How do you win in poker?

The Game Room
41

That’s true - in rocks, paper, and scissors. It’s not true in poker.

Here’s an example. You’re playing poker against Robbie the Robot. Robbie has been programmed to know all the probabilities. But he has not been programmed to practice any form of deceit. He will always play strictly by the numbers. He will only bet if he calculates the odds are in his favor.

You have a hand with four queens. They’re visible on the table so Robbie knows you have four queens.

Robbie’s visible hand is three kings. He also has a hole card which you can’t see. But you know that your hand will beat any hand Robbie can have unless his hole card is the fourth king. And you know the odds of that are slim. So you play the odds and raise, in expectation that Robbie will fold.

But then Robbie doesn’t fall. He calls and raises. What do you do now? Remember, Robbie only plays the probabilities. He can see that you have four queens. And he can see his hole card, which you cannot. And he is raising so he must be calculating that the odds favor his winning the hand.

At this point, what are you going to conclude Robbie’s hold card is? And based on that conclusion, are you going to call or fold?

This demonstrates how rational play works. Now here’s how irrational play works.

Robbie had a short circuit. His hole card is a four. His hand is three kings and he can see your hand is four queens. There is no rational way that Robbie can expect to win this hand. But the short circuit has damaged his brain and is making him do the opposite of what he should do. He is acting irrationally and is betting more money on a hand he knows he has no chance of winning.

But as pointed out above, you are expecting Robbie to be rational so you fold. Robbie just won money by acting irrationally that he would have lost if he had been acting rationally. Irrational play can beat rational play.

You might argue that we don’t play poker with robots. We play poker with human beings and we expect the possibility of irrational play. And you’re right.

But if your claim was correct - that “irrational play can never beat rational play” - then we wouldn’t expect irrational play. We would always use rational play if it always beat irrational play. The reason we expect the possibility of irrational play is because it does sometimes get used. And the reason it sometimes gets used is because sometimes it beats rational play.

Three kings and a four cannot beat four queens in a rational play. But three kings and a four can beat four queens in an irrational play.

42

I don’t think we’re using the word “irrational” the same way. A proper bluff is not “irrational.” It’s a rational play whereby the bluffer knows that she has a positive expected value from her play.

Irrational play, by definition, is random and done without regard to expected value. If you simply bluff random amounts at random times, you’ll lose all your money.

[QUOTE=Little Nemo]
But my point is that while you have to know the math, you have to also go beyond the math in order to win on a regular basis.
[/QUOTE]

Well, it depends who you’re playing against.
f you look around the table and see faces you’ve seen on TV poker shows, and the guy next to you is Doyle Brunson and he’s saying “Stop looking at my face and play yoour cards, son” then yeah, you mix up your play a little or else you’ll never get paid for your good hands.

If you’re playing fish and tourists at a low stakes game, I assure you they will bring plenty of irrationality into the equation for you. You won’t win a ton playing ABC poker - actually, depending on the rake, you can hardly make anything at all at lower stakes - but you’ll still be one of the more successful players at the table, and it’s how a person must learn.

43

You’re positing a very poorly-programmed robot, incorrectly defining that poor programming to be “rational”, and then using that to claim that rational play is inferior. What’s that supposed to prove? You might as well just posit a robot that always raises as heavily as it can and then folds at the end, and claim that that robot is rational.

44

Exactly. Sklansky has a great chapter in his book, Theory of Poker, that touches on game theory and why bluffing needs to be part of one’s game.

Annie Duke wrote a pretty good book on NLHE tourneys that says pretty much the same thing and shows why it’s true. How bluffing shouldn’t be considered a ‘ghetto’ part of your game but a legitimate part of it.

That said, your bluff has to make sense, i.e. it should tell a coherent story. If your actions in the hand don’t support your story then your opponent is likely to take notice.

Sent from my LGLS990 using Tapatalk

45

Agreed. But you were the one who wrote this:

I’ve been pointing out that if you only play by the probabilities, you’ll probably lose to somebody who occasionally doesn’t play by the probabilities.

46

Again, I am disputing this. Admittedly, I have not given a completely rigorous argument, but imagine that you are able to play an optimal (mixed) minimax strategy. This will ensure an average level of winnings no matter what your opponent does.

47

In no limit holdem especially a lot of pros joke (only semi jokingly) that you don’t really need to know what cards you have to play well. After getting to know the table they will figure out the player’s various temperaments and strategies. At that point they can put pressure on someone when they think he’s weak, bluff someone who always plays tight, call every time someone with an obvious tell is bluffing, raise every hand preflop at a table of timid players and just keep scooping up blinds till someone finally re-raises and then fold, making a profit overall with this strategy, all without ever looking at their cards.

I wasn’t really exaggerating that many pros could beat a table of amateurs without knowing what cards they have in their hands given enough time. Once you assume everyone is on the same level in terms of understanding the odds, pot odds, position, probabilities, etc. the cards become of secondary importance to how you play them. Especially in no limit games where immense pressure can be applied at any moment making a snap call into a very difficult decision. The huge pots and tournament wins very often go to someone with what you would consider a really crappy hand if they were just playing by the odds and nothing else.

So as a new player you learn really quickly not to play something like J 5 offsuit but as an experienced player you might use a J 5 offsuit to take down the biggest pot of the night, just depending on the situation at the table at that moment in time.

48

Okay, use my example above.

The other players has four queens showing so both of you know that. You have three kings showing so both of you see that. And you have a hold card which you can see but the other player can’t.

According to mathematical probabilities, this is a very simple hand. You look at your hold card. If it’s a king, you have a hundred percent certainty of beating four queens. If it’s not a king, you have a hundred percent certainty of losing to four queens. According to mathematical probabilities, you should raise if you have a king and fold if you don’t. And you’ll play it that way every time. That’s what playing the mathematical probabilities means.

Do you think playing like that is the best way to play?

49

Playing blind is certainly an element of Three-Card Brag :slight_smile:

No. To me, “playing the probabilities” means play the actual optimal strategy which in this case will be of course to bet with a winning hand, but with a losing hand I will still bet with some small, but non-zero, probability we can calculate. Since you don’t know what I have, you won’t know whether to call or fold and it doesn’t matter which you choose.

50

That’s not necessarily true, no. Even granting your statement the most charitable interpretation possible, and assuming by “doesn’t play by the probabilities” you mean “sometimes does things other than just betting strong hands and folding weak ones” it depends what you mean. Are we talking about a donkey who just does random shit? Play ABC poker and you will usually beat them. They will make mistakes and you will profit. Are we talking about Phil Ivey? Play ABC poker and you will usually lose, because Phil isn’t making many mistakes. So which ones are you talking about? Because believe me, I am never going to play Phil Ivey. If he sits down at my table, I’m moving. Nobody in this thread can afford to play his stakes anyway.

But in truth, playing the probabilities is how you win in poker - even when you consider bluffing and deception, which requires an understanding of the probabilities to do in a sensible manner. There is a time when the probabilities say a bluff has a positive expectation, and a time when they say it doesn’t. The extent to which you must use deception is proportional to the skill of your opponents. Against drunk tourists, you really don’t need to bluff more than once in a blue moon. Against skilled opponents your game must be much more dynamic.

51

If you have four Queens showing and bet into someone who has three Kings showing after all the cards have been dealt, you have not played rationally … you have made a mistake.

52

Abraham Lincoln used to ask people a riddle: How many legs would a donkey have if you call its tail a leg? And people would generally say the answer is a donkey would have five legs if you called its tail a leg. And then Lincoln would tell them they were wrong; a donkey has four legs and calling a tail a leg doesn’t make it a leg.

53

And a raccoon can’t drive a car even if you take away the train conductors’ M&Ms.

54

First of all, you don’t play to maximize your chance of winning a hand; you play to maximize the expected value of your winnings, and if you go all in the instant you see your hand, everyone else is going to just fold, so it might be better to not bet as heavily on a sure-winning hand. And second, a hand of three kings does not have a 0% chance of beating a hand of four queens, because the other player might fold. If you “play the probabilities”, and do it correctly, taking these factors into account, then you will win.

Or, to put it another way: There are in fact computer programs now that will consistently beat humans. How can a computer’s play possibly not be rational?

55

No, you’re wrong. Three kings will never beat four queens. That will never happen in the history of poker.

What will happen in some situation is that a player holding four queens will fold against a player holding three kings. But that doesn’t mean the three kings beat the four queens. It means the first player folded when he shouldn’t have.

That’s the point I’m making over and over and over again. A poker based purely on mathematical probabilities does not include bluffing; be definition, it’s purely mathematical probabilities. That’s what those words mean.

If you claim you’re playing purely mathematical probabilities but you also sometimes bluff, then you’re using the wrong term. You’re played a mixed strategy that is mostly mathematical probabilities with some bluffing. Your strategy is not purely mathematical.

56

Not according to, you know, a dictionary.

Do you have an example of anyone else using the language in this way?

There are at least four posters here who think your own English mathematical usage happens to be less than authoritative. Something as simple as a mixed strategy Nash equilibrium, right on up to more complex concepts like semi-separating perfect Bayesian equilibria, all rely on some randomization of actions in order to cloud the information so that the resulting play is not too predictable. I have never once heard of these randomizations, or the resulting game theoretical equilibria, as being somehow outside of mathematics. What, exactly, is “not purely mathematical” about a simple mixed-strategy Nash equilibrium?

This cuts back to your bizarre statement earlier in the thread:

Players in a standard mixed strategy Nash equilibrium are assumed to be perfectly rational. Always.

This is directly from Nash’s famous original paper, when he proved the existence of an “equilibrium point” (his term) using mixed strategies for “highly rational” players. It is not some mysterious burst of irrationality, according to standard game theory, that causes players to randomize their actions. The solution – what we call today a Nash equilibrium – is interesting precisely because it is their very rationality which can demand randomized playing.

I don’t personally see how the single most famous result in all of mathematical game theory is “not purely mathematical”. That’s a new one on me, personally. Never heard that before.

57

I used to regularly play in a Dealer’s Choice game which included a guy who won a WSOP bracelet in Lowball. I had little experience at that game and was certainly outclassed by some of the other players when Lowball was chosen. Routinely sitting out rounds of a game a player didn’t like was frowned upon and there were always some very weak ‘donators’ in the game, so I studied up (Caro, Sklansky, Malmuth) and came up with a purely mathematical strategy for Lowball which worked quite well. Part of the strategy is to bluff with a pair of 4s or higher to give you the proper bluffing frequency.

It’s pure math, no skill, and showed a profit in games up to at least $20/$40 in public cardrooms, maybe higher with some very live players in the game.

58

It includes whatever I’ve built into the model I’m using to calculate probabilities. I could leave bluffing out of my model, but why would I? It’s not like it’s tough to model bluffing mathematically.

59

Yes, in the dictionary.

Of course I do. I’ve already quoted it. I’ll do it again.

There’s somebody saying you shouldn’t bluff. That’s somebody saying you should only raise when you have a stronger hand.

It’s not an issue about what this strategy is called. The issue is whether this strategy works. It doesn’t. Arguing about what its name is isn’t going to change that.

60

He’s saying not to bluff and only play with a strong hand as part of a “rudimentary strategy for the beginner.” I don’t think he’s saying that is a good strategy for the long term.

But until you understand the nuances of the game and have a lot of hands under your belt you wouldn’t know when a bluff might have a probability of being profitable. It’s not a good idea to just sit down and start bluffing at the wrong moments in your early days of poker.

You have to see a lot of hands - a whole lot of hands - no matter what books you read before you understand when and why bluffing is an important part of play and when the best moments to do it are. The safest way to not go broke during that long learning period is to play pretty much exactly like RickJay recommends. You won’t make a lot of money unless you just have a lot of luck but you probably won’t lose a lot either.

I agree in a sense it isn’t always mathematically “rational” for that moment in a vacuum. Some people bluff every time they get a certain hand, maybe 6 3 or something. By all probabilities that is a horrible hand to raise preflop. But it is a rational decision to work a bluff into their play on a regular but unpredictable basis like that. First of all everyone may just fold and give you their antes. but if you get caught making that bluff that is great - you lose a small raise and buy the doubt of every player at the table the next time you’re raising.

If it all works out someone who saw you get caught bluffing with a 6 3 three times during a session might come over the top with a huge re-raise next time you raise with AA. In the end all those little irrational bluffs are part of a rational mathematical strategy that pays off in the long run.