Beyond Sklansky-Chubukov
 

In his book NLHTAP, David Sklansky asks (and aswers) the question of what is the biggest stack we can profitably go all in with if we accidentally flip our cards over when being on the small blind in heads-up situation.

I played around with this intriguing idea just to come up with another very similiar problem that I believe to be much more realistic.

Instead of flipping over our cards, what if we just told our opponent (before looking at our cards ourselves) the range of hands we are going to go all-in with? What would then be the all-in range that maximized the expected value (for some fixed stack size)?

So, instead of knowing our exact cards, our opponent would only know the range of hands we are moving in with. All other factors would be identical to the problem presented by Sklansky.

This would be more realistic because against observant opponents they may very well have a good idea of the range of hands we are playing with, but unless we actually do flip our cards over they are still not going to know our exact holdings in any particular hand (assuming of course we don't have any obvious betting patterns or other tells that would reveal our cards).

For the sake of simplicity, I suggest the range to be defined in the order of the Sklansky-Chubukov rankings.

Has anyone tried to solve this kind of problem before?

If so, I would like to compare the results I got from my calculations and make sure I didn't make any mistakes. And if not, I'd like to hear what others think of the idea and if it would be worth to study it further and publish some of the results I got.

Re: Beyond Sklansky-Chubukov
 

Interesting... I'll crunch a few numbers.

Re: Beyond Sklansky-Chubukov
 

I was just searching the archives for S-C stuff yesterday, and came across something like this that was proposed. I'll look for a link.

~MagicMan

Re: Beyond Sklansky-Chubukov
 

Unless I am not completely understanding your idea, wouldnt this be covered under calling all-in raises in the previous chapter? So if you tell villain what your range is preflop for going all-in and than stick to your word after looking at your card, villian simply has to compare your range to the pot odds for that range to determine if calling is profitable or not.

Re: Beyond Sklansky-Chubukov
 

You are absolutely right that that is how I assumed the opponent would be playing.

However, the question I was asking was what should our all-in range be if we wanted to maximize our EV (assuming our opponent plays perfectly like you described).

Re: Beyond Sklansky-Chubukov
 

Who is Chubokov?

Re: Beyond Sklansky-Chubukov
 

I've got a spreadsheet where I can type in the hand range of either the pusher or the caller, stack sizes, and blinds. It will tell me which hands have a +EV for calling (against the pusher's range) or pushing (against the callers range).

Re: Beyond Sklansky-Chubukov
 

bump, has this been done already?

Re: Beyond Sklansky-Chubukov
 

AKA Karlson on 2+2.

Re: Beyond Sklansky-Chubukov
 

The SC ratings are flawed unless you are playing against a random hand. You should look up the rating of your hand against his calling range, which is what you are trying to do. This will determine your range to go allin with.

This is why a hand like A9o does well in SC rating, even though it is dominated most of the time. 45s is a better allin hand imo while it does no do very well in SC ratings. This while A9o has a rating of 81.7 45s had only 4.85.

Against AA-JJ, AK-AJ, KQs... 45s has a 33.5% equity and A9o 27.5%... See how this does not correlate to the SC numnbers? The part after the explanation tells you when to use the ratings and when not to.

GL

Re: Beyond Sklansky-Chubukov
 

[ QUOTE ]
The SC ratings are flawed unless you are playing against a random hand. You should look up the rating of your hand against his calling range, which is what you are trying to do. This will determine your range to go allin with.

[/ QUOTE ]

They're not flawed, but you can exploit your opponent better if he is not playing perfectly. Figure out what his range is, and apply what you know correctly. Problem here is playing "perfectly" though, at least as asked:

[ QUOTE ]
However, the question I was asking was what should our all-in range be if we wanted to maximize our EV (assuming our opponent plays perfectly like you described).

[/ QUOTE ]

... I take it we're talking typical short stack jam/fold strategy only.

The equity issues above are balanced out by the frequency which an opposing player is able to correctly call a given hand.

[ QUOTE ]
This is why a hand like A9o does well in SC rating, even though it is dominated most of the time. 45s is a better allin hand imo while it does no do very well in SC ratings. This while A9o has a rating of 81.7 45s had only 4.85.

[/ QUOTE ]

A9 is a vastly, vastly better pushing hand than 54s, because almost any hand is correct to call 54s - it doesn't require much money in the pot for all hands to correctly call. The problem here is that you're putting a random hand on a range. If you're jamming from the SB you'd be treating his holding as a random hand anyway. If you saw a card or some kind of tell where you could guess a range like AA-KK, AK-AJ, KQs, then you already know you can do better than just working by the SCs. Otherwise, you could just as well see how A9o and 54s do against hands like 95, 55, 77, and seeing how much better A9o does.

Obligatory mention of SNG PT goes here. IF you're in this SS jam/fold scenario and can put ranges on opponent, mess with SNG PT.

Re: Beyond Sklansky-Chubukov
 

Maybe the word flawed was wrong, I'm not a native speaker. I mean they are not very suitable to use in many cases.

Here is why, and it also answers why I can put my opponent on a range.

I'm shortstack and he isn't, but I can still hurt him enough to call with descent hands. Therefore he will not call me with a random hand, he will choose to call me with a certain range of hands so it is correct to put him on a range. If he folds it does not matter what you hold, only if he calls. I think this range he calls with does (relatively) better against a weak ace, compared to a suited connector, since a weak ace will be very likely dominated while a suited connector is not. Same goes for weak kings or many not so great broadway hands.

In other words; the SC numbers are correct to use against a random hand, but they ignore that he is only going to call with a certain range and fold the rest. Therefore it is better to push with hands that do better against his plausible calling range than with hands that have good SC numbers. Many times these hands will be quite similar, but easily dominated hands, small pocket pairs and suited connectors will differ imo.

GL

PS: SNG PT is part of poker tracker? I'm using poker office, I might switch because I miss some features I'd like in PO...

Re: Beyond Sklansky-Chubukov
 

[ QUOTE ]
In other words; the SC numbers are correct to use against a random hand, but they ignore that he is only going to call with a certain range and fold the rest.

[/ QUOTE ]
That's all true - but with SC's, his calling range doesn't make a difference at all. You can correctly jam any hand if the effective stack is less than the SC number. Even if you are woefully behind when you've been (correctly) called with an easily-dominated hand such as A3o, it will be outweighed by all of the times your opponent cannot (correctly) call. He just won't get the hands he needs to call often enough, at least when talking long term probability.

SNG PT = Sit N' Go Power Tools, it's a stand-alone program, and it's perfect for what you're talking about. If you can put your opponent on a calling range, say you know his endgame tendencies, you can plug them in the calculator. http://sitngo-analyzer.com/feedback.html, you can use a limited version for free, but full version is worth every penny.

Re: Beyond Sklansky-Chubukov
 

[ QUOTE ]
In other words; the SC numbers are correct to use against a random hand, but they ignore that he is only going to call with a certain range and fold the rest.

[/ QUOTE ]
They assume your opponent has perfect knowledge, and sees your cards, so will only make +ve EV calls!

It is opposite of random hand. It tells you something but not what you want exactly. It says, Push In with A9 is more profitable than folding, even if your opponent knows you have A9 not AK or AA..QQ.

So how can you do better by removing hands, based on a presumed calling range, that's adapted to an imperfect basket of hands?

Remember, S-C number only applies to SB v BB pushes, it says to divide for BSB play, open raising into 2 blinds.

If you're open raising on a steal from mid pos or cutoff in a tourney, then your risk running into many more Aces, that ppl call with, than SB v BB. So the general point, that fold equity play benefits from being a few % better, if you're called by big Aces and big pairs is true.

Re: Beyond Sklansky-Chubukov
 

The problem with SC numbers is that they do not formulate a complete strategy that also includes bluffing.

Re: Beyond Sklansky-Chubukov
 

Gonso: Thanks for pointing out SNG PT! I'll have a look. I see how you can play a game theoretical perfect game with SC numbers, but it is very easy to figure out a better strategy by putting him on a calling range.

RobNotss: I see how you can put the big blind on a random hand and therefore assume a random hand. But still, when he calls your hand does matter when he folds it doesn't. That range is different. If perfect information for your opponent is assumed, yes a game theoretical optimal strategy would be best. But same as above; luckilly we can use better than optimal strategies, because our opponents are not seeing our cards and therefore making mistakes.

Re: Beyond Sklansky-Chubukov
 

The opponent actually looks at your cards, before deciding what action to take! And it is still correct to move in pre-flop, so your hand does matter, and his perfect play, still offers you a profit.

How can you do better, by removing hands from S-C, when you have all the advantages SB v BB of BB not seeing your cards and making mistakes?