Two Plus Two Publishing LLC
Two Plus Two Publishing LLC
 

Go Back   Two Plus Two Poker Forums > >

Notices

Heads Up NL Discussion of heads up NL Texas Hold'em cash games

Reply
 
Thread Tools Display Modes
Old 01-11-2014, 03:48 PM   #126
ron1n
centurion
 
Join Date: Nov 2013
Posts: 188
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by kaby View Post
...it's a totally arbitrary number
I don't understand how it can be arbitrary(?!?). Just playing around with CardrunnersEV will show that when applying max-exploit to a strategy which, in some random decision node, folds more then threshold, the betting/raising range in that spot of the exploitative strategy shifts to 100%. So its clearly not arbitrary...

On the other hand, I can accept arguments on how the EV gained by switching to the 100% betting/raising range as a result of max-exploit on that particular street may not generate the highest EV overall for the exploitative strategy (as far as I can tell CardrunnersEV's max-exploit doesn't work for multi-street/full tree) but I would say it will def be higher then the EV of not switching to betting/raising more as a bluff in that spot.

Last edited by ron1n; 01-11-2014 at 03:53 PM.
ron1n is offline   Reply With Quote
Old 01-11-2014, 04:27 PM   #127
GntlmnsHndshk
enthusiast
 
Join Date: Jan 2014
Location: www.runitonce.com
Posts: 81
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by ron1n View Post
I don't understand how it can be arbitrary(?!?). Just playing around with CardrunnersEV will show that when applying max-exploit to a strategy which, in some random decision node, folds more then threshold, the betting/raising range in that spot of the exploitative strategy shifts to 100%. So its clearly not arbitrary...
Set up a situation in CREV, have the oop player bet flop, turn and then checks the river. The in position player then bets half pot on the river. What frequency do you think the oop player needs to defend?

Check your results in CREV and get back to me.
GntlmnsHndshk is offline   Reply With Quote
Old 01-11-2014, 04:45 PM   #128
ron1n
centurion
 
Join Date: Nov 2013
Posts: 188
Re: 100NL Coolers or bad play?

I dont understand exactly what are you trying to say. Create the situation yourself and put it here...
ron1n is offline   Reply With Quote
Old 01-11-2014, 04:51 PM   #129
fityfmi
old hand
 
fityfmi's Avatar
 
Join Date: Jan 2013
Location: Only the ladder is real.
Posts: 1,678
Re: 100NL Coolers or bad play?

If I defend 55-60% of my hands in the BB when facing a minraise, would you call me tight, loose, or something in between?
fityfmi is offline   Reply With Quote
Old 01-11-2014, 05:32 PM   #130
GntlmnsHndshk
enthusiast
 
Join Date: Jan 2014
Location: www.runitonce.com
Posts: 81
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by ron1n View Post
I dont understand exactly what are you trying to say. Create the situation yourself and put it here...
PREFLOP:
SB/BU opens to 2x, BB 3-bets to 7.5x. BU flats.

FLOP ($15) Q T 6
BB bets half pot. BU calls.

TURN ($30) Q 7 6 2
BB bets half pot. BU calls.

RIVER TURN ($60) Q T 6 2 2
BB checks. BU bets $30.

How often should BB defend here?

Come up some reasonable ranges for the river.
GntlmnsHndshk is offline   Reply With Quote
Old 01-11-2014, 05:32 PM   #131
czechvengeance
old hand
 
czechvengeance's Avatar
 
Join Date: Mar 2012
Location: perfecting my bumhunting skillz
Posts: 1,800
Re: 100NL Coolers or bad play?

@fityfmi: tight
czechvengeance is offline   Reply With Quote
Old 01-11-2014, 07:01 PM   #132
ron1n
centurion
 
Join Date: Nov 2013
Posts: 188
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by GntlmnsHndshk View Post
PREFLOP:
SB/BU opens to 2x, BB 3-bets to 7.5x. BU flats.

FLOP ($15) Q T 6
BB bets half pot. BU calls.

TURN ($30) Q 7 6 2
BB bets half pot. BU calls.

RIVER TURN ($60) Q T 6 2 2
BB checks. BU bets $30.

How often should BB defend here?

Come up some reasonable ranges for the river.
BB's strategy, in order to be *in the spirit* of GTO, should be to call with up to 66% of his bluffcatchers if that means he will reach the threshold of 66% of his entire range by doing so OR to call with all his bluffcatchers if those represent less then 66% of his entire range.

Last edited by ron1n; 01-11-2014 at 07:06 PM.
ron1n is offline   Reply With Quote
Old 01-11-2014, 07:09 PM   #133
GntlmnsHndshk
enthusiast
 
Join Date: Jan 2014
Location: www.runitonce.com
Posts: 81
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by ron1n View Post
BB's strategy, in order to be *in the spirit* of GTO, should be to call with up to 66% of his bluffcatchers if that means he will reach the threshold of 66% of his entire range by doing so OR to call with all his bluffcatchers if those represent less then 66% of his entire range.
So you say, did you put it into CREV and solve it?
GntlmnsHndshk is offline   Reply With Quote
Old 01-11-2014, 07:14 PM   #134
Rei Ayanami
Carpal \'Tunnel
 
Rei Ayanami's Avatar
 
Join Date: Aug 2012
Posts: 11,524
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by ron1n View Post
BB's strategy, in order to be *in the spirit* of GTO, should be to call with up to 66% of his bluffcatchers ...
This isn't in the spirit of GTO at all. Defending ~66.67% of our bluffcatchers would ensure half-street indifference (of villain's "air")*, but poker hands have more than a half-street of play.

ETA: It might turn out that the solution is to defend ~ that % of them, but the crude sort of calculation you're referencing is often way off.

ETA 2: *Provided we can't raise.

Last edited by Rei Ayanami; 01-11-2014 at 07:43 PM.
Rei Ayanami is offline   Reply With Quote
Old 01-11-2014, 08:49 PM   #135
fityfmi
old hand
 
fityfmi's Avatar
 
Join Date: Jan 2013
Location: Only the ladder is real.
Posts: 1,678
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by czechvengeance View Post
@fityfmi: tight
I had a feeling this was the case. Then I think there is no wonder Spladle will have to fold enough on the flop to make it profitable to cbet any two. I'm not saying that makes defending your BB with 70-80% of hands wrong though.
fityfmi is offline   Reply With Quote
Old 01-12-2014, 02:09 AM   #136
GntlmnsHndshk
enthusiast
 
Join Date: Jan 2014
Location: www.runitonce.com
Posts: 81
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by Rei Ayanami View Post
This isn't in the spirit of GTO at all. Defending ~66.67% of our bluffcatchers would ensure half-street indifference (of villain's "air")*, but poker hands have more than a half-street of play.

ETA: It might turn out that the solution is to defend ~ that % of them, but the crude sort of calculation you're referencing is often way off.

ETA 2: *Provided we can't raise.
fwiw, as i've defined the question, it is kinda a half street game. But this doesn't change the fact that you don't need to defend enough even if it's on the river and it's a half street game.

When you think about the weakest part of the BU/SBs range on the river, the weakest hand is going to have decent equity against the BBs range. When most people fabricate a half street game, they make the bluffs have 0% equity, but this isn't always the case as is here. So the BB doesn't need to defend enough to make the BUs bluffs have EV of 0, the BB only needs to defend enough to make the SB indifferent between betting and checking -- which can be quite a bit less.
GntlmnsHndshk is offline   Reply With Quote
Old 01-12-2014, 02:18 AM   #137
Rei Ayanami
Carpal \'Tunnel
 
Rei Ayanami's Avatar
 
Join Date: Aug 2012
Posts: 11,524
Re: 100NL Coolers or bad play?

Yah, that too.
Rei Ayanami is offline   Reply With Quote
Old 01-12-2014, 05:48 AM   #138
JudgeHoldem1848
veteran
 
Join Date: Jun 2012
Posts: 2,463
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by fityfmi View Post
I had a feeling this was the case. Then I think there is no wonder Spladle will have to fold enough on the flop to make it profitable to cbet any two. I'm not saying that makes defending your BB with 70-80% of hands wrong though.
You should really do a game tree in CREV. Take the villain you have the most hands against and start plugging ranges into it. Just estimate one bet size for each decision, it's close enough and a sh**ton simpler. It takes a while but you'll quickly see how bad folding a hand like J7o is to a min raise pf even considering how often that hand gets folded otf vs a cbet.
JudgeHoldem1848 is offline   Reply With Quote
Old 01-12-2014, 11:33 AM   #139
ron1n
centurion
 
Join Date: Nov 2013
Posts: 188
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by GntlmnsHndshk View Post
PREFLOP:
SB/BU opens to 2x, BB 3-bets to 7.5x. BU flats.

FLOP ($15) Q T 6
BB bets half pot. BU calls.

TURN ($30) Q 7 6 2
BB bets half pot. BU calls.

RIVER TURN ($60) Q T 6 2 2
BB checks. BU bets
OK. Suppose you play this river vs 3 unknown opponents: BU1 bets $60; BU2 bets $30 and BU3 bets $1.

a) Does your river defending range changes from one to another?

b) If yes, why?
ron1n is offline   Reply With Quote
Old 01-12-2014, 02:18 PM   #140
GntlmnsHndshk
enthusiast
 
Join Date: Jan 2014
Location: www.runitonce.com
Posts: 81
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by ron1n View Post
OK. Suppose you play this river vs 3 unknown opponents: BU1 bets $60; BU2 bets $30 and BU3 bets $1.

a) Does your river defending range changes from one to another?

b) If yes, why?
yes, your river defending range changes based on your opponents size because you're going to need to make you're going to need to defend enough so your opponent is indifferent to betting vs checking with some hand(s). The smaller the BUs bet, the more likely we're going to need to have a wider defending range.
GntlmnsHndshk is offline   Reply With Quote
Old 01-12-2014, 02:48 PM   #141
ron1n
centurion
 
Join Date: Nov 2013
Posts: 188
Re: 100NL Coolers or bad play?

So you agree that the 1-alpha threshold (where alpha=pot/(pot+bet)) forces you to change your defend range.

And somehow using this thresholds as starting point to develop a [bayesian] balanced strategy vs common bet sizes in common spots it's still outrageous.
ron1n is offline   Reply With Quote
Old 01-12-2014, 03:18 PM   #142
GntlmnsHndshk
enthusiast
 
Join Date: Jan 2014
Location: www.runitonce.com
Posts: 81
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by ron1n View Post
So you agree that the 1-alpha threshold (where alpha=pot/(pot+bet)) forces you to change your defend range.

And somehow using this thresholds as starting point to develop a [bayesian] balanced strategy vs common bet sizes in common spots it's still outrageous.
where did I say anything about a 1-alpha threshold?
GntlmnsHndshk is offline   Reply With Quote
Old 01-12-2014, 03:23 PM   #143
icoon
grinder
 
icoon's Avatar
 
Join Date: Sep 2010
Posts: 678
Re: 100NL Coolers or bad play?

The 1-alpha threshold is only relevant in a very specific situation, namely a river spot where one player has a polar range with sufficient bluffs and the other player has a blufcatcher. You can not generalize this result to other situations. (Even in most river situations it would be completely wrong to defend 1-alpha).

We can safely assume that in every spot we never want to defend a wider range against a bigger betsize than against a smaller betsize, but just how often we want to defend depends on the exact situation. We simply want to defend with the range that minimizes the ev of our opponents maximally exploitative strategy.In the polar vs blufcather on the river this happens to be 1-alpha, in almost every other situation this will be a totally different number.

What we can do is check if the ideas we have about SBs and BBs strategys are consistent with each other. For example, if we would somehow know for sure that SBs GTO strategy involves openfolding some hands on the button preflop, than we also know that BBs strategy should be such that SB can not play those hands he folds profitably for any other line than folding, thus we can deduce that BB should defend >50% time against a minraise, >33% of the time against a 3x, play aggresively enough against a limp, etc. If we know for sure that SBs GTO strategy involves folding every hand worse than 63o, continuing with every hand better than 63o, and folding and minraising with 63o a nonzero % of the time, we can deduce that BB should play such that SB is indifferent between folding and minraising 63o. If we know for sure that SBs strategy involves openfolding no hands on the button, than we can only deduce that BB should play such that SB can profitably play any hand. We cant figure out the exact defending percentage withouth solving the game, withouth making more assumptions it could be 80% or 4% for all we know. Like Spladle said many times, there is no fundamental reason why the SB shouldnt be able to bet 100% of hands profitably in a lot of spots. So,defending exactly 50% preflop against a minraise or 33% against a 3x is completely arbitrary.

Last edited by icoon; 01-12-2014 at 03:49 PM.
icoon is offline   Reply With Quote
Old 01-12-2014, 11:11 PM   #144
Spladle
Pooh-Bah
 
Join Date: Nov 2004
Posts: 5,518
Re: 100NL Coolers or bad play?

Quote:
Originally Posted by ron1n View Post
So it will! If BB turn lead its only value then SB defend should be top x% of his range where x is ~threshold.
No, this is wrong. The SB should defend those hands that can profitably continue against the BB's betting range and fold those hands that cannot. The threshold you keep referring to is irrelevant.

Quote:
Originally Posted by ron1n View Post
Lets see what happens: BB leads turn with only value which is no more then 20% (top pairs+)
No, you can't just arbitrarily set the "value" range at top pairs+; the range that the BB can profitably bet the turn with is a function of the range that the SB checks behind on the flop with. The weaker that range is after the turn comes, the more frequently the BB can bet for value, and vice versa.

Quote:
Originally Posted by ron1n View Post
- SB will have to defend some %
No, this is wrong, the SB doesn't have to defend any particular %.

Quote:
Originally Posted by ron1n View Post
lets round it up to pot bet and 50% defend, to not make BB bluff all his hands.
Because almost every hand will still have some equity when called, the SB calling 50% of the time is not sufficient to prevent the BB from profitably semi-bluffing with any hand he can actually hold. However, even if the BB can profitably bet all of his hands after the SB checks behind on the flop, that doesn't necessarily mean that he should. Just as the SB should almost certainly check back the flop sometimes even when he can profitably bet ATC, the BB should check the turn sometimes (after the SB checks behind on the flop) even if he could profitably bet with ATC.

Poker is more complicated than you think it is.

Quote:
Originally Posted by ron1n View Post
And I suppose BB will also barrel river close to 100 % where SB folds another 50%. The rest 80% when BB doesnt leads the turn, they see a river and basically SB wins around 70% of the time (obv quite simplistic but it will do the job).

So we have:

EV_SB_whenBBleads= timesBBleads * (timesSBdefendOTT*amoutlostOTT + timesSBdefendsOTT*timesSBdefendsOTR*amountlostOTR)
EV_SB_whenBBleads= 0.20*(0.5*(-4) + 0.5*0.5*(-12)) = 0.2*(-5)= -1

EV_SB_whenBBchecks= timesBBchecks * equity * pot= 2.24
EV_SB_whenBBchecks= 0.80*0.70*4= 2.24
---------------------------------------
EV_SB= +1.24
What job is this calculation supposed to do? It has no relevance whatsoever to poker.
Spladle is offline   Reply With Quote
Old 01-14-2014, 06:02 AM   #145
ron1n
centurion
 
Join Date: Nov 2013
Posts: 188
Re: 100NL Coolers or bad play?

Oh well I think realise now where my mistake was...

In the AKQ, BB check/calls with the K at the frequency (50%) of the threshold otherwise SB should deviate from betting at equilibrium. But BB's call freq coincides with the threshold because he doesn't have hands that are losing vs. SB's bluffs.

IMO, the AKQ toy game should be presented rather like the AKQTJ toy game where BB has {K,T,J} and SB has {A,Q}.

Now its becoming clear that BB should call only with his bluff catchers at a freq of 50% to be at equilibrium! So in this new AKQJT game, BB's TOTAL check/call freq is ~16.7%.


Nevertheless, what i said was not technically untrue because in many spots, usually in those where we didn't bet anywhere, our entire range is at least a bluff catcher. But this doesn't obv apply when we already bet somewhere and polarised our range.

2 scenarios:
We are in SB (srp pot) and

1) BB leads flop we call, BB leads turn
2) BB c/c flop and leads the turn.

In case (1) our total call freq should coincide with the 1-alpha threshold, both on the flop and on the turn, but in the 2nd case on the turn this is no longer true because we polarised our ranges by betting the flop and so the amount of bluff catchers we arrive on the turn no longer coincide with our entire range.
ron1n is offline   Reply With Quote
Old 01-14-2014, 10:44 AM   #146
ron1n
centurion
 
Join Date: Nov 2013
Posts: 188
Re: 100NL Coolers or bad play?

Also this quote is from "Applications of No-Limit Holdem" by Matthew Janda (recommended by sauce) - its about raising a flop bet:

Quote:
While the previous section addressed how to design a balanced flop raising range which consists of the proper ratio of value to bluff raises, we have not yet considered how strong a hand must be to effectively value raise the flop. To do this, we now must calculate what percentage of our opponent’s flop betting range should go to showdown if we raise the flop and bet the turn and river. If we raise our opponent’s 6 big blind flop bet to 18 big blinds, he must defend at least 43.8 percent of his flop betting range to prevent us from being able to profitably bluff any two cards.

(14)(1-X) - (18)(X) = 0

=> X=0.438

where
14 is the size of the pot after our opponent bets,
18 is the size of our bet, and
X is the percent of the flop betting range that ouropponent must defend with.
How do you guys comment on that?
ron1n is offline   Reply With Quote
Old 01-14-2014, 11:05 AM   #147
Rei Ayanami
Carpal \'Tunnel
 
Rei Ayanami's Avatar
 
Join Date: Aug 2012
Posts: 11,524
Re: 100NL Coolers or bad play?

Those are raises, not bets. That isn't just a minor detail--it drastically changes the nature of the situation.

Think about it.
Rei Ayanami is offline   Reply With Quote
Old 01-14-2014, 12:25 PM   #148
ron1n
centurion
 
Join Date: Nov 2013
Posts: 188
Re: 100NL Coolers or bad play?

Ok,
same book about facing a bet:

Quote:
For example, suppose we want to compare how much money our opponent must put into the pot if we bet 0.5 PSB, 1 PSB, or 2 PSB on the river and our opponent does not want us to be able to profitably bluff any two cards.

1.For 0.5 PSB, our opponent must call or raise 66.7 percent of the time to keep us indifferent to bluffing. He thus ends up putting in on average at least 0.334 pot-sized bets.

0.334 = (0.5)(0.667)

2.For 1 PSB, our opponent must call or raise 50 percent of the time to keep us indifferent to bluffing. He thus ends up putting in on average at least 0.5 pot-sized bets.

0.5 = (1)(0.5)

3.For 2 PSB, our opponent must call or raise at least 33.4 percent of the time to keep us indifferent to bluffing. He thus ends up putting in at least 0.668 pot-sized bets.

0.668 = (2)(0.334)
about defending vs a bet in position
Quote:
So to start, let’s first examine how often we would have tofold to a flop bet before our opponent can profitably bet twounwinnable cards To help do this, the following equation shows,in terms of pot-sized bets, how often a bluff needs to suceed toassure that the overall bet is guaranteed a certain profit.

Minimum bluff success rate =
(bet size in PSB) / (bet size in PSB + 1)
or
Y=X/(X+1)
where
X is the bet size in terms of pot-sized bets, and
Y is the frequency the bluff must succeed to show an immediate profit.

For example, suppose our opponent bets 50 percent of thepot. This means bluff rate should be 33.3 percent of the time

0.333 = 0.5/(0.5 + 1)

Similarly, the proper defending frequency can be solved forwhen expressed in terms of the opponent’s flop bet sizing.

Y=1/(1+X)

where
X is the bet size in terms of pot-sized bets.
Y is the frequency the potential caller must defend toprevent his opponent from being able to profitably bluff with pure air, and

So in this example where X = 0.5 (pot sized bets) the potentialcaller needs to defend 66.7 percent of the time.

0.667=1/(1+0.5)

Last edited by ron1n; 01-14-2014 at 12:33 PM.
ron1n is offline   Reply With Quote
Old 01-14-2014, 12:42 PM   #149
Rei Ayanami
Carpal \'Tunnel
 
Rei Ayanami's Avatar
 
Join Date: Aug 2012
Posts: 11,524
Re: 100NL Coolers or bad play?

You're framing those excerpts in a misleading way.

https://forumserver.twoplustwo.com/sh...&postcount=193

https://forumserver.twoplustwo.com/sh...7&postcount=25
Rei Ayanami is offline   Reply With Quote
Old 01-14-2014, 05:39 PM   #150
Mr Blonde
Pooh-Bah
 
Mr Blonde's Avatar
 
Join Date: Sep 2012
Posts: 3,661
Re: 100NL Coolers or bad play?

Damn spladle is doing some serious work in this thread. Very enlightening
Mr Blonde is offline   Reply With Quote

Reply
      

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


Forum Jump


All times are GMT -4. The time now is 09:38 AM.


Powered by vBulletin®
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
Copyright © 2008-2017, Two Plus Two Interactive
 
 
Poker Players - Streaming Live Online