T9s river x/r-bluff

No problem, glad I could help

Hope to see you in the coaching tomorrow

Originally posted by floaty
SB is a bad lag sorta. Is the bluff ok?
PokerStars - $0.25/$0.50 (6 max) - Holdem - 3 players
Hand converted by PokerTracker 3

Hero (BB): $8.39
BTN: $24.78
SB: $7.09

SB posts SB $0.10, Hero posts BB $0.25

Pre Flop: ($0.35) Hero has T♠ 9♠

fold, SB raises to $0.50, Hero calls $0.25

Flop: ($1.00, 2 players) 6♦ J♦ 8♣
SB bets $0.25, Hero raises to $0.50, SB calls $0.25

Turn: ($2.00, 2 players) 2♥
SB bets $0.50, Hero calls $0.50

River: ($3.00, 2 players) K♦
SB bets $0.50, Hero raises to $1.00,

Spoiler

foldHero wins $3.84

I guess I'd better talk a little bit about the hand hadn't I? Sorry floaty, got a bit carried away there:

I agree w/ Leader that this is pretty much a nailed on 3-bet if you're playing a split-range strategy, which for SB vs BB I think is absolutely fine. You're in position and the SB is opening wide so the ranges are extrememly wide and, as long as you protect your calling range somewhat splitting it isn't bad.

As played, raise on flop is standard but unless you have some kind of read on your opponent that his donk bet means something made then I don't see any reason not to bluff-raise the turn here. The only reason not to raise here is if you never raise vs a turn donk bet, otherwise your raising range will just be hugely value-heavy and he can actually exploit you by donk/folding all his weak hands on the turn.

On the river, it looks a good card to represent but one of the problems is that we block many of the hands we want to fold (i.e. QT, Q9 for straight draws) and he really shouldn't be folding a decent value betting hand here even on a card like that (I'd also expect him to give up with some of his worse VB hands)

Now if you can tell me that his donks mean "weak made hand looking to "bet to see where he is" then obviously this play makes sense and I like it but that's pretty specific rather than Bad LAG

Summary:

Pre-Flop: 3-bet

Flop: Fine as played, if he donked into you after you 3-bet I'd call and raise the turn IP

Turn: Raise as played unless you have no raising range on the turn vs a donk bet

River: Not a huge fan as we block a lot of draws we'd want to fold, the FD has come in and in theory there shouldn't be many value-betting hands he's folding to your raise

However with a specific read could change my stance on the river.

ooh, the hand made it to page 2 or something made it to page 2, anyway. I didnt raise the hand on the turn cause i felt it was a "weaker" draw than say a flush draw but can still hit a 9 or T and be good for a call down and I didnt think raising all my draws at this point would be valid since I had so many of em compared to the value range, i wouldve raised all my J+ A8 but probably just called down weaker made hands since my read was that he was on an 8 or a 6 or a pp around there or even A2 picking up the 2 there... but I actually was considering raising but decided not to cause i felt it was overbluffing here. Turns out when i look it up in equilab that would make it ~33% draws. But yea, if I only raise my fds here im not raising enough, but isnt raising all draws overdoing it? iguess that if ide raise K8 Q8 that would be some extra combos there, but still alot of my range is draws at this point.

about the read, the bad lag, I was just lazy when i wrote it, i didnt have many hands on him, think he was 32/18 or something and not overly aggressive and not very sd-bound either, but over such a small sample that I couldnt really say for sure. but I think he played the hand really bad so I kinda went with the "bad lag"description from there, but no, i didnt have a specific read about him donking weak hands in this spot, but it was more of a general read on "players like him" or players he was starting to look like...

I guess I'd better talk a little bit about the hand hadn't I? Sorry floaty, got a bit carried away there:

No worries, Im glad a discussion came out of it

but Im still kinda curious, so Im gonna be annoying and continue here. Say I wouldve played the hand better, 3b on the bb here and he'd call. And it had gone he checks, I bet, he raises, I call. he bets turn, should i raise here then? and what is a good devided range for 3b in the bb here?

You don't need to raise all your draws on this board no, but a T-High OESD is pretty much mandatory, if you're going to cut bluffs to raise take out things like QT/Q9 (i.e. gutshots with some high card value) and your A/K-hgh FDs since they have litle fold equity vs what you want to fold

Just to come up with a bit of a range assuming you're not 3-betting the BB and defending around 85% of hands we come up with something like this to raise the flop with:

Combonator Output combonator.com
Board: 6 J 8

Entire range: 975

Grouped: 400 combos 41.0%
Ungrouped: 575 combos 59.0%

Group 1: 279 combos 69.8% (28.6% total)
Group 2: 121 combos 30.3% (12.4% total)

Group 1: Pair of eights or better

88+, 66, AJs, A8s, AJo, A8o, KJs, K8s, KJo, K8o, QJs, Q8s, QJo,
Q8o, J2s+, J2o+, T8s, T8o, 98s, 98o, 82s+, 85o+

Group 2: Queen-high or worse with flush draw, jack-high or worse with gutshot or better

QdTd, Qd9d, Qd7d, Qd5d, Qd4d, Qd3d, Qd2d, T9s, T7s, Td5d, Td4d,
Td3d, Td2d, T9o, T7o, 97s, 95s, 9d4d, 9d3d, 9d2d, 97o, 95o, 75s-
74s, 7d3d, 7d2d, 75o, 54s, 5d3d, 5d2d, 54o, 4d3d, 4d2d, 3d2d

Now to answer your 2nd question:

Yes is the answer.

He's going to, or should have a load of draws in his flop x/r range. Of course we block some of them but T9 here is actually ideal. From an exploitative sense you may have fake outs vs a PP or an 8 (unlikely but possible) and from a balance persective you have one of the worst high-card value hands in your 3-betting range and you have a strong draw, perfect.

As far as a divided range, that's a toughie because I know my 3-betting range is split and weighted (i.e. I 3-bet 25% of the time w/ a certain hand etc)

When I read the OP (copied below for convenience), I said to myself that the hand was pretty well played. After reading all the brilliant subsequent analyzes, I wasn't entirely sure anymore.

Originally posted by floaty
SB is a bad lag sorta. Is the bluff ok?
PokerStars - $0.25/$0.50 (6 max) - Holdem - 3 players
Hand converted by PokerTracker 3

Hero (BB): $8.39
BTN: $24.78
SB: $7.09

SB posts SB $0.10, Hero posts BB $0.25

Pre Flop: ($0.35) Hero has T♠ 9♠

fold, SB raises to $0.50, Hero calls $0.25

Flop: ($1.00, 2 players) 6♦ J♦ 8♣
SB bets $0.25, Hero raises to $0.50, SB calls $0.25

Turn: ($2.00, 2 players) 2♥
SB bets $0.50, Hero calls $0.50

River: ($3.00, 2 players) K♦
SB bets $0.50, Hero raises to $1.00,

Spoiler

foldHero wins $3.84

But what if one turn things around again, and supposes that optimal strategy versus an average opponent at these stakes is a blend between GTO and old fashioned poker psychology, hand reading, value oriented bets and aggression? Then the play in the hand, as played, can't be very far from the optimum.

So, assuming that the opponent isn't too brilliant, I believe that the best line of play is the line that has the best possibility to make the opponent fold, combined with the possibility to bail out cheaply if the cards fall off badly. (One needs not win every battle, only the war.)

Preflop: A 3-bet (which I b t w agree with in general) is for balancing only, or, if seen as an isolated hand, for deception only. We are surely behind. From the old fashioned POV, it is a clear call for (drawing) value. From the psychological (and mathematical) point of view, it might be easier to bluff the opponent off the hand later on by just calling, since the pot is smaller. It is decidedly cheaper to bail out if the flop comes badly.

Flop: What would be a better spot to semi-bluff raise? The opponent might (given that he is not so brilliant) fold lots of stuff he shouldn't if he knew what we had. (Sklansky reasoning, out of fashion, but nonetheless correct.)

Turn: This is a tricky spot. The opponent just confirmed that he isn't all that brilliant. If he is decent, he'd know that we would bet if he checks. (The free card raise is out of fashion, at least vs one opponent.) Now, this is where we have to decide whether to invest 3 bets into what will probably be a bluff (we have some 8 outs) or take our chances with the draw and optionally bluff on the river.

If the analysis says that he will not fold anything with value, then a raise here is senseless.

If the analysis says that he is more likely to fold to a raise turn - bet river sequence, he can still hit on the river and come out donking again.

River: The bluff raise gives 3.5:1. On the one hand, those are poor odds. On the other hand, his odds of 4.5:1 for a call aren't the best in the world either. Surely he can fold a pair here, he will see all sorts of bogeymen, since a river raise always looks pretty strong. We can even represent (to him, since he isn't brilliant) a slowplayed monster. Indeed, we block a number of hands he should be bluffing with, but does he know precisely which hands he should bluff with? Also, his turn donk (to the extent that it is "readable") doesn't exactly scream Q9.

With his irrational turn donk he has quite likely committed himself to bet 100% of his range on the river. He might even be bluffing with A high given which is a pretty bad move.

/***/

I have tried, for fun, to analyze the hand from a psychological perspective as opposed to a technical analysis. The conclusions differ a bit. I don't argue with the technical analysis, but a technical analysis isn't all inclusive. In NL there are situations where the technical analysis decidedly takes the back seat. The psychological aspects of the game are still there (to a much lesser extent) in FL - or it wouldn't be poker between humans.

/Johan =

Originally posted by Boomer2k10

Originally posted by Leader22

Originally posted by floaty

Originally posted by Leader22

We agree there. The problem is that he's unlikely to fold any of that imo.

But I have to have some bluffs right? If I only raise flushes or 2pairs then my whole range is value oriented and there would be no reason for him to call down with mid-pairs... but perhaps I dont have to think that way in this spot against this opponent...

That is the problem with the game theory way of thinking at low limits. There's a prescription sometimes with less experienced players that good play is mostly about balance. That's generally not true below 10/20. The players below that are very often poor and exploitable in clear, consistent ways, and you will rarely be precisely counter exploited by excellent opponents at these limits. Therefore, the focus is not on ensuring against leaks in your game, but maximizing the value of your opponents faults.

So how are we to play against an unknown opponent then if not to try and play in a balanced way?

Of course exploitative play is better against players who have obvious leaks, no-one has ever said otherwise, and if anyone has they're wrong, however poker is a game of incomplete information especially if you are unfamiliar with your opponent so just assuming he's not going to adapt or play well against you because you're not in a 10/20 game is a pretty big presumption.

Yes this guy has donked the turn on a brick, something a "standard" player is unlikely ever to do and has been labelled as a "bad lag" but in what way? Without anything more specific how are we supposed to take an optimal exploitative line?

The reason I disagree with that thinking is there really aren't unknowns. We can say a lot about how the population players act. Thus, we can make adjustments against players before they even play a hand.

When it comes to counter adjustment, I would make a few quick points. The frequency which the spot arises is a very significant factor in if we will face a counter adjustment. The quality of the opponent is a key factor the frequency and accuracy of the counter adjustments we will face. The likelihood of facing a counter adjustment relates significantly to how clear, profitable, and frequent the original adjustment is amongst the players in the games, and finally, I'm skeptical of the argument that, in practice, we must preemptively adjust our play based on the presumption of future adaptation as there are many players that do not adjust or who we will see infrequently enough for such concerns to be irrelevant.

Sorry Leader I can't say I agree with this

You're saying you disagree with pre-eptimvely exploitating/adjusting based on the fact our opponent will adjust and yet you're willing to pre-emptively exploit based on the average player at the limit?

I just don't like this line of thinking for anything above micros, I just feel it totally caps your ability, especially now the games are getting tougher, and results in this artificial equilibrium we see in the games where there are spots which people play horribly but that's ok but everyone plays them just as badly so the rake wins

Maybe I'm being stubborn in my beliefs but for me I would much prefer to know where the balance point is and react based on individuals rather than just assume everyone's making the same mistakes and base my gameplan around that because I know at least a handful of players that would crush me out of sight if I made those assumptions

Originally posted by Boomer2k10
Sorry Leader I can't say I agree with this

You're saying you disagree with pre-eptimvely exploitating/adjusting based on the fact our opponent will adjust and yet you're willing to pre-emptively exploit based on the average player at the limit?

If you know that the average pf open from UTG is 15% but you believe the correct pf open is 19%, are you going to assume that an unknown opens 19% and react accordingly?

I just don't like this line of thinking for anything above micros, I just feel it totally caps your ability, especially now the games are getting tougher, and results in this artificial equilibrium we see in the games where there are spots which people play horribly but that's ok but everyone plays them just as badly so the rake wins

Maybe I'm being stubborn in my beliefs but for me I would much prefer to know where the balance point is and react based on individuals rather than just assume everyone's making the same mistakes and base my gameplan around that because I know at least a handful of players that would crush me out of sight if I made those assumptions

Playing exploitively first doesn't preclude an advanced understanding of game theory. It would merely subordinate the execution of gto lines to those spots where it is truly needed.

I don't agree that our strategy should be based around avoiding making mistakes in the very rare case that we happen upon an excellent player which we've never seen before. Even in this rare case, the effect is limited by our observation of the player and subsequent adjustment to gto.

I also think it's an overly broad representation of my position to say I'm suggesting that we should "just assume everyone's making the same mistakes." I'm suggesting that based on an analytical look at different situations we assume players act in a certain way until there is evidence otherwise. In more technical terms, I'm suggesting we adopt a Bayesian approach where prior experience and data are updated by current actions of a specific player.

To show the argument for this approach, let's reexamine the underlying principle of game theory which is that we know nothing about the opponents' strategies, and thus should seek to minimize our loss in all spots by assuming our opponents play perfectly. If we know anything of relevance (even in the presence of uncertainty*), we should adjust our strategy to maximally exploit the game. Thus, in order to believe that we should always default to gto first in poker, we must believe that past hands played by other players bare no relationship to future hands played by new players. It follows then that there is no population of new players, and that every new player that enters the game basically adopts a completely random approach initially. The only way to conclusively disprove this would be through cross-validation of say players styles in their first hundred hands, which would take a little too much time atm. However, I'm about as confident as it's possible to be without having done it that such a study would in fact disprove that players adopt a random approach in their first hands. Thus, I'm confident that in the presence of past data we should adjust based on that data because it is predictive of future players' actions.

*A simple example of this would be if I gave you a coin that you didn't know was fair and had you flip it. Let's say it comes up heads. If we where to bet on the next flip, it can be shown that you should bet on heads.

If everything would be so black and white as you present it... Look back at Boomer's post - he said nothing about GTO - talked about a game theoratical approach, that is completely different (tho it is based on GTO). But anyway, let's just stick to what you said...

If the average PF open from UTG is 15%, does that suggest that everyone raises 15% from UTG? What if half the people raise 11% and half of them 19% or 99% of the players have a 19% opening range, and the 1% has whatever-that-makes-the-average. Should you assume they have a 15% opening range? Obviously not - and here lies the absolutely moronic PEE-thinking that spreads like a virus. Statistics are good tools to push the exact meaning we want into anything that has to do with math. Statistics is a nice part of mathematics, however statistical argument is the most dirty, mean, manipulative and misleading kind of argument anyone ever invented. It's filled with biases (that tend to be overlooked for the sake of proving a point), misinterpretations and willful misinterpretations.

So back to that average 15%, there are infinite different scenarios where that number can come from. Of course, there's only one that is true - but do we know that? Do we have enough of a proof to say that this exactly is how an average player plays? How about if we know most of the player pool, and an unknown turns up. Should we assume he plays "average", or should we assume he's an "average fish", or should we weigh it, but how much then? I'd be really surprised if anyone had the answer for the above questions.

But let's say even that's settled. There are I-don't-know-how-many-million different scenarios that can come up in a HU game (don't even talk about 6max). Do we know the general tendencies in all the cases? Surely not, but we can guess, we can implicate, deduct. Do we have evidence that our deductions are correct? Can we put a number on "how often this action means, that in a different spot he's gonna do that"? Put some of these together, and you'll come up with a number way below what we can actually assume as a "general tendency".

And even if we know all the average tendencies, can we calculate the best response for all of them in real time? Cause "people rarely bluffraise the river" doesn't mean we should fold, since they should rarely bluffraise the river (except for certain situations). That and several other PEE-catchphrases in themselves are not arguments, but used as such. But in a lot of other cases, even if people tend to over- or underdo something, it doesn't mean that we have an exact answer (since the nature of EV doesn't work solely on a % basis).

To know all this, you have to have a representative-enough database of a representative-enough number of players, you also need... oh wait, we failed at the first point. Since datamining is illegal, furthermore very very expensive, I don't think anyone has hundred-billion hands stored on their hard drive to be able to do a fair analysis on all this, nor the computing power to analyse every situation. Of course, we can simplify stuff by guessing - but isn't the whole point of science not to be guessing? Or should we just guess how correct our guesses are... never mind.

But hey, let's assume, this still works out for us (that obviously it doesn't). We're done with the first hand. How about the second one? How did the first hand affect the player? Will he deduce something out of the results? Or he isn't leveling at all? Did it make him tilt in a positive or negative way? Did something happen on other tables that could've made him tilt or just think otherwise? Is he being bothered now by his wife, or losing focus cause he's being blown by a thai hooker? (Ok, the latter is less likely than I want to, but still, you get my point.)

Now if you have all this, and a lot more figured out, you can play perfect exploitation. The counterargument here is that GTO-play hasn't been solved, so how can we base stuff on that? It's easy - we know a lot about how GTO works, we just don't know the exact solution - therefore we can work it into our game partially. It doesn't mean we lose value overall, it only means that we admit that general tendecies cannot be known to the extent we need. I truly believe everyone that states he knows and can work with it perfectly, is either a liar or delusional.

I don't say there is no place for exploitative play, or even player tendencies in poker, cause there is. But that place is much more limited than most people think. People shoot themselves in the foot by oversimplifying the perfect exploitative play to the point, where it makes less sense than just defaulting to a good GTO-based strategy. And it's not the people's fault, since the human mind can't do any better, nor the current state of a home-PC. It's those people's fault who preach it - strictly IMO.

tl;dr version: there's an old lady, who is 66 years and 364 days old (and no, it's not a leap year) - do you wanna bet she dies today, cause the average life expectancy is 67?

Originally posted by madorjan
If everything would be so black and white as you present it... Look back at Boomer's post - he said nothing about GTO - talked about a game theoratical approach, that is completely different (tho it is based on GTO). But anyway, let's just stick to what you said...

If the average PF open from UTG is 15%, does that suggest that everyone raises 15% from UTG? What if half the people raise 11% and half of them 19% or 99% of the players have a 19% opening range, and the 1% has whatever-that-makes-the-average. Should you assume they have a 15% opening range? Obviously not - and here lies the absolutely moronic PEE-thinking that spreads like a virus. Statistics are good tools to push the exact meaning we want into anything that has to do with math. Statistics is a nice part of mathematics, however statistical argument is the most dirty, mean, manipulative and misleading kind of argument anyone ever invented. It's filled with biases (that tend to be overlooked for the sake of proving a point), misinterpretations and willful misinterpretations.

So back to that average 15%, there are infinite different scenarios where that number can come from. Of course, there's only one that is true - but do we know that? Do we have enough of a proof to say that this exactly is how an average player plays? How about if we know most of the player pool, and an unknown turns up. Should we assume he plays "average", or should we assume he's an "average fish", or should we weigh it, but how much then? I'd be really surprised if anyone had the answer for the above questions.

But let's say even that's settled. There are I-don't-know-how-many-million different scenarios that can come up in a HU game (don't even talk about 6max). Do we know the general tendencies in all the cases? Surely not, but we can guess, we can implicate, deduct. Do we have evidence that our deductions are correct? Can we put a number on "how often this action means, that in a different spot he's gonna do that"? Put some of these together, and you'll come up with a number way below what we can actually assume as a "general tendency".

And even if we know all the average tendencies, can we calculate the best response for all of them in real time? Cause "people rarely bluffraise the river" doesn't mean we should fold, since they should rarely bluffraise the river (except for certain situations). That and several other PEE-catchphrases in themselves are not arguments, but used as such. But in a lot of other cases, even if people tend to over- or underdo something, it doesn't mean that we have an exact answer (since the nature of EV doesn't work solely on a % basis).

To know all this, you have to have a representative-enough database of a representative-enough number of players, you also need... oh wait, we failed at the first point. Since datamining is illegal, furthermore very very expensive, I don't think anyone has hundred-billion hands stored on their hard drive to be able to do a fair analysis on all this, nor the computing power to analyse every situation. Of course, we can simplify stuff by guessing - but isn't the whole point of science not to be guessing? Or should we just guess how correct our guesses are... never mind.

But hey, let's assume, this still works out for us (that obviously it doesn't). We're done with the first hand. How about the second one? How did the first hand affect the player? Will he deduce something out of the results? Or he isn't leveling at all? Did it make him tilt in a positive or negative way? Did something happen on other tables that could've made him tilt or just think otherwise? Is he being bothered now by his wife, or losing focus cause he's being blown by a thai hooker? (Ok, the latter is less likely than I want to, but still, you get my point.)

Now if you have all this, and a lot more figured out, you can play perfect exploitation. The counterargument here is that GTO-play hasn't been solved, so how can we base stuff on that? It's easy - we know a lot about how GTO works, we just don't know the exact solution - therefore we can work it into our game partially. It doesn't mean we lose value overall, it only means that we admit that general tendecies cannot be known to the extent we need. I truly believe everyone that states he knows and can work with it perfectly, is either a liar or delusional.

I don't say there is no place for exploitative play, or even player tendencies in poker, cause there is. But that place is much more limited than most people think. People shoot themselves in the foot by oversimplifying the perfect exploitative play to the point, where it makes less sense than just defaulting to a good GTO-based strategy. And it's not the people's fault, since the human mind can't do any better, nor the current state of a home-PC. It's those people's fault who preach it - strictly IMO.

tl;dr version: there's an old lady, who is 66 years and 364 days old (and no, it's not a leap year) - do you wanna bet she dies today, cause the average life expectancy is 67?

Before I get started, it always seems like these discussions about gto and exploitive play take on a certain divisive character. To be absolutely clear, my position is never that people on the other side of this debate are horrible or suck. It has been clearly shown that a player can win at the top levels using that approach.

-----

The positions as I understand them are either a) start with GTO and adjust based on individual players or b) start with GTO and adjust based on all available relevant information. Thus we disagree about what's relevant and to what degree, not about taking a game theory approach. Perhaps I wasn't clear enough in about GTO still being the base.

I have to call you out on the 19% example. If 99% have 19% open and 1% have something else that pfr would be -381%.

What strikes most regarding the argument about the open percent is not that you don't bring up the draw backs, which for the record I think are able to be mitigated with distributional theory, because you do, but rather it begs the question: what's your solution? You're really going to assume it's 19%? What if the average was 1%? What if you'd played a million hands and no one had ever opened UTG? Is there any level of evidence that would have you change your strategy? If there is, then it's just a matter of assessing that evidence and determining its relevance.

The argument you make about statistics is a fairly common one especially outside of poker in politics or economics. It is, however, based on the false premise that statistical arguments or analyses are not understandable by people not trained in the field. Statistics is math. It's not less "pure" then game theory or real analysis. Its theorems are proven with algebra, calculus and real analysis just like group theory or combinatorics. Statistics is the way math interacts with the real world. If you think your being manipulated, show how. You don't have to take my word for anything just like I didn't with the 19% average example.

I agree that poker is highly complex and that many actions we take will be imperfectly derived from determinations about other spots. However, your argument seems to be that if we can't figure out every spot using data, then it's of no value. I don't agree. Overall, it seems like you see my position as "game theoretic approach is inferior to a data analysis approach" where my actual position is "game theoretic approach is supplemented by a data analysis approach."

Your argument about execution is valid regardless of the way we determine our strategy. We often have less then ten seconds to make a decision. It's not any easier to derive the GTO solution plus adjustments in ten seconds than it is to query a database and do a statistical analysis as both are impossible. My approach to this problem has thus been to train reactions away from the table to enable instinctual reactions in real time.

Your discussion of sample sizes is not well informed. It seems to derive from your argument that if we can't determine everything were better off determining nothing, which I addressed above. I've seen hundreds of examples both in and out of poker where small samples have shown significant results.

You discuss never having "the exact answer," which is true as statistics is incapable of ever giving an exact answer since statistical analysis assumes uncertainty. However, this fails to account for the facts that there are mathematical lines which can be drawn in any spot and that the results of analysis on observed data can be bound within realistic limits. When those realistic limits fall above or below the boundry, you have an answer with high confidence.

Some of your questions in the seventh paragraph touch on the critique of statistics that you can pick an individual case and determine its deviance by close observation in a way that would otherwise be unknowable. Clearly, if you have more information you will do better at predicting the outcome, but that doesn't invalidate that determinations can be made about the population as a whole without that information and that those determinations can increase ev overall even if there is a minority (in magnitude not number) of individuals within the population for whom the play is -ev.

You touch on that players make adjustments. The OP seems to me to be a great example of why this concept is getting too much emphasis. Because we don't bluff here, we now face the prospect that this player will subsequently fold more? However I haven't seen any evidence that this player (or any player) is likely to make such an adjustment, and evidence could be produced if it was true.

My position is not that we can play "perfect exploitation." Decisions are determined based on the balance of the evidence and sometimes we will determine incorrectly. The question is what method determines the correct play most often. Also, a lot is known about statistics and it's applications. In fact most all the work on game theory in academia is now in statistics and computer science departments. Game theory is a primary focus in statistical training after the first year of grad school. This is the book we used if you need proof: http://www.amazon.com/Statistical-Decision-Bayesian-Analysis-Statistics/dp/0387960988

I disagree with you that what people think is not their fault. Making arguments never hurts anyone. Everyone here is capable of reason and discernment. If they choose to believe something to their detriment, that's on them. This is of course a very small example of the much larger argument that Europeans and Americans often have.

Average life expectancy is measured from birth. Thus, the average life expectancy for someone who is 66 years and 364 days is not 67 since we know they are 66 years and 364 days. If it was, then yes I would bet you whatever you want though since it's even money I would take the other side. I don't like cheering for people to die.

So in summery, I don't think we disagree that much on the theory or approach in practice. It seems to me that you believe past data is of very limited value where as I believe it's of significant value. As I said above, players are responsible for their approach to the game, they should weigh the arguments and examples and decide for themselves. Either way though I think they'll do fine if they apply themselves.

Before I get started, it always seems like these discussions about gto and exploitive play take on a certain personal character. To be absolutely clear, my position is never that people on the other side of this debate are horrible or suck. It has been clearly shown that a player can win at the top levels using that approach.

Have to agree completely, I didn't want to be personal at all, sorry if I did. I believe people make it personal, cause (as we seemingly agree on) it's more of a belief than actual evidence. Still, I think we need these kind of discussions, since (as we seemingly don't agree) most players don't have the capability to decide which side to take, or what to believe in. And honestly, this is one of the best posts I've read from you ever, so it was worth it regardless.

I don't think that either is superior to the other objectively, I just think that the statistics-based approach has much more negative effects (see the beloved PEE), can cause more harm, especially in a learning environment. If you consider maxEV as the ultimate goal, and assume a great statistician player, probably your approach is better. If, however, you consider hourly rate (and a good balanced player), I'd be really surprised if it wasn't hugely in balanced players' favor. Now the question of maxEV vs hourly is always an issue, but learning to be a great statistician I believe is much harder than learning balance-based math.

I have to call you out on the 19% example. If 99% have 19% open and 1% have something else that pfr would be -281%.

What strikes most regarding the argument about the open percent is not that you don't bring up the draw backs, which for the record I think are able to be mitigated with distributional theory, because you do, but rather it begs the question: what's your solution? You're really going to assume it's 19%? What if the average was 1%? What if you'd played a million hands and no one had ever opened UTG? Is there any level of evidence that would have you change your strategy? If there is, then it's just a matter of assessing that evidence and determining its relevance.

Nice catch with the 19% stuff, but you get my point regardless.♦

My solution is pretty simple: either don't assume anything, play your best approximation of GTO, or (what I tend to do) assume something wishy-washy about general tendencies (e.g. probably doesn't fold top 5%, probably doesn't raise bottom 60% or so), and build a strategy against this unknown but narrowed range.

My strategy in general, and what I (maybe wrongly) call balanced strategy is nothing like GTO play - it doesn't assume the set of all strategies randomly distributed, it weighs them in a certain way. Obviously so far it alignes with your view. The difference is that my approach tries to minimalize the damage that false assumptions lead to, while yours just take it in the face. The way I minimalize damage is the use of some concepts learnt from GTO. So all in all, I believe the main difference is the personal answer to the question: how confident are you in your presumptions. I failed and saw some really good players failing in their assumptions, and so far only seen one, who didn't.

BTW a really good exercise I always mention in these balance vs exploit arguments is that take certain non-standard situatuions, and ask yourself about the probability of something exact happening. E.g. how often donkbets mean top pair+? The two thing every one of the players realized after guessing and looking it up: they were not even close to the correct answer, and they didn't have a big enough sample size on these exact situations.

The argument you make about statistics is a fairly common one especially outside of poker in politics or economics. It is, however, based on the false premise that statistical arguments or analyses are not understandable by people not trained in the field. Statistics is math. It's not less "pure" then game theory or real analysis. Its theorems are proven with algebra, calculus and real analysis just like group theory or combinatorics. Statistics is the way math interacts with the real world. If you think your being manipulated, show how. You don't have to take my word for anything just like I didn't with the 19% average example.

I don't say statistics don't have a place in real life or math or poker. The thing I hate about statistics is that it works against the standard human mind. It works with probability - a thing that human mind is programmed to be incapable of accepting. Understanding - it may happen, but never accepting. You have to be a genius with really out-of-the-box thinking and an extraordinary mind to be able to accept it. When 10 reds come in a row in roulette, your mind tells you that a black will come, even if you're aware of gambler's phallacy. You look for patterns, rules, regularity, balance, etc. It's an evolutionary need for survival, since nature mostly works in predictable way. Probability theory, and therefore statistics works against this instinct of ours, so most people's mind is limited in terms of understanding all of that. (And you can see this kind of behaviour all around the forums, starting from downswing-bitching through winrate-questions to hand analysis.) It's not a surprise it's used for deceptive and mean purposes all around the world.

Your argument about execution is valid regardless of the way we determine our strategy. We often have less then ten seconds to make a decision. It's not any easier to derive the GTO solution plus adjustments in ten seconds than it is to query a database and do a statistical analysis as both are impossible. My approach to this problem has thus been to train reactions away from the table to enable instinctual reactions in real time.

This is absolutely true, however, I never argued that GTO is better or more executable. A balance-based (or GTO-based) strategy is. Determining where you are at your range, and how you want to deviate from the standard play is much easier than thinking through the opponent's possible range, your equity vs that range, etc. It's much more static, robotic if you like, less room for creativity at the tables, etc. If we wanna go by empirical evidence, my borderline of playing my best game was on 4 6-max tables when I played exploitatively, with the balance-approach I can play 4 HU tables with around the same amount of focus.

You discuss never having "the exact answer,"...

By "exact answer" I meant the statistical "exact answer" - meaning e.g. a weighed range, not an exact binary range.

Clearly, if you have more information you will do better at predicting the outcome, but that doesn't invalidate that determinations can be made about the population as a whole without that information and that those determinations can increase ev overall even if there is a minority (in magnitude not number) of individuals within the population for whom the play is -ev.

Yeah, what I touched on later is that it matters a lot how +ev and -ev those cases are. A certain example is a hand in my ex-blog where I called a river raise with the second nuts, when the nuts was 1 combo. The vast majority of the people who saw the post said that it's probably more than 50% chance Villain is an overplayer, so 3betting is the best play there. The problem is that in this case 50% is actually not enough, we need the guy to be an overplayer 66% (assuming we don't fold to a cap in a thousand BB pot) of the times. And tho it's fairly simple if you look at it as an equity number (with weighed ranges), it seems that people can't be abstractive enough to immediately see that. It goes back to the same argument that I made in the UTG-open example.

You touch on that players make adjustments.

I'd like to rephrase this. I touch on that player's decisions are not static. They can be willful deviations (adjustments), but also uncontrolled stuff (like the uncontrollable thai hooker). That gives an other incertain variable to our equation, that makes it even harder to determine anything close to what we need.

I disagree with you that what people think is not their fault. Making arguments never hurts anyone. Everyone here is capable of reason and discernment. If they choose to believe something to their detriment, that's on them. This is of course a very small example of the much larger argument that Europeans and Americans often have.

Yep, probably the standard EU vs US stuff, so I won't even start an argument.♦

All in all, I think it's a great discussion, and also shows how similar the two approach is. And obviously I was joking with the average life expectancy stuff, pointing out how some people (not you obv) use statistics. I'm happy that we can so easily come to terms, and it doesn't turn into a bloodbath as these discussions usually do. I really appreciate your input.

tl;dr version: it's a question of belief, go away!