Example of buble situation where ICM calcs may be inadequate?

[sergsz]

Situation from party 109:

Blinds 100/200

Hero ~ 1000
Button ~5000
SB ~ 2000
BB ~2000

Does it make sense for the hero to push with a wider range of hands than ICM calcs would suggest to avoid a situation (after the blinds hit him) where hero's stack would be considerably smaller than the other stacks, reducing hero's FE precipitously and probably causing the mid-stacks to tighten up and "wait out" the hero?

In the actual hand, I pushed Q8o, and lost to the big stacks's 99.

ANy thoughts on the general scenario and the particular hand?

[pergesu]

This is a hand where it depends entirely on your opponents' calling range. With the average range defined in SNGPT, this is a clear push. If they're maniacs, it's a clear fold.

ICM says this is a push in most cases. I think you just unfortunately ran into a big hand, so now you're looking to see if that was wrong. It wasn't. You play poker goot.

[gasgod]

I haven't done ICM, and I don't have SGA, but just considering stack sizes, this looks like a poor push. Big stack can call with a wide range of hands, because losing won't severely affect his tourney equity. Your stack isn't small enough to make it attractive to him to keep you alive. And he knows the blinds will sit this one out, so there's no risk behind him.

Big stack can't lose a lot of equity, but both mid stacks stand to gain a ton. That means you must be the one losing it. It's a zero sum game.

GG

[DasLeben]

[ QUOTE ]
This is a hand where it depends entirely on your opponents' calling range. With the average range defined in SNGPT, this is a clear push. If they're maniacs, it's a clear fold.

ICM says this is a push in most cases. I think you just unfortunately ran into a big hand, so now you're looking to see if that was wrong. It wasn't. You play poker goot.

[/ QUOTE ]

I agree. With 5xBB, this is a pretty clear push unless they're going to call you with trash. In that case, gasgod is clearly correct in folding. *shrugs*

[gasgod]

I would argue for a wide calling range for big stack. Folding gives him 37%, Calling and losing gives him 33%, while calling and winning gives him 41.8%. So, he has more to gain than he has to lose. I'd be interested in what SGA gives for a calling range of best 55% of hands (i.e. down to 45%), which is close to his break even point.

GG

[eastbay]

[ QUOTE ]
I would argue for a wide calling range for big stack. Folding gives him 37%, Calling and losing gives him 33%, while calling and winning gives him 41.8%. So, he has more to gain than he has to lose. I'd be interested in what SGA gives for a calling range of best 55% of hands (i.e. down to 45%), which is close to his break even point.

GG

[/ QUOTE ]

I think you're making a mistake here. Button has to win 45% of the time with his hand against your push range. This is not the same as saying he should call with 55% of his hands. He has to put you on a range, and then figure which hands of his are 45% winners against it.

If he puts you on any two, 63% of his hands are better than 45% winners (22+,A2+,K2+,Q2+,J4o+,J2s+,T6o+,T4s+,97o+,96s+,86s +).

If he puts you on a loose range like 22+,A2+,K2+,Q2s+,Q6o+,J7s+,J9o+,T8s+,98s, then "only" 28% of his hands are good enough to win 45% or better: 22+,A2+,K9o+,K6s+,QJo,QTs+.

Using the second estimate, and going back to the original question, putting the other two guys on a tight range of 66+,ATs+,AJo+, your +$EV hands are:

22+,A2+,K9o+,K5s+,QTo+,Q8s+,JTo,J8s+,T8s+,97s+,87s

Q8o doesn't make the cut, but it doesn't miss by a mile either.

All of this assumes a lot about button, though. You might do better to just stick him with a common sense read, although I generally like the second estimate pretty well for a typical $55.

eastbay

[eastbay]

You can account for the UTG effect by discounting your equity against which to compare your push.

Maybe something like the average of what you've got before and after posting your next blind is a reasonable way to do it.

eastbay

[microbet]

Nice point. How about discounting your equity a little something everytime you get further out of position?

I don't gamble with my children, but I'd almost be willing to bet my first born you've thought of this.

[eastbay]

[ QUOTE ]
Nice point. How about discounting your equity a little something everytime you get further out of position?

I don't gamble with my children, but I'd almost be willing to bet my first born you've thought of this.

[/ QUOTE ]

Your children are indeed safe. Look for it in future SGA releases as an option for $EV valuations. The first implementation is a linear discounting from 0 on the button to 50/50 from UTG (average of pre/post blind $EV).


It occurs to me that this probably should be applied to all players, not just you. I have a little more thinking to do about this still.

eastbay

[curtains]

[ QUOTE ]
Situation from party 109:

Blinds 100/200

Hero ~ 1000
Button ~5000
SB ~ 2000
BB ~2000

Does it make sense for the hero to push with a wider range of hands than ICM calcs would suggest to avoid a situation (after the blinds hit him) where hero's stack would be considerably smaller than the other stacks, reducing hero's FE precipitously and probably causing the mid-stacks to tighten up and "wait out" the hero?

In the actual hand, I pushed Q8o, and lost to the big stacks's 99.

ANy thoughts on the general scenario and the particular hand?

[/ QUOTE ]

One thing for sure is that Im pushing Q8o here.

[gasgod]

[ QUOTE ]
[ QUOTE ]
I would argue for a wide calling range for big stack. Folding gives him 37%, Calling and losing gives him 33%, while calling and winning gives him 41.8%. So, he has more to gain than he has to lose. I'd be interested in what SGA gives for a calling range of best 55% of hands (i.e. down to 45%), which is close to his break even point.

GG

[/ QUOTE ]

I think you're making a mistake here. Button has to win 45% of the time with his hand against your push range. This is not the same as saying he should call with 55% of his hands. He has to put you on a range, and then figure which hands of his are 45% winners against it.

[/ QUOTE ]

Yes, I realized that when I went to bed, but I hoped it would make little difference. /images/graemlins/smile.gif

[ QUOTE ]
If he puts you on any two, 63% of his hands are better than 45% winners (22+,A2+,K2+,Q2+,J4o+,J2s+,T6o+,T4s+,97o+,96s+,86s +).

If he puts you on a loose range like 22+,A2+,K2+,Q2s+,Q6o+,J7s+,J9o+,T8s+,98s, then "only" 28% of his hands are good enough to win 45% or better: 22+,A2+,K9o+,K6s+,QJo,QTs+.

Using the second estimate, and going back to the original question, putting the other two guys on a tight range of 66+,ATs+,AJo+, your +$EV hands are:

22+,A2+,K9o+,K5s+,QTo+,Q8s+,JTo,J8s+,T8s+,97s+,87s

Q8o doesn't make the cut, but it doesn't miss by a mile either.

All of this assumes a lot about button, though. You might do better to just stick him with a common sense read, although I generally like the second estimate pretty well for a typical $55.

eastbay

[/ QUOTE ]

Excellent analysis, Thanks.

GG

[Bataglin]

I think you did what you had to do. You can't win'em all.