Varying bet sizes

[ACW]

I've frequently seen people state that you should always bet / raise by the same amount to avoid giving a tell by the size of your bet.

While this may be a sensible practical approach, it seems highly sub-optimal to me.

Firstly, it is surely correct to allow public information (such as position, number of players in the pot, board texture, etc) to influence bet size. I suspect many of the "always bet the same amount" crowd would agree with this.

I'd go a step further and say that you SHOULD allow your hand to influence your bet size. Clearly when betting a draw into a large field, you want them all to call. When betting the nuts, you want callers. When betting a hand which is ahead but vulnerable, you want to charge the others players as much as possible, or have them fold.
So with some hands it is correct to bet small, with others correct to bet big.

Now we have to deal with the problem this poses in that everyone knows what we have - so we must mix it up a bit.
The situation is surely analagous to the classic game theory example of how often to bluff as a function of bet and pot size. When we bet small, it will be optimal to have a draw a% of the time, the nuts b%, a good but vulnerable hand c% and a bluff or semi-bluff d% of the time.
Similarly we could assign percentages for hand types for a pot size bet, on overbet, a check-raise, check-call.
I don't believe the optimum solution to this problem is "always bet the same amount regardless of your hand".

Has anyone ever seen anything written on this question? Any ideas how often you should have the clearly implied hand and how often a different hand type? I'd guess you should bet the hand in the optimal way something like 70-80% of the time, and throw in different bet sizes around 20-30%, but I could be way off.

[Kopefire]

The "always bet the same" crowd gets the advice from T.J. Cloutier . . . who, if you watch at all, varies his bet size.

Frankly, I think not using the fact that you can bet any amount you want to over the required minimum in NL more or less gives up one of the huge weapons available to you.

I strongly suspect that Cloutier gives this advice out in the hope that lots of people follow it, and then come play in tourney's that he's entered in . . .

[jayrutz2]

The way understood this is that it is primarily applicable to opening, before the flop raises, especially in tournaments vs ring games, and more so in early to mid rounds than later rounds, depending on your stack size, on the theory of eeping opponent guessing on anything from pocket 2s to AKs, AA or even lower suited connectors. This allows you to adjust your action both in subsequent preflop betting and post flop rounds.

Comments?

[pzhon]

Mike Caro suggests betting the strength of your hand on average, making random variations around the true strength. This is wrong, but it may be less wrong than betting the same amount each time.

[Dov]

I tend to vary my bet sizes according to the reads I have on the player(s) in the hand with me. I know what size bet player A will call with or fold, while player B may fold to any bet over a certain size. I will keep them in their comfort zone when I want a call, and make them uncomfortable to varying degrees when I either want them to fold or just prefer that they fold.

I tend to bet in such a way as to try to get the pot HU with whoever I deem most vulnerable to me in that pot. If I can't do it, I sometimes have to give up the hand or take a larger risk than I would have liked.

If I have too much trouble with this, I usually decide that the game is too tough for me and move on. I like to own the tables I play at.

Dov

PS - This discussion is obviously about NL.

[patrick dicaprio]

i think the point is that you certainly shoudlnt routinely do things like raise 4BB with medium pairs but raise 3BB with high pairs and mnraise with flush draws or something. one thing that can work is varying your bet sizes at random, utilizing a randomizing device such as the cards or the secondhand on a watch.

please note that i havent tried this, but i have been thinking about whether this could work if you play in a game with the same players all the time such as a home game. if i come up with a system and try it out i will let you know.

Pat

[ACW]

The principle still applies preflop - with drawing hands you'd like to build the pot while keeping people in, but with hands like QQ you want to get heads-up or buy the pot right there. Also you'd rather charge more for a call with JJ than with AA.

It would seem optimal to generally use appropriate bet sizes, mixing it up just enough to create uncertainty in the minds of the opponents.

[golFUR]

There was something about this post that didn't quite sit right with me. It took an analogy to make it clear.

(I'm going to end up formalizing this preface... First, I play at UB, home of the 'bet pot' button, that colors my commentary a bit. Second, as I've said elsewhere (but don't expect anyone to necessarily remember) I'm more of the school of 'poker by feel' than 'poker by numbers'. Find my other posts if you are interested...)

Poker is 50% cards and 50% the people playing them. I didn't get that from anywhere, I made it up, feel free to substitute your own numbers in. The point is, it is so much easier to study the card side of it, to apply numbers and formulas and such, that often the people side of it is either neglected, relegated to another forum or worse, subsumed into the numbers debate. This is what it seems to me you've done.

[ QUOTE ]
When we bet small, it will be optimal to have a draw a% of the time, the nuts b%, a good but vulnerable hand c% and a bluff or semi-bluff d% of the time.
Similarly we could assign percentages for hand types for a pot size bet, on overbet, a check-raise, check-call.
I don't believe the optimum solution to this problem is "always bet the same amount regardless of your hand".

[/ QUOTE ]

This is akin to painting by numbers. "In Picasso's book it says that any Fall Landscape painting should be minimum 30% red ochre and 20% yellow safron. What do I do when I want to incorporate a lake though? Can I use Dali's formula for reflections to cover some of those requirements?"

Simply first, yes, betting the same amount isn't any good. If I raise 3.5BB in late position with my pocket Ts, play the hand out and showdown my Ts I have no cause to be surprised or upset when 15 hands later my pocket Ks lose to 9s or an AQ if I raised the same preflop. Your opponents are trying to read you the same as you are them, if the only thing they have to go on so far is "3.5BB roughly equates to Ts" you are begging for trouble if you raise the same later with 6s or Ks.

So we acknowledge that varying the bet amount is both useful and necessary. It allows us to put information that is beneficial to us out there into play. If we vary the amounts semi-randomly we can maintain the obfuscation that 'same size bets no matter' hopes to gain for us. This decision, to vary the amounts of the bets based on pocket, position and a semi-random function, while couched in the language of the numbers game is more realistically a function of the people game. When we choose to start playing 'creatively' we have to finish the plays that way as well. Which is where the problem comes in in this case.

You seem to want a formula for how to finish hands based on how they were begun. Maybe I am misreading a bit. Let me know. I am certainly detecting too much of one language in a question that belongs in the other though.

From another tack then...

[ QUOTE ]
Clearly when betting a draw into a large field, you want them all to call. When betting the nuts, you want callers. When betting a hand which is ahead but vulnerable, you want to charge the others players as much as possible, or have them fold.
So with some hands it is correct to bet small, with others correct to bet big.

[/ QUOTE ]

To an extent, yes, when betting a draw to the nuts into a big field you want callers and you don't want to invest too much in case it doesn't hit. Who says you have to hit to take it down though? This is the leap it seems needs to be made. By going the extra step into the 'creative' side of the game and playing around with different bet amounts you are playing around with the people as well. You've gone through the numbers and out the other side. Why give up there? Why immediately turn back around and ask for more numbers on what to do now? You are manipulating them with your preflop bets, keep it up!

[ QUOTE ]
Has anyone ever seen anything written on this question? Any ideas how often you should have the clearly implied hand and how often a different hand type? I'd guess you should bet the hand in the optimal way something like 70-80% of the time, and throw in different bet sizes around 20-30%, but I could be way off.

[/ QUOTE ]

I'd guess you should bet the hand the proper way at the proper time and only find out if that means 70/30 or 30/70 in retrospect. To attempt to plan your game in this much detail is losing the forest for the trees. To sit and keep a tally of how many multiway pots I've been in with pockets of range X, to 'plan' on bluffing x% of the time... surely this is giving up a great deal more money than failing to vary your bets by a few tens or hundreds of chips here and there. Just because your draw missed doesn't necessarily mean anyone else's hit. If you decided, for whatever reason, that this was a good place to get tricky and raise preflop with just a suited A, and you then missed the suit and missed the A... who says the play is over? Continue reading the board and more importantly the players and their reaction to it, the hand might not be over...

I'm afraid this isn't as coherant an answer as I intended. There isn't yet, or I haven't found yet, the right language and terms I need. I believe the painting analogy is useful though. I can reduce a painting to pages of detailed information chock full of numbers. I can reduce it to diagrams full of angles and lengths, ratios of colors to open space etc. This doesn't change the fact that fundamentally, painting is an artistic endeavor dependant far more on creative input than numerical clarity. The output as well, like an elegant chess solution or a monster trap paying off, can be a numerical summary but is more appropriately an artistic expression.

[ACW]

Thanks for the comments - you've clearly put a lot of thought into this.

I agree with the point you seem to be trying to make - that the question of bet sizes is less important than being able to use good judgement, and the need to follow through on later rounds. I wasn't trying to diminish those factors at all.

When you make a small raise from late position with a suited connector, and three of a different suit flop, clearly the fact that you advertised a drawing hand could enable you to steal the pot if it's checked to you. And yes, this may take two shots to fold everyone, and if a fourth suited card hits, you may have to give up. There's no substitute for good judgement and a good read on the opponents in these situations.

I wasn't thinking in terms of "What formula should I use?". Rather I think this situation is analagous to the heads-up bluffing situation where game theory dictates the optimum bluff ratio in the absence of a read. Of course, if the opponent is a calling station, the optimum suddenly becomes to never bluff, and if he's weak tight, it becomes more attractive to bluff.
Similarly, if I know that, in the absence of other information, it's correct to pot bet with the nuts 70% of the time (for example), it doesn't change the fact that at a table full of calling stations I should increase that to 100%. If the table is full of players who will fold to a bet but attempt to steal if checked to, then the optimum could in extreme cases change to always checking. It does help though to know where the starting point is before making the player dependent adjustments. That's what I was trying to get at.

[golFUR]

Okay, yeah, I may have answered a slightly different question. When you are chasing a grail you tend to see cups everywhere...

[ QUOTE ]
It does help though to know where the starting point is before making the player dependent adjustments. That's what I was trying to get at.

[/ QUOTE ]

I think I came close to this in my first post though, and depending on where you are in your game, you might have the answers already in PokerTracker. I would tend to want to avoid someone else's idea of what is a good percentage until I determined through play where I was already at. If I found in analyzing my own games that I was nearer 50% while some pro had set the bar at 70%... It might just be a style thing that I have no desire to change. Maybe I want to put too much intuition into the game but I think the concepts we are discussing here are rather style dependant and player dependant. We can state rather emphatically the odds of the case J coming out on the turn, there isn't really any argument over how to arrive at that number, it is the same number no matter who has Js in the pocket. How often should *I* bet pot with x pocket on x board though? That is a much different question, it isn't straight math but depends as much on my reads as how others are reading me, it is player dependant.

So, analyze your own game, decide whether or not you are 'comfortable' with where you are at now, and then adjust where necessary.

Something I tried may work for you as well. When I first started learning the basics of poker the style that resulted was rather predictable. I was a rock. I learned about preflop hand selection and protecting your pockets and position all at about the same time. Half way through a SnG with me you'd have me perfectly pegged. If I raised in early position it was guaranteed to be a big pair or suited slick, never ever a small pair or offsuit blackjack. Not to say that game didn't work, it did, it made me good money, but I wanted to learn more.

Even if it ended up that that was my game for good, I figured the only way to learn to play against maniacs was to play as one for a bit, to learn the way they think. So I created in my head a few different 'styles' with different rules for each. I'd resolve at the beginning of the SnG to play one style or another. The maniac style was great fun btw.

After a few months of this I found myself comfortable switching styles at will, able to recognize other styles played against me etc. That fluidity in my game extended down from the macro to the micro, I gained a better understanding of each play from each position. So when asked about %s over all, from my perspective, the immediate question in return was, 'for what style'? How do you play? If you are a maniac you'll be bettin the hell out of everything, if you are a rock you won't... Player dependant statistics as opposed to adding up missing cards and doing a little arithmatic.

[Dov]

I'm with Golfur on this one. Once you know the theory and get some experience, you should definitely experiment to understand the other players in the game and the strengths and weaknesses of different styles.

As a simple example of this, play a game of chess with someone, and then switch colors halfway through the game. You will be amazed at the holes you see in your own defenses and the strength of the opponent's position. What you thought was obvious isn't necessarily.

[ACW]

[ QUOTE ]
I'm with Golfur on this one. Once you know the theory and get some experience, you should definitely experiment to understand the other players in the game and the strengths and weaknesses of different styles.


[/ QUOTE ]

I agree. It's probably something I should try again - I haven't done this since reading "Killer Poker". In this Jon Vorhaus recommends trying out a variety of styles to see how others react when you're maniacal, and build discipline when playing as a rock.

On the rock side, he advocates small poison - not playing any card under a 9, no exceptions. I kept it going for about an hour, picking up small gains (I rarely got called), then decided that completing to a single raise from the BB with pocket 7's was justified. I lost all my earlier gains and then some to a set over set. I almost felt it was punishment for my lack of discipline!

Another style that's great fun is to always raise the first hand posted by a new player at the table. This worked well for me at limit - I haven't tried it at no limit.

One I've seen someone else try is to turn up and burn two buy-ins through suicidal all-ins then turn into a solid rock. It's amazing how many people still had him down as a nutter and called him down with second pair even once it seemed obvious what was going on. It worked well for him.

P.S. Thanks for the great responses! It seems the answer to the original question is as I suspected - it depends!

[Dov]

I also enjoyed Killer Poker. The chapter on the Antifish concept contains, at least for me, the secret of poker. Keep them always ucomfortable and guessing.

Good post. I hope I was able to help a llittle.

BTW, I do that raising your own blind in NL all the time. It works better at a tight table where everyone is already familiar with the styles of the others and have already established a rhythm. When you upset the rhythm, it takes them a while to readjust. Some of them don't do it in time and lose their whole stacks - usually to me /images/graemlins/laugh.gif