ICM/SNGPT rambling thoughts(long)

[AliasMrJones]

Look, +real-prizepool-money-dollars-EV is not the same thing as +chip-EV. I think some people are confusing what ICM does. It models your share of the prize pool, not the chips. Something that gets you a bigger share of the prize pool (real money dollars) can never be a bad thing.

The assumptions about calling ranges for remaining players is crucial to the model and if players are widening their calling ranges because you're pushing too much that will affect the outcome, but ICM lets you make that adjustment. The fact is, though, if you are following ICM, your pushes are dictated by a number of factors and you will be pusing less the farther from the SB you get. And/or you will eventually be called and either double up or bust out. ICM is not just modeling your fold equity, it is modeling your total equity so even when you're called it isn't a bad thing. You will be ahead and win sometimes and you will be behind and suck out sometimes and double up. In any case your real-prizemoney-equity will go up. You do want a bigger share of the prizemoney, right?

You can certainly question the assumptions or the model itself, but I don't think people use it as a crutch or lean on it too heavily. In fact, I think just the opposite. Not enough SnG players get as much out of it as they could. (In fact, most SnG players don't use it at all...)

[ZeroPointMachine]

I'd like to thank everybody for the great input.

I'm not afraid of being wrong or admitting it. I get learned good that way.

Today I got learned real good.

My argument for passing small +$EV situations for better ones later is completly false.

Other than special circumstances that ICM cannot calculate (i.e. blinds increasing or tiny stacks)the only argument for passing on marginally +$EV situations is the difficulty in making these calculations perfectly. A small change in calling ranges can change some plays from +$EV to -$EV. In these cases it can be argued that you can't be sure it is positive and there is nothing wrong with passing it up.

[ QUOTE ]
Don't play above your bankroll, and squeeze as much EV out of your play as possible.

[/ QUOTE ]Isn't it possible that you can expect to be given better EV situations later?

You and I play heads up. First hand I show you AQ of clubs and go all-in. You hold 6d6s. Do you call? Calling has positive expectation, but probably not nearly as much as merely getting me to sit down at the table with you.

Anyway, I think there is an opportunity cost that is not factored into EV calculations.

[Nicholasp27]

1) if u push many times in a row, u'll need to change villian's calling range, which will change the hands that make it +ev

2) .1% shouldn't be used, as there should be a more +ev point before the tourney is over



all tourney long, we make decisions every hand which have some + or - $ev attached to them (we can't calculate the +/- $ev of every hand in a tourney right now) and if u add all of those up, u hope it's a positive number...

well say it's the bubble and all of your moves have actually added up to -$0.40 If you push a +$0.10 hand and get called and bust, you leave the tourney having lost money...But if you forgoe this +$0.10 push, then you are now at -$0.50...if the next hand you get a +$0.60 push, then it's better that you pushed that rather than get impatient and push the +$0.10, as even if you bust, at least you end up better than with the +$0.10 push

we can't calculate the ev of every move you made during the tourney, so we can ignore all of the previous moves and just focus on maximizing $ev once we can calculate it (when <10bb and people fold or push to you)...go through tournaments in the software and add up your moves from all of the hands that can be calculated...in a game, use your judgement on whether u think that the total would be higher if u push your current hand and risk busting or if you wait for a higher ev hand




anyway, ICM is a model, but you have to use discretion in whether or not you think there will be a better place to push and whether or not you have villian's calling range correct...also, you may think ICM is underestimating the +ev of a push due to the fact that blinds go up next hand...lots of factors can make our application of ICM for a given hand wrong, even if the tool is accurate and the math correct

[Nicholasp27]

[ QUOTE ]
I some what agree with you on this, but I believe the OP is saying the IMC is not perfect and could leading you to -$EV by taking lots of very small $EV pushes.

[/ QUOTE ]

as long as you correctly compensate for the fact that you took those small edge pushes, then pushing every small $EV will yield a positive number

the problem comes from

a) not changing opp range accurately to reflect their new standards based on previous hands
b) sure it is a + number, but is it the most positive you could have had? you may not be maximizing your ev if you take every small edge


but remember, if u pass up a +.2 ev push, then you have to subtract that from your running total...i can fold 10 +.2 ev hands in a row and then get a +1.5 ev hand due to higher blinds and my now lowered stack (and maybe better hand as well) but that totals to -.5 total...so i should have taken some of those +.2 ev hands

so that's why +.5 is a pretty decent number...it's hard to get a hand +ev enough to overcome 5+ passes on >+.5 pushes

[AliasMrJones]

[ QUOTE ]
well say it's the bubble and all of your moves have actually added up to -$0.40 If you push a +$0.10 hand and get called and bust, you leave the tourney having lost money...But if you forgoe this +$0.10 push, then you are now at -$0.50...if the next hand you get a +$0.60 push, then it's better that you pushed that rather than get impatient and push the +$0.10, as even if you bust, at least you end up better than with the +$0.10 push

[/ QUOTE ]

Again, I think you are confusing real-money EV with chip EV. What you say above might be true of tournament chip EV, but it can't possibly be true of real-money-prize-pool EV.

ICM (or at least SnGPT) takes into account your prize money equity if you fold, your prize money equity if you push and everyone folds, your prize money equity if you push, get called and win and your prize money equity if you push, get called and lose and weights each according to how often each of those outcomes is going to occur. If it says it is worth 2% more of the prize pool if you push than if you fold, having taken into account all of those things, then you gain 2% more of the prize pool if you push and you do not if you fold. That 2% doesn't come back. There is no I get more +$EV from a later decision so I'll forgo this one. ICM has already taken into account the situation if you fold and based on all of the outcomes/probabilities has determined that it is worth 2% more of the prize pool to push here than to fold. Period. End of story. The assumptions input could be wrong and the ICM model itself might have flaws, but assuming you believe the assumptions and the model are good, fold and try for a push later is already in the model and the model says it is worth more to push here than to fold and try for a push later.

Later on you might get an even more +$EV situation. But, that is an independent situation. Yes, if you had pushed earlier and lost you might be out and not able to take advantage of this opportunity, but, again, ICM/SnGPT in effect took this possibility into account by virtue of the ICM model and determined it was worth more real money to push at the earlier point.

Tournament-chip-EV is not equal to prize-pool-EV. The more +real-prize-pool-dollar-EV situations you take advantage of, the more real-prize-pool-money you will win in the long term. Period.

[Nicholasp27]

no, i'm not confusing the two...i'm talking purely prize pool, not chips...i'm just thinking about it differently

and read your last statement: " The more +real-prize-pool-dollar-EV situations you take advantage of, the more real-prize-pool-money you will win in the long term. Period."

exactly

if u pass up a +$0.20 push and then get a +$0.50 push and take that, then you followed your last statement and made more money...

and i know that .2 doesn't come back, which is why i subtracted it from my running total

if u push a .2% ev hand then the next hand may end up being 0% or -.1 or +.1 ev due to the looser calling range of villian, whereas it would have been +.5 if u had folded the .2% ev...

this isn't probability theory with independent trials...what u do one hand DOES affect the next...not the cards but villian's actions...

u can't just push every +.1% ev situation with impunity unless you can be >99.9% accurate in your assessment of villian's exact calling range based on previous hands (and not just prior hand, but whole tourney, how his other table he has open just tilted him or how his g/f coming over makes him a maniac so he can go get some, etc...u can NEVER predict opp's calling range with 99.9%+ accuracy, so therefore u may not want to take every +.1% ev push, as you aren't accurate enough to know that +.1 isn't a -.3

u also don't know the future so u don't know if the sum of folding this hand + pushing next hand > or < pushing this hand and pushing/folding next hand

i don't have sngpt on this comp, but it'd be pretty easy to come up with an example where you have a +.1 ev push and then show the next hand where if u had pushed that +.1, then it's <ev than if u had folded that previous +.1 hand

[The once and future king]

If you had an +.5%EV push but knew your next hand would be a +1.5%EV opportunity you would obvioulsy not take the first one due to the when you are broke you are done principle.

Early on in the tourney one may pass small +EV situations because one expects to have greater EV situations later and does not want to risk the opportunity to take those EV situations . Once we are on the bubble though suddenly we must take any EV situation however small?????

Obviously other factors are important before pushing on the bubble other than the mere existance of a +EV situation.

[HesseJam]

This is a very good discussion.

The way I see it:

The OP certainly has a point that push decisions cannot be viewed as independent coinflips. The EV of a push under a given assumption of calling ranges of your opponent is the EV from Eastbays tool + the difference between the EVs of future push decisions when you did push your hand before and when you didn't (good luck calculating this!).

Moreover, I also have the feeling that the 0.5% threshhold is set arbitrarily. One could ask why at all having a threshhold? The only justification I could see is that you should avoid a coinflip because you can do better than that.
How can you do better than a coinflip? Because you play better than your opponents and you should not go below your edge. So, I think that this should be the major variable for setting the threshhold: If you are in a game where your opponents make many mistakes you should take your time and wait for the better opportunities, thus the threshhold should be higher (but how much?). If you feel your competition as just as good as you (or even better) you could (should) take the coinflip.

Raptor is right that playing a certain hand right culminates in getting right the calling ranges of your opponents. But he is only almost there. You have to factor in what your decision will mean to the future calling ranges and how this will affect your EV on your future weighted average decisions. If opponents loosen up, your average EV will very often go down which means that if you stay with the threshhold you will have less situations where you can push for profit.

[AliasMrJones]

[ QUOTE ]
no, i'm not confusing the two...i'm talking purely prize pool, not chips...i'm just thinking about it differently

and read your last statement: " The more +real-prize-pool-dollar-EV situations you take advantage of, the more real-prize-pool-money you will win in the long term. Period."

exactly

if u pass up a +$0.20 push and then get a +$0.50 push and take that, then you followed your last statement and made more money...

and i know that .2 doesn't come back, which is why i subtracted it from my running total

if u push a .2% ev hand then the next hand may end up being 0% or -.1 or +.1 ev due to the looser calling range of villian, whereas it would have been +.5 if u had folded the .2% ev...

[/ QUOTE ]

But, here's the problem with the above...you're assuming you're going to GET that higher +$EV situation later. SnGPT has ALREADY taken that into account with the ICM model and determined that it is worth more prize money equity to take this opportunity now than to fold and hopefully get a better opportunity later. That is the entire basis for the ICM model. What is my prize money equity based on the current stack sizes. You might get a more +$EV situation later...you might not...you might get a more +$EV situation later and lose. ICM models these various chances and comes up with prize money equity.

If you push 10 hands in a row the ICM model doesn't change. What does change is the calling range of your opponents. That is a different issue. One hand affects the actions of your opponents, but not the model. You can change the input parameters to account for this. In that sense you are right about your actions now affecting future outcomes. A push now could widen opponents calling range.

But, that is a separate matter. The idea that you can pass up a smaller +$EV situation now so that you will survive to take advantage of a larger +$EV situation later is already accounted for as part of the ICM model and is therefore already factored into the current +$EV differential between pushing and folding. So, the notion that you can do that is illogical.

The model doesn't account for player skillz, though. It assumes each player will get equal +$EV opportunities over time, which is true. But, your opponents may not take equal advantage of these opportunities. If your opponents, do not take advantage of their +$EV opportunities, I think it would make sense to use a higher threshold. SnGPT already uses +.05 as the threshold. It might be worth discussing what threshold is appropriate for opponents who play incorrectly, like often limping or raising 3xBB with <10xBB, for instance.