A shocking discovery..... shocking to me at least

[cartman]

This may be a little premature, but I have been doing some pokerstove research and it appears at least at this point that:

Heads up against a preflop raiser, we should essentially never fold a pair prior to the river for one bet assuming he will autobet the flop and the turn

Am I crazy?


Thanks,
Cartman

[Yarney]

Depends on the player, but with most aggressive players this is the case, depending on the board and how much paint is on it.

-Yarney

[___1___]

cartman,

Ummm...I really don't think so.

So, UTG+1 who is 20/14/4 raises and I call in the BB with T [img]/images/graemlins/spade.gif[/img]9 [img]/images/graemlins/spade.gif[/img].

Flop comes A [img]/images/graemlins/heart.gif[/img]K [img]/images/graemlins/diamond.gif[/img]9 [img]/images/graemlins/club.gif[/img]

We're calling down with this? Just one example but you get the picture...

Edit: My point is just that we can usually narrow an opponents range down to a point where calling down with a pair, regardless of whether opponent auto-bets flop and turn, is incorrect in many instances.

___1___

[cartman]

An AK board is one of the only exceptions I think. Maybe my methodology is flawed. See if this makes sense:

Assume opponents raising range from his specific position this hand is 20%. From that I estimated his range to be:
A8+, KT+, QJ, A3s+, K8s+, Q9s+, J9s+, 55.

I have 23.

To estimate the chances that I am ahead on the turn when the board is AQ28 for instance, I just entered a river card of a 4 which I thought was the ultimate blank. Pokerstove claims that if I showed down my 23 on the final board of AQ284, I would have the winner 21.2% of the time. So I think that is a reasonable estimate of how often I am ahead on the turn. (This is the key assumption to the entire endeavor in my opinion. Is it valid?)

Assuming he open-raised preflop, only I call in the big blind, and I check call the flop, then after he bets the turn the pot will contain 4.25BB. So if he bets the river (of course he won't always) I am getting 5.25:2 odds to call down. If I estimate that I have 5 outs when I'm behind and that he has 6 outs when I'm ahead, then I need to be ahead on the turn 22.2% of the time.

That's awfully close to the 21.2% estimate that I get from pokerstove. Changing the Q to a K in my example, plunges this figure to 15.7% and makes it a clear fold, but on an AQ or AJ board it looks like a toss up (pokerstove gives 21.2% for both) and on an AT (28.4%) or a KQ (32.7%) board it is a very clear call.

Is there a flaw in my method?

Is my assumption above valid?

What do you guys think?


Thanks,
Cartman

[___1___]

[ QUOTE ]
An AK board is one of the only exceptions I think. Maybe my methodology is flawed. See if this makes sense:

Assume opponents raising range from his specific position this hand is 20%. From that I estimated his range to be:
A8+, KT+, QJ, A3s+, K8s+, Q9s+, J9s+, 55.

I have 23.

To estimate the chances that I am ahead on the turn when the board is AQ28 for instance, I just entered a river card of a 4 which I thought was the ultimate blank. Pokerstove claims that if I showed down my 23 on the final board of AQ284, I would have the winner 21.2% of the time. So I think that is a reasonable estimate of how often I am ahead on the turn. (This is the key assumption to the entire endeavor in my opinion. Is it valid?)

Assuming he open-raised preflop, only I call in the big blind, and I check call the flop, then after he bets the turn the pot will contain 4.25BB. So if he bets the river (of course he won't always) I am getting 5.25:2 odds to call down. If I estimate that I have 5 outs when I'm behind and that he has 6 outs when I'm ahead, then I need to be ahead on the turn 22.2% of the time.

That's awfully close to the 21.2% estimate that I get from pokerstove. Changing the Q to a K in my example, plunges this figure to 15.7% and makes it a clear fold, but on an AQ or AJ board it looks like a toss up (pokerstove gives 21.2% for both) and on an AT (28.4%) or a KQ (32.7%) board it is a very clear call.

Is there a flaw in my method?

Is my assumption above valid?

What do you guys think?



[/ QUOTE ]

I do think the method and assumptions look pretty accurate. I'm just not sure if or how you can translate this methodology into real-life situations. Honestly, in your 23 example, how many people can bring themselves to call down in such a situation? I don't know if I could.

The person to really as about this is Peter_Rus. He's done similar analysis with regard to calling an open-raise out of the BB heads-up...

___1___

Shouldnt we also consider how often the guy "fires the last barrel" ?

[cartman]

[ QUOTE ]
Shouldnt we also consider how often the guy "fires the last barrel" ?

[/ QUOTE ]

Definitely we should. Some opponents will also not automatically bet the turn. On each street, we must re-estimate his hand range given that he bet. Against an opponent who will autobet the flop and the turn, his estimated hand range is the same as his estimated preflop range.

My indication that we are getting 5.25:2 odds on a call down is somewhat incorrect because of the "firing the third barrel" problem. If we always call the river when we call the turn (probably not a good idea) and he only bets the river when we are behind, then I guess we are actually only getting 4.25:2 on a call down, which would require us to be ahead far more often on the turn (27.8% instead of the 22.2% that getting 5.25:2 requires).

Of course he will not actually play perfectly on the river as he will sometimes bet with worse hands and check with better hands. But we will also sometimes fold a better hand. This is all very inexact and highly dependent on our continual assessment of his hand range on each street given that he bet.

My speculation is that most of the potential +EV that I miss by incorrectly folding weak pairs on the turn is from situations in which I would call the turn and then he would check behind on the river with nothing and I win. Against an opponent who always fires the third barrel, then the analysis above would be dead on in my opinion.

Even if he plays perfectly the river, though, I think the preliminary conclusions are still staggering. Unless I'm making a mistake in my reasoning, even if he plays perfectly on the river thereby reducing our odds to 4.25:2 to call down, we should call down with any hand that we estimate has at least 27.8% chance of being ahead on the turn. That means calling the turn bet with 23 on an AT28 board intending to also call a river bet against an opponent who will autobet the turn.

This is so astonishing to me that I feel like I must still be missing something. I suspect the correct move may be to call the turn bet and usually fold if he follows through on the river.

What do you guys think?

Thanks,
Cartman

[redbeard]

regarding the 4 as a blank on the river and the result is you win 21.2% of the time your "key assumption" i have to believe this to be correct. the 4 didn't help you at all and has very little chance of improving him so i would have to say this is a great assumption. my next question is the flop call on a AQ2 board. does poker stove indicate that this should be called? if not then we wouldn't even get to the turn question. i do understand what you are getting at though. i think that heads up i too often underestimate the value of even the smallest pair on a board that is ace high. this could revolutionize my game quite a bit. i look forward to hearing how everyone else weighs in on this.

[etizzle]

i assume you meant 55+?

[cartman]

[ QUOTE ]
i assume you meant 55+?

[/ QUOTE ]

Right. Thank you, I mistyped.

Cartman