Trickier Sit N Go Question

[leaponthis]

[ QUOTE ]
when you are one of two short stacks on the bubble, it is more important to prevent the other player winning pots than it is to win them yourself.


[/ QUOTE ]

Now you see, this is something that one can bring to the table that is more portable than the ICM calculator. I like this, although, I believe, that I need to think about it a bit. It would also be nice to get Sklansky's opinion. But...

leaponthis

[blackize]

[ QUOTE ]
Now you see, this is something that one can bring to the table that is more portable than the ICM calculator. I like this, although, I believe, that I need to think about it a bit. It would also be nice to get Sklansky's opinion. But...


[/ QUOTE ]

Those of us who have played a ton of SNGs seriously, we have memorized a lot of common ICM scenarios and we can pretty easily extrapolate others.

[leaponthis]

[ QUOTE ]
In other words, he and the STT forum are in complete agreement about that.


[/ QUOTE ]

Well maybe if you had a concept to think about you would not have had to do analysis on a "Ton" of sng's. When I first learned how to play poker I picked up Seven Card Stud For Advanced Players by Sklansky and Malmuth. That book contains a detailed explanation for almost every poker tactic one needs to be a competent winning poker player. From that book for example I learned when and why to check raise. It saved me a lot of analysis and pain.

leaponthis

[blackize]

SNGs are too counterintuitive to have something that simple you can take to the table.

Someone goes all in preflop in a cash game and you have Aces. You call every time because you cannot possibly be beat and you cannot possibly be a larger favorite preflop. In a SNG you still cannot possibly be beat and you cannot possibly be a larger favorite preflop, but it is still wrong to call in some situations.

[citanul]

leap,

you're simply entirely wrong (when you say that maybe if we reduced things to "concepts" we wouldn't have to do tons of analysis), and i'll offer a very quick, somewhat hand-waving argument/example of why.

late game play in stts is by and large a math problem.

each time you make a play, you ask yourself:

"is the value of the chips that i expect to have when i follow through with this play greater than the value of the chips that i could expect to have from any other alternative?"

in many situations where you must ask yourself whether or not to move all in, for example. when you look at this, you look at

Expected value of the stack if we fold,

and

Expected value of our stack if we move all in

assuming we are covered, this is:

(probability that everyone folds)(value of stack if they all fold)+
(probability of getting called)(probability of winning if called)(expected value of the stack we are left with if we are called and win)

when we look at the example in this thread, the most important thing to consider is that by leaving ourselves the chance to fold if we need to, we have preserved the value trapped in those remaining chips in situations where we would most like to have them.

anyway, back to my main example:

pretty obviously, even if we can perfectly estimate how our opponents call (and that is not a simple task), knowing both the probability of winning if called by a range, and the expected value of our stack when we fold, and the expected value of the stack we get by stealing the blinds, the expected value of the stack we get by doubling up, etc, are all things that really, you can't know unless you've done your homework.

if one were to try to make a very simplified example of a sng hand, it would look like the standard Chubukov-Sklansky question. Certainly it isn't easilly suggested that by understanding a "concept" one would be able to come up with the Chubukov-Sklansky numbers, or even a reasonable approximation, if one needed them on the fly while playing poker at a table. again, the C-S numbers are about the most simplified you can make a stt late-game question.

to be more precise for your understanding, there are plenty of situations one could quickly draw up where it would be correct to move all in with K9o but not K8o, or K9s but not K9o. This is not the sort of thing that a "concept" would help with. shortly, late game stt play is almost entirely "brute force mathematics."

i hope that clarifies a bit.

citanul

[leaponthis]

[ QUOTE ]
to be more precise for your understanding, there are plenty of situations one could quickly draw up where it would be correct to move all in with K9o but not K8o, or K9s but not K9o.

[/ QUOTE ]

Since you are being precise please define "plenty of situations". I ask mainly because I do not buy the "plenty of situations". If, in fact, there were plenty of situations that you descrbe, then it would follow that there would be a general concept for them. I also do not buy that the individual needs to do tons of analysis in lieu of using a proven general concept to accomplish close to if not the same results.

leaponthis

[blackize]

The best concept I can think of for STTs has already been coined by Dan Harrington. He called it "first in vigorish" and used it to illustrate that you should push looser than you should call.

Taking it a step further to determine exactly what is profitable against a range requires ICM examination.

[citanul]

[ QUOTE ]
Since you are being precise please define "plenty of situations". I ask mainly because I do not buy the "plenty of situations". If, in fact, there were plenty of situations that you descrbe, then it would follow that there would be a general concept for them. I also do not buy that the individual needs to do tons of analysis in lieu of using a proven general concept to accomplish close to if not the same results.

leaponthis

[/ QUOTE ]

leap,

without too much work, one could generate nearly arbitrarilly many such situations. the point of my post, which i thought was quite explicit, is that because they are so mathematically precise "concept" loses out to "reality." while the concept of "first in vigorish" is great, there's often a cutoff on what hands you can profitably move all in with.

take a look at this term (not perfectly quoted i'm sure since i'm not quoting it) from my last post:

When you move in one thing that is very important to you is this:

(probability that you get called)(probability that you win when you are called)(number of chips that you will have if you get called and win)

not only are some of the hands one should move all in with counter-intuitive to the normal person (moving all in with T9s is frequently better than moving in with A2o), but when you grind through the actual mathematics, say you know perfectly what your probability of getting called is, and what hands they are going to do it with, then what you need is to construct a set of hands which wins the proper percentage of the time against those hands, and fold the rest.

while this produces "quite a few" move all in with any two cards" situations, obviously (since K9o is just a better hand than K8o, and K9s is just a better hand than K9o, by many different metrics, but here explicitly by the Chubukov-Sklansky numbers) there are going to be times where the better hand is above the cutoff and the worse hand is below it. one can construct such situatoins ad nauseum.

just because something happens a lot doesn't mean you can make a generalization. in particular, it would be a pretty pointless generalization to make to say for each pair of hands next to each other in the chubukov-sklansky rankings when the probability works out like yadda yadda, we play the top hand and don't play the next one.

instead, the more genral rule is effectively the rule of first in vigorish, which was mentioned by a poster above, somewhat explained by harrington, and in pretty thorough detail i believe explained by me above.

the rule, or generalization, one more time, for you: (note, there will be exceptions to this generalization, as there are for all good ones, and honestly, this too is simplified)

when in a late game sit and go tournament situation where your decision is binarily moving all in or folding, the relevant consideration is:

(1) If you choose to fold:

What will be the value of your stack if you fold, where value is calculated by a model such as the independent chip model, which assigns prize pool equity based on stack size?

and

(2) If you choose to move all in:

What is the expected value of your stack if you move all in, where the value is calculated using the same models.

The expected size of your stack will be:

(Probability that everyone else folds)x(Your starting stack + blinds + antes) + (Probability that you get called)(Probability that you win when called)(Stack size if you get called and win)

If you look at the expected value of each, and figure out which one gives you the better expected value, you do that one.

In general, it is correct to move in far, far more than most people think, because the probability that you get called is frequently quite low, and in holdem the odds that you are a huge dog are low.

This is my last long post in this thread. Vince, I wish you could understand why I'm flustered here. Some things are just not "generalizable." You can write up nice statements like the one in the preceding paragraph, but really when you're playing at a real poker game against people who have not only done that, but sat down and played with the simulators enough to actually know which hands perform well in which situations, and which situations are actually "push any two" spots, the theorist will get taken apart, and quickly. The more familiar and precise with their knowledge of this material your opponents are, the worse you will fare.

citanul

[leaponthis]

Listen I agree that there is no substitue for the precise mathematical answer. I agree with that. But you create your own problem when you demand precision in situations such as the one Sklansky describes. My gosh your own statement should tell you that. K,9 you play one way and K,8 you play another. Do you truly believe that you will be able to be precise enough during a poker game to determine the distinction between the two?

In Sklansky's example the best and accepted, I believe, option to proceed is to call. Agreed? Your ICM model shows it to be optimal. But what makes it optimal is that you are trying to increase your chance of winning bt ganging up on the BB and giving yourself an option to continue should an unfavorable situation arise. (SB raises and BB folds).

So rather than make a statement of which I am not sure i will ask a question? In Sklansky's scenario what two cards do you fold according to the ICM?

If the answer is none then I do not agree with your arguement.

leaponthis

[citanul]

[ QUOTE ]
So rather than make a statement of which I am not sure i will ask a question? In Sklansky's scenario what two cards do you fold according to the ICM?

If the answer is none then I do not agree with your arguement.

leaponthis

[/ QUOTE ]

leap,

[quote[Four good players left. Two have large stacks. 300-600 blinds. Big blind has 100 behind him (700 total). UTG folds. You are on button with 1100 and have KJ offsuit. Call, raise or fold?

[/ QUOTE ]

here's how poor the problem is for generalization:

depending on the definition of "large stacks," and in fact which stack is larger, the button or utg, SOME good players if you fold any hand on the button will fold almost every hand to the bb. they will do this so they can rob the other big stack on the next hand, and potentially for a long time. that is, they keep the bubble alive so that they can exploit it, for free chips. hell, some people will even just limp the sb and dump to the sb if they miss and things like that, depending on the stack sizes.

in sklansky's question as well, obviouly it makes no sense to limp an incredibly poor hand in many scenarios. as i said above, often the sb has no really good incentive to try to knock out the 4th place player, so he could easilly fold the sb, leaving you pot commited to crippling yourself with a poor hand.

so basically what i'm saying is that while there may or may not be hands good enough to make moving all in actually superior to just calling, expectations wise, with the exact stack you have in this loose scenario, there are almost certainly hands bad enough you should fold them.

continuing, depending on how big "big" is for the other playerse, and their style of play, etc, if one were to change your 1100 to some other number, all these situations change drastically. this should be obvious as well. if you can make it through the blinds next orbit easy, if you can't, if you have tons of chips, etc.

i hope this sheds more light on the non generalizability of such spots.

c