Out Now
×

25/50 HU, flop/turn decision

Posted by

Posted by posted in High Stakes

25/50 HU, flop/turn decision

BB: $6947.50
SB: $7206.18 (Hero)
Preflop ($75.00) (2 Players)
Hero was dealt 8 T 4 K
Hero raises to $125, BB calls $75
Flop ($275.00) 8 4 K (2 Players)
BB checks, Hero bets $150, BB raises to $699.50, Hero calls $549.50
Turn ($1674.00) 8 4 K 9 (2 Players)
BB checks
Same good aggressive villain from my previous post..This was the first time he c/r me in the match (even though we haven't played that many hands), so I figured he has something here..What's better, flatting or 3betting the flop? I prefer flatting and eveluating turn..If I 3bet, I assume we are going with our hand, but I think if we get it in on the flop, we can't be ahead...I decided to call flop, and he checks the turn.. Help pls

23 Comments

Loading 23 Comments...

Phil Galfond 12 years ago
3betting the flop would be mostly a protection play. You'd get it in if you have to, but given that you block all the sets, it's very likely he'll just fold, or get it in flipping with you.

I think 3betting flop is fine, and I think calling flop is fine.

From a metagame standpoint, there's something very nice about being able to 3bet him the first time he x/r flop in your match. For that reason, I'd definitely choose to 3bet.

Once he checks the turn, he's either giving up with a weak draw (far and away the most likely, in my opinion, since the board is now so draw heavy that he'd likely feel his turn bet has no FE), x/r with a hand better than yours (not likely), x/r with a big draw, x/c with a big draw, or check-deciding with a hand like K8 or AK75ss or AK4xss (2nd most likely maybe).

I think you should go ahead and just bet for protection... probably calling it off if he jams, but not being too happy about it. It's just so likely that he's giving up with a pair+gutter or a weak flushdraw, and not letting him see a river with those hands in a $1600 pot is very valuable.
GameTheory 12 years ago
I have to disagree with Phil on 3-betting the flop. I think getting it in for 140bb on the flop with this holding is really bad. This was the first flop check/raise and against a reasonable range of sets and strong draws we have only around 26% equity:
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hKd 26.51% 903,716 273,163
KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR 73.49% 2,746,821 273,163

Plus on a spade turn our equity drops to around 10%, so we can safely fold. And also having this hand in our bet/callingrange makes our range much stronger on blank turns since we are expected to be 3-betting our sets very often.

Once he checks on the turn I do agree with Phil that he is not likely to be strong and that you should bet and not being to happy about calling if he jams.
TJ Serdar 12 years ago
Couple questions about the syntax...

567[Js-As]:ss means 567+J high-A high FD

K[Js-As][9-A]:ss ????

[Js-As]K[5-7][5-7]:ss!RR ???

Phil Galfond 12 years ago
Good post, GT.

I'm not very adept with the syntax, but I believe you're not factoring preflop decisions in? Many KKxx combos will 3bet, and depending on villain, some 44 and 88 combos will fold. He also won't have many AKxx nfd hands with 3 overs to the 8, as most would 3b pre.

I also think you included many TP+fd hands that are too weak to x/r- not that that will help our equity by removing them.

(please correct and embarrass me if I'm misreading the syntax)

We also can't safely fold a spade turn, at least until we approximate his flop x/r bluff frequency.
Augqie 12 years ago
Adjusting for opponents preflop starting range (tried ranges between 20%!5% - 40%!12%), and/or making his flop range tighter like Phil suggested, hero's equity lies between 26% and 33%. See for details below.

Omaha Hi Simulation ?
876,580 trials (Exhaustive)
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hKd 27.49% 217,756 46,485
(30%!8%):(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,AsK[5-7][5-7]:ss!RR) 72.51% 612,339 46,485
Edit · Link · 2+2 · Deuces Cracked · LeggoPoker

Omaha Hi Simulation ?
1,132,420 trials (Exhaustive)
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hKd 32.89% 345,260 54,351
(40%!12%):(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR) 67.11% 732,809 54,351
Edit · Link · 2+2 · Deuces Cracked · LeggoPoker

Omaha Hi Simulation ?
813,440 trials (Exhaustive)
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hKd 32.62% 244,720 41,241
(30%!12%):(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR) 67.38% 527,479 41,241
Edit · Link · 2+2 · Deuces Cracked · LeggoPoker

Omaha Hi Simulation ?
1,146,360 trials (Exhaustive)
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hKd 30.02% 326,276 35,706
(20%!5%):(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR) 69.98% 784,378 35,706
Edit · Link · 2+2 · Deuces Cracked · LeggoPoker

Omaha Hi Simulation ?
1,531,760 trials (Exhaustive)
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hKd 30.06% 430,048 60,718
(30%!5%):(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR) 69.94% 1,040,994 60,718
Edit · Link · 2+2 · Deuces Cracked · LeggoPoker

Omaha Hi Simulation ?
1,175,060 trials (Exhaustive)
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hKd 33.30% 365,939 50,661
(30%!8%):(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR) 66.70% 758,460 50,661
Edit · Link · 2+2 · Deuces Cracked · LeggoPoker

Omaha Hi Simulation ?
3,923,700 trials (Exhaustive)
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hKd 26.51% 903,716 273,163
KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR 73.49% 2,746,821 273,163
Edit · Link · 2+2 · Deuces Cracked · LeggoPoker
GameTheory 12 years ago
567[Js-As]:ss = 567 with J high or higher flushdraw.
K[Js-As][9-A]:ss = KXY with J high or higher flushdraw and one overcard to the 8.
[Js-As]K[5-7][5-7]:ss!RR = KXYZ with J high or higher flushdraw and a gutshot.
steezy 12 years ago
GameTheory-
I'm on the same page as you about 3betting this flop and getting it in; however, Phil merely suggested 3-betting this flop, not necessarily getting it in. Also, he seems to infer there's some value of having this in hero's flop 3betting range.

Phil-
Would you elaborate on the purpose of 3betting the flop with a hand like this?
GameTheory 12 years ago
Phil is right about that many KKxx, AK[8-Q][8-Q]:ss 3-bet preflop. And that many weak pair + fd combo's won't play for stacks. But I tried to give a simple range, so in order to get close to the real equity you have to add some hands too much here and remove some hands there.

Here we only take rainbow KKxx, tp+nfd hands and more wrap+fd, and put a 10% 3-bet range, 50% defending range:
board: 8s4cKs
Hand Equity Wins Ties
8dTd4hK 26.29% 138,390 38,756
50%!10%:(KK:wxyz,88,44,K84,K8:ss,567[7s-As]:ss,K[As][9-A]:ss,[As]K[5-7][5-7]:ss!RR) 73.71% 422,854 38,756

Equity is still very close to 26%, but calling on a spade is not. Op "figured he has something here" so he shouldn't be facing too many bluffs. That makes AsK[9-Q] without another spade a very nice bluff c/f hand obv.

board: 8s4cKs 2s
Hand Equity Wins Ties
8dTd4hK 9.85% 11,638 5,636
50%!10%:(KK:wxyz,88,44,K84,K8:ss,567[7s-As]:ss,K[As][9-A]:ss,[As]K[5-7][5-7]:ss!RR) 90.15% 129,446 5,636


@ steezy Phil said: "You'd get it in if you have to, (...) get it in flipping with you." So I interpret that as instructing to get it in after you 3-bet the flop.

And I really fail to see the value of having this hand in your flop 3-bet range, you miss out on folding spade turns and you also weaken your range on blank turns by a ton if you always get in sets and 3 pair hands.
However against very aggressive opponents that c/r very frequent here I would advocate making a small raise to around $1525 with a wide range that could include this hand.



Phil Galfond 12 years ago
I think our main disagreement comes down to villain's bluff frequency. OP said that it was the first 3b in the match over a small sample, so he figured he had something. OP also called him good/agressive.

Having K84 in our hand drastically reduces his combinations of value x/r's, and I don't think that the fact he didn't x/r yet in the match should sway us towards assuming that he's not bluffing much, especially if we've seen him play aggressively elsewhere. (which is why I still don't want to fold spade turns)

I guess I can't argue with sims vs. his get-in range, so even if his semibluff frequency is high, you guys are saying if we DO 3bet, we can't profitably call off the rest (assuming he never rebluffs).

Perhaps this is an adjustment in my game due to the bluffy nature of high stakes games, or perhaps it's just a leak, and I'm overplaying. Still, I'd rather leak by overplaying like this and making their lives tough early in a HU match.
GameTheory 12 years ago
I'm not sure if I ever said that his x/r bluff frequency _must_ be low, even though OP figured he had something based on his frequent x/r history and the fact his opponent is risking almost double the size of the pot if he does this with air.

But me and Phil clearly disagree on how to construct ranges against this x/r. From what I read he wants to get it in K84x and sets, potentially some other strong hands like AK[8-Q] with the nutflushdraw. So then his calling range becomes something like Kx + fd, inside wraps, and some more, I would guess JJTT type hands and strong backdoor draws. So that means that on a blank turn his range is very weak and completely capped. And also on a spade turn he has a lot of flushes in his range.

In general if this opponent is playing good aggressive, with Phils ranges we become too weak on blank turns and too strong on spade turns. That's why I think choosing this hand from your range to call on a spade turn is a mistake.

Given my last range and the KT84 of OP we're only against his 73% equity get it in range 0.58%, so that means he should fold a ton to a flop 3-bet or else get it in extremely wide. That makes me rather bluff a ton hands like As8, As4, 84XX, pair + gutshot, pair + clubs. But I still don't want to raise this hand, since it would be a waste to either fold it or get it in with 26%.
And also note that 84xx reduces the combinations of value x/r's almost as strong as K84, especially since Phil said that most KKxx combo's 3-bet preflop!



Sauce123 11 years, 11 months ago
GT- It's not a fair argumentative strategy to take it for granted OOP only gets it in with .58% of hands, and then use that to argue that getting it in with KT84 is bad. I don't think Phil would agree that getting it in with .58% of your flop range is a sound strategy, and I wouldn't either. Or maybe I'm misunderstanding you?
GameTheory 11 years, 11 months ago
Ben you make a fair point, I will elaborate on this:
KK,88,44,K84,K8:ss,567[7s-As]:ss,K[As][9-A]:ss,[As]K[5-7][5-7]:ss!RR, with mostly nutdraws and made hands is 10.48% preflop and 4.40% on the flop.
The I argued that KT84 has very bad equity against this range, in fact it has 21.03%
Then Phil argued that most strong hands in this range 3-bet preflop, this would lead to fewer KKxx combos and fewer AK[8-A][8-A]:ss hands. I agree with Phil that this is a valid assumption, then me and Augqie ran simulations in which the KT84 still has bad equity (26-33%).

Now this 0.58% range was only given that the flop was K84 and that we hold KT84 and the assumptions that Phil suggested himself. Given our hand and the flop only 55% of all total hands are still possible.

As a part of his floprange his get it in range is much larger than 0.58%, since if he calls with 50%!10% his calling range on the flop is only 21.88%. And as part of his flop check/raise range this part becomes even bigger. I never meant to imply that the opponent would fold 99.42% of his range to a 3-bet or anything close to that. The way it was worded it was easy to misunderstood what was meant.

The main purpose of this was to argue with Phil about the card removal effects of having 84(K) in your hand. I counted frequencies of the get it in range when holding 84xx and K84x respectively, to show that the difference was small. The probability to be up against the get it in range goes down less than 30% if we go from 84xx to K84x. Having the naked As is almost as good as having a K. Having an 8 has stronger card removal effects than having a K!


Phil argued that KT84 was good to 3-bet it in, because of card removal effects and game flow. I disagreed with Phil for 3 reasons:
1. KT84 has bad equity to put in 140bb in a single raised pot against even a range with many weak hands as KJs7s5, 3-betting and and folding would be wasting equity.
2. When Phils 3-bet get it in range would contain all sets and 3 pair hands his calling range would become very weak and capped on blank turns. Flatting this hand would be good for his range.
3. Given that many KKxx and AKss type hands 3-bet preflop 84xx becomes a much better candidate for 3-betting the flop as a (blocker) bluff since its equity and playability isn't good enough to continue. It has card removal effects that are only slightly worse than K84x.
Sauce123 11 years, 11 months ago
GT-

Phil's "assumption" that villain 3bets preflop isn't really an assumption at all. I think we all agree that 3betting a non 0 percentage 140bbs deep vs a button open is optimal, and I think all agree that 3betting ranges in PLO should be fairly linear. We might disagree whether the Odds Oracle top 10% is the best 3betting range (I know I would) but because PLO hands include four cards it shouldn't drastically effect the accuracy of your sims whether we 3bet 8%, 12%, 10%, Odds Oracle 10%, or a slightly more playable 10%. But, it's just a mistake to fail to modify villain's flop range by subtracting his 3betting range.

It's also a mistake to fail to modify villain's x/r range by our hand's cards in a case where his x/r range is set and 2PR heavy and we hold T2P ourselves. It makes a big difference whether we hold KT84 or 2222 here.

After taking into account these modifications to villain's range, your sims seem to say villain gets in .58% of his flop range given Phil's dead cards, and against such a range we have between 26 and 33% equity. I'm not sure this should convince us that 3bet calling KT84 is dominated.

If villain, gets in .58% of his flop range, this means a few things. First, if he x/r flop somewhat widely, this leaves him either folding or calling our 3bet with a high frequency. Say his flop x/r range is a fairly conservative 12% when modified by our dead cards. Against our 3bet, he's getting it in with .58%/12% of his flop x/r range, or 4.8% of the time. If that's true, then either (a) he's calling a much larger amount of his x/r range than he's 4betting, perhaps 5-8% of his flop range or ~10-16 times as often as he is 4betting, or (b) he's folding a ton to the 3bet making it's value as a protection bet very high (as well as perhaps making him exploitable to 3bet bluffs).

Considerations like these suggest to me it isn't plausible to argue that the fact we only have 25-30% equity all in against a .58% flopped range entails we shouldn't 3bet. I'm not disagreeing with your conclusion, actually, (I don't think we should 3bet either) but I think you're undermining your argument by choosing an implausible flop strategy for villain- you basically assume villain is playing near clairvoyantly against Phil's exact hand.



GameTheory 11 years, 11 months ago
First I totally agree that to get the most accurate ranges we need to adjust for preflop 3-betting, therefore I said Phils assumptions are valid. My first range didn't adjust for preflop 3-betting, I added unrealisticly many hands for the opponent and still our equity was very bad, 26%. So I felt not much need to make more accurate ranges, since it was super obvious that it would not suddenly become 46%.

We can agree on that Odds Oracle or PPT 10% ranges don't perfectly reflect 3-betting, since strictly linear 3-betting makes your flatting range very unbalanced. A more accurate approach would be using weighted 10% ranges, where the top 10% range sometimes flats. And also some hands that are not top 10% such as 4567 get 3-bet by most regs. The result of this is that the "3-betting effect" gets smaller, so the first range was more accurate than I and Phil first thought.


We should not modify our opponents x/r range to our hand. What we should do is estimate his total x/r range and the get it in part of his x/r range. Then in order to determine the strategy with the highest expectation with our current hand we should calculate the relative frequencies with which our opponents gets it in if we 3-bet the flop and how good our hand plays on turns and how much equity we have etc etc. This was exactly what I was doing, I was calculating absolute probability that that our opponent had his get it in range given our KT84 holding and a suggested 84xx. I'm not sure why you talk about "also a mistake" when this was exactly what I was doing.

What I think Phil is also is greatly overestimating how strong the card removal effect for 3 pair here is. He implicitly makes the assumption that it is very two pair and set heavy and that the bluff part of his range does not contain many of the cards that he blocks.

For instance if our opponent splits his flop x/r range in two parts:
1) his get it in range: 50%!10%:(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR)
2) his bluff range 50%!10%:(K!(44,8,ss,[A-9][Q-T]))
I used the wider get it in range with more non nutflushdraws since it favors Phils argument best to have our opponent stack off light here. This range is actually 0.70% instead of 0.58% so still very close.
His bluff range contains entirely weak Kxxx hands with little equity and bad playability. Of course this this might not be very accurate but it is not unreasonable to use a blocker for topset.

Given that we hold 8dTd4hKd the probability that our opponent is bluffing is now 2638/(2638+1898) = 58%
Now we turn the Kd into the Qd, giving us an 84xx unplayable hand:
Given that we hold 8dTd4hQd the probability that our opponent is bluffing is now 4065/(4065+2611) = 61%

So in this case 3-betting the flop wih QT84 yields 3% more folds!



But back to our KT84 hand on the flop.
50%!10%:(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[Js-As][9-A]:ss,[Js-As]K[5-7][5-7]:ss!RR) is a 0.70% range given our hand, we had 33% against this range.
His total flop range of 50%!10% given we hold 8dTd4hKd was 21.88% like I mentioned in my previous post.
His total x/r now becomes 1.65%, that means if he always checks and we always bet we face a x/r 7.5%

Now the EV of 3-bet getting it in equals:
(+$775*58% +($13895*33%-6947.50)*42%) = -$543
which is clearly worse than folding, since folding is -$275.



You suggested we face a x/r 12%, so I'll widen his x/r bluff range:
2) 50%!10%:(K!(44,8,A,ss))
Then his total x/r range given our hand becomes 2.65%, so we face a x/r 12.1%.
Then given our hand he is bluffing 5368/(5368+1898) = 74%.
Now the EV of 3-bet getting it in equals:
(+$775*74% +($13895*33%-6947.50)*26%) = -$41

Which assumptions are the most reasonable onces? You and Phil must be much better equipped to estimate ranges of 5000 PLO regs. I would say that 74% x/r bluffrange given we hold KT84, which would be 71% if we held 2222, is too many bluffs. OP figured "he had something here".

And I don't think stacking off wider or defending a lot against flop 3-bets by flatting is a plausible strategy either since that would imply he is gonna play a $3000-4000 pot on the turn with TP + 4th nutflushdraws or weaker.
Sauce123 11 years, 11 months ago
GT-

When we're analyzing the strategy to take with a unique hand like KT84, it's good analysis to take into account our dead cards. When we're analyzing how to play our whole range, it's good analysis to take into account the removal effects of our entire ranges for taking various actions given various exact hands- for example, if we bluff with 2222 we need relatively fewer bluffs to make villain indifferent to continuing with his x/r range than if we bluff with 84xx.

In doing good analysis, it's also important to realize that imprecision in the inputs leads to imprecision in the outputs. If we put villain on a very specific x/r range containing a ton of spade draws for instance, we shouldn't be surprised that on analysis we'll find that bluffing with Asxxx has a high EV. But that won't mean that bluffing with Asxxx is a deep truth about the spot and is part of optimal play, rather we should realize we've overfit the data unless we can justify our choice of a x/r range by looking at EVs in the rest of the game tree as well as adjustments villain can make if he knows our choice of Asxxx rebluffs.

Here's a contrived example for a more abstract case. It's easy to see that a x/r range which doesn't contain sets isn't at equilibrium in PLO with deep stacks. My opponent, (knowing I don't hold made hands) will realize that even very weak made hands like bottom pair have enough equity to continue against my range. He'll also know he can pot it in on any blank turn and I'll be forced to fold. So, he'll just bet flop when he has decent equity against my draws and move me off of them when they miss and fold when they hit. If he can do this, I can unilaterally improve my EV by x/r sets and x/r them again on blank turns since he'll be going nuts on those cards, so x/r flop without made hands isn't optimal.

To avoid overfitting (and wasting time doing overly precise analysis) it's important to describe your reasons for composing ranges in a certain way. Once you do that, people can more meaningfully agree or disagree with you and at least agree on the problem which is at issue. In your original analysis (by not specifying a flop x/r frequency) and only getting it in .58% on the flop, I noted that you would be left either folding or smooth calling the 3bet a lot more often than you would be raising it, and that this would add a lot of EV to Phil's 3bet even if he was crushed the (very) few times he got all the money in.

Now that I know you're x/r around 8% and getting it in around .58%, I know you have around a 14:1 ratio between your x/r range and your get it in range. That's a little low for my taste, but not unreasonable. The problem is that your inputs won't say much about Phil's- people in high stakes games almost always x/r flop 12-20% in a spot like this (rightly or wrongly). By specifying the inputs to your analysis it's easier to see your conclusion which is that 3betting KT84 against an 8% x/r range is spewing. But we can get to that conclusion a lot quicker doing a lot less analysis. Hell, villain flops T2P+ or a mega draw nearly 8% of the time, of course it's not right to get it in with dry T2P against that 140 deep in position.

Aleksandra ZenFish 11 years, 11 months ago
Quotes~ Hell, villain flops T2P+ or a mega draw nearly 8% of the time~ ( and that is 1 in 10 approx ) and Phil's- people in high stakes games almost always x/r flop 12-20% in a spot like this (rightly or wrongly). plus added card removal effect..doesnt this 3 things weight heavily toward a decision to 3 bet flop?
Also "important to realize that imprecision in the inputs leads to imprecision in the outputs" makes equity higher then GT suggests as well, so i kinda feel like First comment covers it perfectly
AlexeiMartov 11 years, 11 months ago
"s~ Hell, villain flops T2P+ or a mega draw nearly 8% of the time~ ( and that is 1 in 10 approx ) and Phil's- people in high stakes games almost always x/r flop 12-20% in a spot like this (rightly or wrongly). plus added card removal effect..doesnt this 3 things weight heavily toward a decision to 3 bet flop" - no it means the other guy is check raising 20% of the flop at max but 8% he has a better hand. this means half the time you get it in vs better (which is bad as your equity vs his range is losing you money).

you might be making enough getting them to fold when u shove to counteract the lost equity when running the money in the middle vs their range half the time, but calling to see the turn will give you information regardless of how good the player is (they simply cannot always bet the turn as their air will not have enough equity on average- thus they will have to check their air sometimes- if they choose to balance this with value hands they will make less with their value hands. the argument is that this makes you more when you are ahead (they check their bad hands, you pot and win, it gets checked down on scare cards and you don't lose more when they have better, you stack off on a blank but don't lose too much there as they have some combos pair + draw stuff now so our equity is kinda flippy). the information in this scenario is pretty valuable as it lets you play very aggressively vs their weak hands and often shuts down the action when you were behind when a scare card comes.
AlexeiMartov 11 years, 11 months ago
3 betting flop is wrong at 7k and fine at 5k most opponents imo. its the pot odds relative to stack size that save u getting it in on flop (combined wit blowing tem off 20% equity hands tat check raised aggro). you're not making value from your equity vs their range, you make your value by blowing them off hands with equity and not losing much when getting it in. calling turn is also best imo for the same reason- you have pot odds to call and a good hand but your hand is not best equity wise vs their range, and you will basically always get to showdown for free when a scare card hits (or call on a total blank which will be pretty breakeven/reasonable result for you relative to getting it in 40/60 on the turn or w/e).
AlexeiMartov 11 years, 11 months ago
also vs some people your hand is best on turn cos they are insane check raisers- in that case, apply same concept, if your hand is 60/40 vs their range you should shove to avoid losing value on rivers.
AlexeiMartov 11 years, 11 months ago
"Now that I know you're x/r around 8% and getting it in around .58%, I know you have around a 14:1 ratio between your x/r range and your get it in range. That's a little low for my taste, but not unreasonable"

what sauce is saying here is the breakeven point for getting K8 in if it loses you a fair bit of money shoving (as their range is stronger) might be somewhere around his area as you make up the money lost in all in equity (eg if you had 35% equity in 200 bb pot, your pot equity is 70 bb so u lose 30) in their folded check raises. getting it in 1/14 times you check raise makes me a bit confused though as this is obviously too much check raising and could be exploited by shoving any 4 even with 0 equity.... i might have this wrong though, clarify?
GameTheory 11 years, 11 months ago
I will first reply to Ben as I think that he is making some mistakes that need to be clarified, Alexei is following up on them.

>>>>>When we're analyzing the strategy to take with a unique hand like KT84, it's good analysis to take into account our dead cards.<<<<<
Correct, this was exactly what I was doing. Recognize my range, so there won't be no mistake.




>>>>>In doing good analysis, it's also important to realize that imprecision in the inputs leads to imprecision in the outputs.<<<<<
I'm very aware of this, I don't think that my conclusions are very sensitive to imprecise assumptions about his get it in range or x/r bluff frequency. If you think otherwise please elaborate.




>>>>>If we put villain on a very specific x/r range containing a ton of spade draws for instance, we shouldn't be surprised that on analysis we'll find that bluffing with Asxxx has a high EV.<<<<<
Sorry but this is false. In fact, the contrary is true!
For instance if we assign the following spade draw heavy x/r range to our opponent:
As*s,KK,88, this is a 5.55% range.
If we block the As we know that he has (As*s,KK,88)!(As), a 2.22% range. So when we hold the As we block around 60% of his range. But that does not imply that bluffing with the naked As is a good play, because now we are sure we are up against sets!

If his range was very air heavy, for instance it would include all no pair no flushdraw hands:
(As*s,KK,88,*!(K,8,ss,4,RR))!(Jd) that would be a 18.28% range while holding the irrelevant Jd blocker. Giving a 3.44 to 1 bluff to value ratio.
Now holding the As blocker
(As*s,KK,88,*!(K,8,ss,4,RR))!(As) becomes a 16.05% range, while holding the As blocker the bluff to value ratio is 7.23, more than twice as high when holding the Jd.

In the air heavy range the As is only in 20% of hands where it was in 60% of hands from the value range.
And yet against the value range using the As blocker was terrible, it is even worse than a random hand since random hands have higher equity against flushdraws than against sets. And against the air heavy range it was a very profitable bluff!






>>>>>To avoid overfitting (and wasting time doing overly precise analysis) it's important to describe your reasons for composing ranges in a certain way. Once you do that, people can more meaningfully agree or disagree with you and at least agree on the problem which is at issue.<<<<<
I like to think that my analysis has much overfitting, neither have I wasted time doing overly precise analysis.
As you can see my x/r bluff ranges are very rough: 50%!10%:(K!(44,8,ss,[A-9][Q-T])) just all _bad_ Kxxx hands. Then I checked to see how sensitive this range was to us holding KT84, QT84 and 2222. I found bluff to value ratios of:
1.30 for KT84
1.44 for QT84
0.83 for 2222

This seems very reasonable because on average the x/r bluff range has over 1 pair on the flop, which seems to be in the ballpark of what you would expect. Plus since by definition the bluff x/r does not get it in the equity of that range is not important. Hence spending more time on improving this range would be wasting time just like you said.

Furthermore I think you and Phil are contradicting eachother:
1. Phil said most KKxx and AK[8-Q]:ss type hands would 3-bet preflop, making the absolute probability that he has a hand from his get it in range on the flop low.
2. Phil said many weak TP + fd would not x/r (get it in) on the flop.
3. You said that opponents x/r this flop between 12 and 20%.
4. You said it was not a fair argumentative strategy of me to say that he would only get it in with TP + 3rd nutflushdraw minimum because that would artificially make our equity too low.





>>>>>Now that I know you're x/r around 8% and getting it in around .58%, I know you have around a 14:1 ratio between your x/r range and your get it in range. That's a little low for my taste, but not unreasonable<<<<<
These numbers are wrong.
Given our hand the range with only the TP+ nut fd hands:
50%!10%:(KK,88,44,K84,K8:ss,AsK*s,567[Js-As]:ss,K[As][9-A]:ss,[As]K[5-7][5-7]:ss!RR) equals 0.59%.
But his preflop calling range is:
50%!10%, which equals only 21.88%!

This is very important and you seem to have missed it. But this is just something that cannot ignore, our hand influences the range that he has on the flop. For instance if we hold your favorite hands 2222 his preflop calling range is (still):
50%!10%, which now equals 27.69%!

Thats why you cannot say "he x/r 12% here" without specifying to what. It must be relative to him calling preflop, our hand and the flop. That is why I worked out examples where we had a 4 to 1 x/r to x/r get it in ratio given your assumption of 12% x/r when we hold this hand and a 0.70% get it in range. With the 0.58/0.59% range this ratio becomes 4.5, but definitely not anywhere close to 14.


>>>>>3betting KT84 against an 8% x/r range is spewing. But we can get to that conclusion a lot quicker doing a lot less analysis. Hell, villain flops T2P+ or a mega draw nearly 8% of the time, of course it's not right to get it in with dry T2P against that 140 deep in position.<<<<<
Where do you get this 8% from? According to my math he has this 0.59% get it in range 2.7% of the time given our holding and 4.8% in general when his preflop calling range is 50%!10%.

For he rest I agree, this is what I said in my first post:

##### I think getting it in for 140bb on the flop with this holding is really bad. This was the first flop check/raise and against a reasonable range of sets and strong draws we have only around 26% equity: ######
I constructed a range that was too wide in my opinion and then our equity was still terrible so I concluded that getting it in for 140bb in a SRP was bad. Then Phil commented on this range and we had to go deeper.....

Be the first to add a comment

Runitonce.com uses cookies to give you the best experience. Learn more about our Cookie Policy