Dennis Bragg: Pushing/Defending In The Blinds: The Next Logical Step

[zkzkz]

So it seems like the next logical step given the past two articles is to take one of the situations and instead of assuming your opponent pushes/calls with xx% of the hands, assume your opponent will use the corresponding table.

Of course then you can rerun the analysis for the corresponding table based on your optimal response. And rerun the first side again with the improved opposite side.

I'm not sure but my hunch is that the two sides will converge to the same table regardless of your starting conditions. In which case you'll have a single set of tables for the "optimal response assuming your opponent plays optimally" Then you could have alternate tables for "assuming your opponents plays 10% tighter and 10% looser".

Your comments at the end of the small blind section noting that with a very small stack you should assume your opponent will (correctly) call with any two cards foreshadow this line of reasoning.

It's possible the two tables will converge to different results depending on the initial conditions. That would be an interesting thing to see. It's also possible they don't converge at all. But I suspect that's not the case.

It would be really interesting the results if your program could be extended to do this analysis.

[Vee Quiva]

I think the article is great for showing the math behind these decisions. Here is the criticism.....

A large portion of the discussion in the one table tournament section deals with bubble play. I feel that the advice would be very different on the bubble versus playing 5 or 6 handed.

Is this a topic for a future magazine or would the author or anyone else care to comment here?

[trojanrabbit]

I've already looked into optimal vs. optimal play and am currently writing an article of my own that comments not only on this, but on bubble play as well. I'll hopefully give some insight into just how much adjustment needs to be made on the bubble in different situations.

[wreckem]

That is an interesting point and I am not sure of the answer. I suspect that the two tables do not converge to the same table simply because the pusher has fold equity. But there probably is an optimal push strategy and a cooresponding optimal call strategy.

I am working on some other simulations right now, but when I get a chance I will try to investigate this some more.

Dennis

[wreckem]

[ QUOTE ]
I feel that the advice would be very different on the bubble versus playing 5 or 6 handed.

[/ QUOTE ]

I do not agree with this statement, but I would be interested to hear why you think the strategy would be different.

Dennis

[wreckem]

rabbit, I am looking forward to your article.

Dennis

[trojanrabbit]

[ QUOTE ]
rabbit, I am looking forward to your article.

[/ QUOTE ]

So am I... /images/graemlins/laugh.gif

I've submitted my first part which looks at how prize money and bubble situations can affect the EV decisions you have to make. It hasn't been accepted yet, but naturally I'm optimistic. As a second part I'm already working on optimal vs. optimal for SB v. BB (yes it usually does converge) both in the cash game as well as how the answers change in different tournament situations.

[Nottom]

[ QUOTE ]
[ QUOTE ]
I feel that the advice would be very different on the bubble versus playing 5 or 6 handed.

[/ QUOTE ]

I do not agree with this statement, but I would be interested to hear why you think the strategy would be different.

Dennis

[/ QUOTE ]

Because the $ value of chips becomes so warped in many bubble situations. You can easily find yourself in a position where your opponent has you covered doesn't look at his cards, pushes, and you would be incorrect to call with hands as good as AK.

[wreckem]

[ QUOTE ]
Because the $ value of chips becomes so warped in many bubble situations. You can easily find yourself in a position where your opponent has you covered doesn't look at his cards, pushes, and you would be incorrect to call with hands as good as AK.

[/ QUOTE ]

Yes, there are special cases where this is true, like when the pusher is the big stack and you are the second biggest stack. Or when another stack is so small he is about to be blinded out.

The tables only consider chip EV, not $EV. However, I still think they apply to most bubble situations.

[Nottom]

[ QUOTE ]
The tables only consider chip EV, not $EV. However, I still think they apply to most bubble situations.

[/ QUOTE ]

Sorry, but they don't. Even in a somewhat common spot where everyone has about the same stacks and the SB is pushing, the difference between ChipEV and $EV is usually pretty big.

For example, You are in a party SNG on the bubble. Everyone has 2000 chips and the blinds are 100/200.

Folded to the SB who pushes blind.
According to the chart you are correct to call with a bit over 50% of your hands (which would be absolutely correct in a HU match)

However a bubble situation warps the chip value so much. Winning the hand is worth about 38% of the prize pool, while losing it gives you nothing. At the start of the hand, your stack was worth exactly 25% of the pool (everyone was even) so you stand to gain only 13% while losing 25%.

As a result even if your opponent is pushing blind you still need a good hand to call, namely: 66+,ATo+,A7s+,KQo,KTs+ (about 12% of hands)

As an aside, if you opponent knows this then it actually becomes correct to push any two from the SB since the BB cannot call often enough to punish you for your loose pushes.

This just shows how different and counter intuitive proper bubble play can be. If the BB does decide to call then it hurts the SB, but it also hurts himself and the remaining players are the grateful recipients of all that lost $EV.

(Note: the STT forum is full of this sort of analysis)

[trojanrabbit]

[ QUOTE ]
As an aside, if you opponent knows this then it actually becomes correct to push any two from the SB since the BB cannot call often enough to punish you for your loose pushes.

This just shows how different and counter intuitive proper bubble play can be. If the BB does decide to call then it hurts the SB, but it also hurts himself and the remaining players are the grateful recipients of all that lost $EV.


[/ QUOTE ]

This is exactly what I'm going to cover in my second article (the first has been accepted but not published). I'll give a lot of examples and explain what's going on.

[wreckem]

Yes, I agree completely with what you are saying. I was focusing on your original example of folding AK. I think it is a rare bubble situation where folding AK to an all-in raise is correct.

[Scuba Chuck]

[ QUOTE ]
Yes, I agree completely with what you are saying. I was focusing on your original example of folding AK. I think it is a rare bubble situation where folding AK to an all-in raise is correct.

[/ QUOTE ]

You must not play very many SNGs.

[Scuba Chuck]

Rabbit, if you don't mind, could you cover your SNG experience?

Thanks in advance, Scuba

[trojanrabbit]

[ QUOTE ]
Rabbit, if you don't mind, could you cover your SNG experience?

[/ QUOTE ]

Only about 200 games, mostly 20+2. But I've tried to make most of my article based on mathematics and game theory rather than judgement. Here's an example of what you'll be getting in my article. Take your example where you have 4 people on the bubble, each with equal stacks. I look at not only the example where the stacks are 10BB, but at shallower stacks as well.

Pushing from the SB:
http://img176.imageshack.us/img176/9759/table31yw.jpg


Calling from the BB:
http://img176.imageshack.us/img176/4736/table45dy.jpg


The cards marked in white mean that you should push with a deeper stack, but not if it gets shallower than indicated. Otherwise it means push/call if the stack size is this much or smaller.

I'll also give other examples on the bubble besides everyone having equal stacks.

[Scuba Chuck]

[ QUOTE ]
--------------------------------------------------------------------------------

Rabbit, if you don't mind, could you cover your SNG experience?


--------------------------------------------------------------------------------



Only about 200 games, mostly 20+2. But I've tried to make most of my article based on mathematics and game theory rather than judgement.

[/ QUOTE ]

Can you comment more about your background in poker in general then? I am trying to find out more about where you are coming from.

Also, just to give my background. I've played about 4,000 SNGs just this year. About 80% $30+3s, and 20% $50+5s (and a couple handfuls of $109s). Anyway, I think your article has some minor flaws. I PM'd Eastbay about your article, and that's why you see his post (And I seriously doubt he skimmed it. It's more likely he read it through once, and that's all). IMO, Eastbay is likely one of the most valuable resources this forum has with regard to the Math and SitNGos. Don't read this to say that I think he is THE supreme authority, but he's the most valuable math resource I'm aware of, and I have great respect for his opinion and skill.

I did read your article, and although I agree with the lionshare of your article, I think that it misses some subtle important issues, which Eastbay said succinctly - "he has not considered payout structure."

Good luck at the tables.
Scuba

[benfranklin]

As I read this thread, trojanrabbit is not Dennis Bragg, but is writing a couple of articles on the same topic that are yet to be published.

[trojanrabbit]

[ QUOTE ]
As I read this thread, trojanrabbit is not Dennis Bragg, but is writing a couple of articles on the same topic that are yet to be published.

[/ QUOTE ]

Correct.