Where ICM is lacking?

[ChipLeader]

A lot of people read (and flamed) my post yesterday about preflop decisions on the bubble of STTs in low stakes touraments. I was told by many to look into and research ICM, which I gladly did (including reading a post by INSTA or something earlier today that has drifted into later pages and i fear wont be coming back.)

First, before anybody gets the wrong idea, i think ICM is an invaluable tool to understand. Do not think I am discrediting it.

I do however, see a few potential problems in a 4 handed situation. Either I dont understand the concept completely correctly yet or there are a few flaws.

A.) Skill is totally negated
If you think you are significantly better than the other 3 players, or have some very solid reads, I feel you can exploit them better making raises and playing poker rather than deciding whether between only 2 plays.

B.) Constantly putting YOURSELF in the line of fire
While it may be a +EV push, taking all these small edges forces you to be the one constantly risking your stack. If I was in a situation where 3 other players were pushing preflops back and forth constantly, i would simply sit back waiting for a monster or til one of them beats another.

C.) Giving up better spots
I think this is my biggest problem with it. As Sklansky said in I forget which book, probably TOP, "A good bet is NOT a good bet if it prevents you from taking a better one later." Pushing because a play is +EV now is still not correct if it means risking a stack now when you can do it later with a better EV.

I think all of these points kinda blend together in ways, but these are the areas where I am struggling when trying to apply ICM. I played a few SnGs earlier just to fiddle with the calculator and try to ingrain the concept better and while it enlightened me as to how many spots I could push and be technically correct, it seemed to advocate too much aggression.

Just my opinions, im sure i have some of this wrong as i am only beginning to get involved in the topic but I would really like to hear as much as I can about this from people who have been using it for a long period of time.

[KyleH186]

ICM doesn't lack. The only time ICM can even be refuted is in a situation where taking a slightly -EV call in order to have a bigger stack and abuse ICM more later on.

Skill in SNGs comes down to how well you know ICM. Don't fool yourself into thinking you can outsmart someone playing an unexploitable strategy.

It is unexploitable. SNGs are a mathematically solved game. They are very different from cash games in this respect.

You can make preflop raises that are all in if the blinds are not big, but once those blinds get up to where you have 10-15 BBs, you better be pushing and folding.


Edit: As far as giving up +EV spots for better ones later, that is a flawed concept. ICM deals with +$EV, which means that you will make more money by making the plays. Folding +$EV hands will lose you real dollar value in the long run, its as simple as that.

[ChipLeader]

[ QUOTE ]
ICM doesn't lack. The only time ICM can even be refuted is in a situation where taking a slightly -EV call in order to have a bigger stack and abuse ICM more later on.

Skill in SNGs comes down to how well you know ICM. Don't fool yourself into thinking you can outsmart someone playing an unexploitable strategy.

It is unexploitable. SNGs are a mathematically solved game. They are very different from cash games in this respect.

You can make preflop raises that are all in if the blinds are not big, but once those blinds get up to where you have 10-15 BBs, you better be pushing and folding.

[/ QUOTE ]

So your answer to my post is "its unexploitable because it is?"

The point I made in the "C.)" section explains why i disagree with what you just said. ICM, as I understand it, does not take into account future opportunities, it only calculates the one at hand. If it is truly exploitable, I would like to hear how. The whole point of the OP was to point out where I think it IS exploitable. If you can show me why im wrong id love to see it.

[KyleH186]

It is unexploitable because if I am shoving correctly, and correctly assessing what hands you will call me with, the only way you can cause me problems is by hurting yourself equally or more and spite calling.

[jukofyork]

[ QUOTE ]
[ QUOTE ]
ICM doesn't lack. The only time ICM can even be refuted is in a situation where taking a slightly -EV call in order to have a bigger stack and abuse ICM more later on.

Skill in SNGs comes down to how well you know ICM. Don't fool yourself into thinking you can outsmart someone playing an unexploitable strategy.

It is unexploitable. SNGs are a mathematically solved game. They are very different from cash games in this respect.

You can make preflop raises that are all in if the blinds are not big, but once those blinds get up to where you have 10-15 BBs, you better be pushing and folding.

[/ QUOTE ]

So your answer to my post is "its unexploitable because it is?"

The point I made in the "C.)" section explains why i disagree with what you just said. ICM, as I understand it, does not take into account future opportunities, it only calculates the one at hand. If it is truly exploitable, I would like to hear how. The whole point of the OP was to point out where I think it IS exploitable. If you can show me why im wrong id love to see it.

[/ QUOTE ]
It's got nothing to do with ICM itself. The problem lies with the fact that current software (SNGPT, SNGWiz, etc) only applies ICM to each of the outcomes at the end of the hand.

Juk [img]/images/graemlins/smile.gif[/img]

[holy32]

Just a quick comment on C).

You're asking "why should I risk my stack now if I can get it in better later". This is where I can see you missed the whole point of ICM.

If you assign ranges correctly (for the sake of this example let's say perfectly) you will make the equivalant of real money dollars when you make a +EV push/call. There is no disputing this point.

If a guy bet the river in a cash game and you can see his cards and know you have him beat, you wouldnt fold either, right ? Right, because it's throwing away money. Same thing is true for +EV situations. If you know a push/call is +EV, you are going to make money in the long run by pushing/calling instead of folding/waiting.

Of course there are situations in which pushing/folding are close. Here is where the "edge" comes into play (sngwiz uses the term, you can call it whatever you want). B/c none of us can assign ranges perfectly and there is always the chance of someone to act behind you busting without you pushing (which is +EV for you), you will sometimes pass up on marginal situations as it is possible they are -EV (if you assigned wrong ranges) or someone might get busted which is more +EV than your push.

[tshort]

[ QUOTE ]
It is unexploitable because if I am shoving correctly, and correctly assessing what hands you will call me with, the only way you can cause me problems is by hurting yourself equally or more and spite calling.

[/ QUOTE ]

Kyle,

You are being short-sighted and the OP is correct. There are spots where even though you probably have the correct range and it is +ICMEV it could be better to make the other play.

[ADLinden]

[ QUOTE ]
[ QUOTE ]
It is unexploitable because if I am shoving correctly, and correctly assessing what hands you will call me with, the only way you can cause me problems is by hurting yourself equally or more and spite calling.

[/ QUOTE ]

Kyle,

You are being short-sighted and the OP is correct. There are spots where even though you probably have the correct range and it is +ICMEV it could be better to make the other play.

[/ QUOTE ]

I think by this you mean tho that you will have a more +EV move in the future, which is essentially still using ICM

[KyleH186]

The whole point of ICM is that it calculates the effect of a short term play on your long term profit. Thus, following ICM cannot be short sighted.

[tshort]

[ QUOTE ]
The whole point of ICM is that it calculates the effect of a short term play on your long term profit. Thus, following ICM cannot be short sighted.

[/ QUOTE ]

ICM models equity in tournaments according to current chip stacks. While it has shown to be close in most situations, ICM does not perfectly model your equity.