Sklansky-Chubukov numbers attacked
 

In Full Tilt's tournament guide, Andy Bloch has a chart (which he says Ferguson also came up with simultaneously and independently) showing what stack sizes should push or fold / call push in a SB v BB game. See p 98-99.

It is the same chart Chen and Ankenman published in Mathematics of Poker. See Jam or Fold tables at p. 136

After conversion, the numbers are different than the S-C numbers. (The Bloch/Chen numbers are expressed in Big Blinds, the S-C in small blinds minus 1 big blind). To convert the numbers, you either double Bloch's and subtract 1, or add 1 to S-C and halve it.

Bloch says his table is more realistic than the S-C numbers. In the S-C game, the BB knows the SB's hand. In Bloch's game the BB does not have perfect information. Bloch says this is more like real poker. As a result, Bloch says suited connected semi-bluffing cards are devalued in the S-C perfect information game.

Has this been discussed? Has David commented on Bloch's game or responded to Bloch's commentary?

Re: Sklansky-Chubukov numbers attacked
 

If you actually read DS's book, you'll find that he explains the S-C numbers are only a guide, which you should adapt to your particular opponent and situation. He also gives a basic idea of how to adapt them to real life scenarios.

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]
If you actually read DS's book, you'll find that he explains the S-C numbers are only a guide, which you should adapt to your particular opponent and situation. He also gives a basic idea of how to adapt them to real life scenarios.

[/ QUOTE ]

You sure do make a lot of assumptions. Of course I have read Sklansky's book. Who doesn't know to adjust to your foes. Who doesn't know push-fold strategy is optimal only at stack sizes of 10-13 BB or less.

My question had nothing to do with the PRACTICAL application of either S-C or Bloch's numbers, which is all your reply addressed.

This is a poker THEORY board. Sklansky created a SB-BB jam or fold game (and Chubokov ran the numbers). Bloch, Ferguson, Chen and Ankenman created a different SB-BB jam or fold game, which Bloch contends is more realistic than Sklansky's game.

My question is whether Sklansky has addressed the criticism of the contstruct of his game.

Re: Sklansky-Chubukov numbers attacked
 

Sklansky states in his book that the S-C number for the "robust" hands (like suited connectors) should be higher because a hand like J2o probably won't call when you shove with 87s. He doesn't change his chart to account for this, but his chart wasn't designed to inform you when to push with hand x. It just shows that you can't lose money when your stack size is =< the S-C numbers and your opponent only calls a push when he's ahead.

Re: Sklansky-Chubukov numbers attacked
 

IIRC, the SC numbers were only supposed to be a rough guide anyway, and more specifically, DS mentioned that a lot of the numbers are effectively underestimated in regards to standard practice. The perfect information bit is kind of rediculous anyway.

Using SNGPT-type approaches (more exploitative push/fold strategies like those mentioned) are generally more effective.

There were a few SC number threads trying to address the SC problems, including one of mine a ways back, and there was never really a good defense. I don't remember DS addressing the issue offhand at any point.

Re: Sklansky-Chubukov numbers attacked
 

The correct way to interpret the SC numbers is that if they say a push is correct in the perfect information game, then folding is dominated by moving all in in the real game, but if they say folding is correct in the perfect information game, then that says nothing about the real game. The numbers therefore do not suggest an actual strategy, and so comparing them with jam or fold unexploitable strategies is comparing apples and oranges.

Re: Sklansky-Chubukov numbers attacked
 

Yes, SC numbers are pretty much a threshold for 'when not to fold', and that's about it.

Re: Sklansky-Chubukov numbers attacked
 

One of the advantages of SC-type numbers is that you can calculate them without necessarily solving the game. For example, I think it's not feasible to solve push-or-fold Kansas City Lowball (too many different hand matchups) but you can still calculate S-C analogues for individual hands.

It just happens that it is feasible to solve push-or-fold NLHE, so the S-C numbers aren't as useful.

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]
Has this been discussed? Has David commented on Bloch's game or responded to Bloch's commentary?

[/ QUOTE ]

It is rare to see David comment on other authors, but I assume that Mason has already hired a hitman to get even with Bloch.

On the more serious side, I'd like to know if A4s is rated correctly in Bloch's tables. It looks like a serious typo to me, especially if you compare the number with A5s and A3s. The problem with such typos is twofold. Finding them is one part, figuring out the right numbers is the other.

Re: Sklansky-Chubukov numbers attacked
 

A4s seems to be correct in the MOP tables, at least i cant spot anything out of line.

Push: A5s-A3s: >50
Call A5s: 30.1, A4s: 25.6,A3s: 24.7

What is A4s value in Blochs table?

Re: Sklansky-Chubukov numbers attacked
 

In terms of SC it would be 857 and it is ranked just behind K-K and ahead of Q-Q! It goes {AA, AKs, AKo, KK, A4s, A5s, QQ...}. I understand that the reason for this is card removal and some sort of bluffing equity, but I am still very surprised.

Btw, Chen/Ankenman mention A5s in their on Jam/Fold also, so it seems that these hands do indeed rank very high.

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]
A4s seems to be correct in the MOP tables, at least i cant spot anything out of line.

Push: A5s-A3s: >50
Call A5s: 30.1, A4s: 25.6,A3s: 24.7

What is A4s value in Blochs table?

[/ QUOTE ]

I wonder how those decimals were computed. For example, call with A4s has no exact treshold, the strategy is mixed between stacksizes 25.3 and 25.69 and is equal [0.1646 call; 0.8354 fold] for stack size 25.6. Why this very value was selected for the table?
http://www.pokerbolide.com/files/images/a4scall.png

Andrzej Nironen

Re: Sklansky-Chubukov numbers attacked
 

Hm, good question.

As its rounded to the last digit, the actual call-% used as threshold may be a bit larger/smaller than 16.5% tho. Maybe they simply picked some arbitrary limit like 20% or 25%?

Its a surprising choice in any case, it would make more sense to pick a high-% for call, and a low-% for push - which would be slightly biased towards exploiting the "average" player. (Here, we call slightly looser than NE, which probably isnt a good idea vs. an average player.)

But i guess thats way beyond practical relevance. I doubt anyone would want to mix actions for intervals of +/-0.15. And i definitely dont memorize those values to the last digit - as it would be pretty useless to know the exact values, when i dont calculate the stacks-to-blind ratio that accurate anyway.

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]

As its rounded to the last digit, the actual call-% used as threshold may be a bit larger/smaller than 16.5% tho. Maybe they simply picked some arbitrary limit like 20% or 25%?


[/ QUOTE ]
Maybe.
<font class="small">Code:</font><hr /><pre>
Stack Call
25.55 0.3035
25.65 0.0581
</pre><hr />

[ QUOTE ]

But i guess thats way beyond practical relevance.


[/ QUOTE ]
Of course, but we speak about theory here [img]/images/graemlins/smile.gif[/img]

Andrzej Nironen

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]
Btw, Chen/Ankenman mention A5s in their on Jam/Fold also, so it seems that these hands do indeed rank very high.

[/ QUOTE ]
I have read and re-read that section dozens of times and I still haven't figured out why A5 (as opposed to Ax with x&gt;5) is in there.

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]
[ QUOTE ]
Btw, Chen/Ankenman mention A5s in their on Jam/Fold also, so it seems that these hands do indeed rank very high.

[/ QUOTE ]
I have read and re-read that section dozens of times and I still haven't figured out why A5 (as opposed to Ax with x&gt;5) is in there.

[/ QUOTE ]

Of all the Axs, A5s has the second-best equity (after ATs) against AA.

Re: Sklansky-Chubukov numbers attacked
 

I understand that it has to be an A because of card removal. I believe you, when you write that A5s has the second best equity (and I could run a simulation if didn't believe you).

What I don't understand is why. What is the reason I would get A5s as superior to A6s or, say, AJs against AA if I ran a simulation?

I did notice that it was ten and five. You can't make any straights witout T and 5. Is it that by having one of those in our hand we remove some straight possibilities?

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]

What I don't understand is why. What is the reason I would get A5s as superior to A6s or, say, AJs against AA if I ran a simulation?

[/ QUOTE ]

A5s can make straights with 2 hole cards, and can make more 1-hole card straights than A4s-A2s. A5s and ATs have identical straight potential.

ATs/A5s is better against AA than AJ+s, because they have a higher potential to make straights by using the T/5 only.

Re: Sklansky-Chubukov numbers attacked
 

Well, the mystery has been cleared. I understand it now. [img]/images/graemlins/cool.gif[/img]

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]
Well, the mystery has been cleared. I understand it now. [img]/images/graemlins/cool.gif[/img]

[/ QUOTE ]

Cool. Now I am adding T5 to my list too since it can make every straight except 1.

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
Btw, Chen/Ankenman mention A5s in their on Jam/Fold also, so it seems that these hands do indeed rank very high.

[/ QUOTE ]
I have read and re-read that section dozens of times and I still haven't figured out why A5 (as opposed to Ax with x&gt;5) is in there.

[/ QUOTE ]

Of all the Axs, A5s has the second-best equity (after ATs) against AA.

[/ QUOTE ]

Thanks Jerrod, but this still doesn't explain why Andy seems to be ranking A4s ahead A5s (and both ahead of QQ in terms of jam/fold against a random hand).

Re: Sklansky-Chubukov numbers attacked
 

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
Btw, Chen/Ankenman mention A5s in their on Jam/Fold also, so it seems that these hands do indeed rank very high.

[/ QUOTE ]
I have read and re-read that section dozens of times and I still haven't figured out why A5 (as opposed to Ax with x&gt;5) is in there.

[/ QUOTE ]

Of all the Axs, A5s has the second-best equity (after ATs) against AA.

[/ QUOTE ]

Thanks Jerrod, but this still doesn't explain why Andy seems to be ranking A4s ahead A5s (and both ahead of QQ in terms of jam/fold against a random hand).

[/ QUOTE ]

Rankings aren't like linear; some hands pop in and drop out at lower stack sizes, or are mixed at different stack sizes, etc. When making "tables" you sort of have to just decide when the hand is "in" and when it is "out." I suspect that it's this process that causes Andy's rankings to be this way.