'd still use Push-charts as a guide for SNGs, but be cognizant of the ICM pressure. e.g. a voice in my head will say "The chart says I can shove X% of hands in this spot, but the ICM pressure on me is quite strong, so I'll jam a little bit tighter", or "There is extreme ICM pressure on the guys to the left of me, so I'll jam a lot wider than the chart says".
One of the things I've noticed in the satties that I play is that many of the bad regs seem to be pretty good at making ChipEV+ shoves and calls (as if they are looking at SnapShove), but they make horrible decisions regarding the ICM effects, as if they're not paying any attention to the overall stack situation or the bubble. I think it's partly because they are also playing standard MTTs at the same time. Satty/SNG jamming ranges can be very different to standard MTT jamming/calling ranges. I mean, I've open jammed or 3-bet jammed 18bb with 7-high a few times in satties. I'd never do that in an MTT, unless I was somehow the chip leader on the final table in a spot where the ICM pressure was huge.

Long time since I did this tourney/ICM stuff, but nobody mentioned it yet...

You can "learn" ICM ranges simply by using the ICM simulator (ICM quizzes) in software like SNGWIZ (assuming that all still exists idk).

It doesn't take long, it's easy to do, it's a little bit fun (not much!) and it will quickly show you that, in many spots, the right decision is quite obvious (and not far off your push fold charts usually).

It will also show you that the CEV differences can be tiny/marginal, so your ICM errors when you make them only have small consequences for your win rate because you simply won't be making any huge ICM howlers and your decisions will be "ICM correct" 80-90% of the time, even if you are not that great at it.

that's exactly the kind of answer i was looking for too. Thanks man. noted.