thought re ICM

[rong]

OK, so we're late in a game, 4 handed on the bubble, big enough blinds and equal enough stack sizes that everyone is playing push fold poker & everyone basically understands ICM, whether or not they are accurate on their ranges is somewhat irrelavent, but lets assume all are somewhat aware.

ICM states that for each position there is a particular range of hands we should be shoving & a range we should be calling with (based on stack & opponents calling ranges etc etc)

Lets assume the following positions means the following ranges for shoving (ignore the figures, its for an example) and we are always either folded to or shoved into:

For pushing:
BB - wont happen as we are either pushed into & fold, or folded to
SB - Push top 70%
Button - Push top 40%
CO - Push top 20%

Now lets concentrate on the small blind position for this example. Push top 70% - Does this mean we should push whenever we get cards that fall into the bracket of the top 70% of cards, or that we should push ATC 70% of the time?

Because I have thought about this, and variance means that over 10 hands, its unlikely that exactly 7 out of those 10 will be in the top 70%. However if our opponents know ICM relatively well and expect that we do also, they will expect that our shoves will be with cards from the appropriate range, and therefore only call with the even tighter appropriate range. So if we are having a bit of a bad run of cards, we can assume that opposition will not know this and therefore we don't actually need these cards. Cos lets face it, villains calling range (perhaps top 15% say) smashes our whole range, regardless of whether or not the cards are in the bracket of top 50-60% or the top 20-30%.

Does this make any sense to anyone at all?
Am I missing the point of what ICM is telling us?

[Fielmann]

Push top 70% - Does this mean we should push whenever we get cards that fall into the bracket of the top 70% of cards, or that we should push ATC 70% of the time?

The former.

However if our opponents know ICM relatively well and expect that we do also, they will expect that our shoves will be with cards from the appropriate range, and therefore only call with the even tighter appropriate range. So if we are having a bit of a bad run of cards, we can assume that opposition will not know this and therefore we don't actually need these cards.

You're levelling one step too much (or too little, idk) here.
That 70% pushing range, you see, is calculated, taking into account BB's calling range, which, in turn, is already considered as being modified by our past behaviour.
Whenever we have a hand that's not in the top 70%, and fold, our opponent's calling range shrinks (e.g. from 15% to 13%), which means that next time we can push with a somewhat weaker range (e.g. top 75%).

[rong]

so essentially, by playing our shoving range, which is based on opponents calling range, we have a nash equilibrium and cannot improve on this situation?

[Fielmann]

so essentially, by playing our shoving range, which is based on opponents calling range, we have a nash equilibrium and cannot improve on this situation?

I think that if everybody at the table
1) correctly estimates other players' shoving ranges and calling ranges
2) correctly calculates their own shoving ranges and calling ranges
then I guess we have equilibrium and casino is the only one who wins.

In real life, however, people make mistakes. By being better at 1) and 2) we make fewer mistakes than them and therefore we win.

I don't really know much about game theory, though. If anyone sees things differently, feel free to say so.

[JR9477]

I've wondered if this is the case in higher level SNG's. Do the players there split dead money from the few fish in the long term?

I agree that most people do eventually make mistakes, whether it be by frustration, desperations(blind increases), outside influences, etc. I've a had SNG's where we're at a stalemate for a while, but eventually someone slips up. Also, sometimes people are making minor mistakes with ICM, and the gains take quite a few orbits to appear.

[taipan168]

I've wondered if this is the case in higher level SNG's. Do the players there split dead money from the few fish in the long term?

I agree that most people do eventually make mistakes, whether it be by frustration, desperations(blind increases), outside influences, etc. I've a had SNG's where we're at a stalemate for a while, but eventually someone slips up. Also, sometimes people are making minor mistakes with ICM, and the gains take quite a few orbits to appear.

This is why I think that the Stars SNGs are dead above the $60 level. Because decisions become pretty simple at SNGs (unlike deep stacked cash games or MTTs), everybody realises that with the rake, there just aren't enough bad players around to contribute to the good players to justify the entry fee, so all the good players end up just handing money to each other (and to Stars).