Bet sizing question

[SeanC]

I'm reading NL HE Theory and Practice, and I understand the concept of maximizing EV when betting a hand you know to be best, but I was wondering if there's a simpler thought process that can be used at the table that the fully expanded process Sklansky gives? Sklansky's process is simple, but it can be hard to run through your head quickly.

By the way, the process covered by Sklansky is:

1. Put your opponent on a range of hands. Assign likelihoods for each holding.

2. Calculate EVs of various sized bets based on the chances of his various holdings. I'd start with big bets and move down on a standard list of something like overbets, pot-sized, 1/2 and 1/3.

3. Choose the bet with the highest EV.

This process makes perfect sense theoretically, but isn't it a bit cumbersome to apply at the table?

[wsock23]

At first, maybe. But after awhile it just becomes second nature. Also, the more hands you see a certain opponent the easier it becomes. Pay attention to everything he plays and calls down.

[Gonso]

Speaking for myself, bet sizing is obviously important, but I don't spend a lot of time working on the math but rather what I'm trying to accomplish and how it works with my style.

For example, when I make a RR from the blinds, I'll generally make it a little more than a pot sized since I'll have to play OOP if called. I'll do this when I'm restealing light as well as when I'm reraising a legit hand for value. To me, at least, it's far too complex to sort out the various EVs of various holdings and their raise sizes - and it's complicated further once you take the playability of certain hands into account.

I'm still trying to get the most EV by making these raises, but I'm more looking at the medium-term. I think it's wrong to view each hand in a vaccuum and evaluate the EV of each particular bet/raise. Some raise sizes are clearly suboptimal individually, no question.

Also, there are psychological and image aspects that bets or raises can have which wreak havoc on the math sometimes. For example, we've all been in spots facing a river bet where we fold to a "suspiciously" small bet, but would have called a larger sized bet... so it's not always a simple curve between EV and raise/bet amounts.

[SeanC]

[ QUOTE ]
Speaking for myself, bet sizing is obviously important, but I don't spend a lot of time working on the math but rather what I'm trying to accomplish and how it works with my style.

For example, when I make a RR from the blinds, I'll generally make it a little more than a pot sized since I'll have to play OOP if called. I'll do this when I'm restealing light as well as when I'm reraising a legit hand for value. To me, at least, it's far too complex to sort out the various EVs of various holdings and their raise sizes - and it's complicated further once you take the playability of certain hands into account.

I'm still trying to get the most EV by making these raises, but I'm more looking at the medium-term. I think it's wrong to view each hand in a vaccuum and evaluate the EV of each particular bet/raise. Some raise sizes are clearly suboptimal individually, no question.

Also, there are psychological and image aspects that bets or raises can have which wreak havoc on the math sometimes. For example, we've all been in spots facing a river bet where we fold to a "suspiciously" small bet, but would have called a larger sized bet... so it's not always a simple curve between EV and raise/bet amounts.

[/ QUOTE ]

Yeah, I agree that trying to constantly calculate EVs is futile. But let's take a situation where you have the nuts, or a hand that you know is best, on the river and you're going to value-bet it. You're fairly certain the villain has top or two-pair. Theoretically, you could run through a quick EV calculation on the bet, starting with the largest plausible bet size based on how he plays. You could then estimate the EVs of a couple smaller options. If you were quick, it probably wouldn't take more than 10 seconds.

It, however, gets tougher if you're looking at a more diverse range of hands. Let's say that, as far as you're concerned, there's a 30% chance he has a busted flush draw, a 50% chance he has TPMK or MPHK and a 20% chance he has a worse hand. This is a bit more complicated and impractical unless you want to sit for a minute before betting, haha. My guess, however, is that the great players have taken these long, drawn out EV calculations and have drilled in the process so extensively that the high EV plays come almost automatically, and they mainly focus on hand reading and psychological aspects. Do you agree?

The reason I'm asking is because my bet sizing has always been based on what I put the villain on, what he puts me on and then what the largest sized bet I "think" he will call. The problem with that method is that the "I think" part wasn't quantified and therefore the EV really isn't being maximized. Sometimes I'd get high, sometimes low--and I didn't know better either way.

I was tending to bet too large when I had a hand and when I was bluffing, which was prompting folds when I had something and occasionally snapping my bluffs for a big loss. I now see that his is a very unprofitable way to play, haha.

[PantsOnFire]

Reading and understanding NL theory is a good foundation. It gives you a lot of the "whys" when you make plays at the table. It also gives you some target goals for situations where you are not sure how proceed.

However, the real meat of NL play is experience and reads. Sure you have guidelines for your play but really the particular situation dictates things like bet sizing.

A good example is having the nuts on the river. You want to make a bet that will maximize your EV. So the range is a very small bet which will win you a little but very often to a very big bet which will win you a lot but not very often (and everything in between). You can assume a call frequency for different size bets and come up with an EV chart.

However, at the table, you need to take in the realities of the situation. Your image is important, your previous bet sizes on the river are important, your read on what villain might be holding is important.

So you have the theory in the back of your mind which is telling you to make a bet that has both a good chance of being called and is big enough to have maximum EV. It turns out that this bet size will be a product of your experience and instinct which is founded on your theory.

A beginner has no concept of this theory and just bets an amount that seems common (or the same amount every time).

A good player will know this theory and think about his bet size more carefully. He will be a little more if thinks he will get a call and a little less if he thinks he won't get a call.

An expert player will know this theory, have a good read on the tendencies of villain, have a good idea of villain's hand, have a good idea of villain's image of him, have a good idea of what villain thinks he is holding and make bet based on instinct and experience that is very close to a maximum EV bet.

[Gonso]

[ QUOTE ]
Let's say that, as far as you're concerned, there's a 30% chance he has a busted flush draw, a 50% chance he has TPMK or MPHK and a 20% chance he has a worse hand. This is a bit more complicated and impractical unless you want to sit for a minute before betting, haha. My guess, however, is that the great players have taken these long, drawn out EV calculations and have drilled in the process so extensively that the high EV plays come almost automatically, and they mainly focus on hand reading and psychological aspects. Do you agree?

The reason I'm asking is because my bet sizing has always been based on what I put the villain on, what he puts me on and then what the largest sized bet I "think" he will call. The problem with that method is that the "I think" part wasn't quantified and therefore the EV really isn't being maximized. Sometimes I'd get high, sometimes low--and I didn't know better either way.

[/ QUOTE ]

Well some of that is easy. For example, a busted flush draw won't call much of anything, so you can disregard that. All of the various marginal hands... it's not as if you can add up the various amounts they'd call for each possible holding, and just average it out into the perfect bet amount. In this case you're selling your hand, and again the likelihood of getting a call is often not proportionate to bet amount. A large bet might look more like a bluff and actually increase your chances of getting a call, for example.

Bet sizing has more relevance on earlier streets in terms of implied and reverse implied odds, overcharging for draws or to protect hands, etc.

[SeanC]

[ QUOTE ]
Reading and understanding NL theory is a good foundation. It gives you a lot of the "whys" when you make plays at the table. It also gives you some target goals for situations where you are not sure how proceed.

However, the real meat of NL play is experience and reads. Sure you have guidelines for your play but really the particular situation dictates things like bet sizing.

A good example is having the nuts on the river. You want to make a bet that will maximize your EV. So the range is a very small bet which will win you a little but very often to a very big bet which will win you a lot but not very often (and everything in between). You can assume a call frequency for different size bets and come up with an EV chart.

However, at the table, you need to take in the realities of the situation. Your image is important, your previous bet sizes on the river are important, your read on what villain might be holding is important.

So you have the theory in the back of your mind which is telling you to make a bet that has both a good chance of being called and is big enough to have maximum EV. It turns out that this bet size will be a product of your experience and instinct which is founded on your theory.

A beginner has no concept of this theory and just bets an amount that seems common (or the same amount every time).

A good player will know this theory and think about his bet size more carefully. He will be a little more if thinks he will get a call and a little less if he thinks he won't get a call.

An expert player will know this theory, have a good read on the tendencies of villain, have a good idea of villain's hand, have a good idea of villain's image of him, have a good idea of what villain thinks he is holding and make bet based on instinct and experience that is very close to a maximum EV bet.

[/ QUOTE ]

Thanks for the reply and I totally agree on all points.

I'm going to have to play around with quite a bit of EV calculations so the decisions will come quicker at the tables.

[SeanC]

[ QUOTE ]
[ QUOTE ]
Let's say that, as far as you're concerned, there's a 30% chance he has a busted flush draw, a 50% chance he has TPMK or MPHK and a 20% chance he has a worse hand. This is a bit more complicated and impractical unless you want to sit for a minute before betting, haha. My guess, however, is that the great players have taken these long, drawn out EV calculations and have drilled in the process so extensively that the high EV plays come almost automatically, and they mainly focus on hand reading and psychological aspects. Do you agree?

The reason I'm asking is because my bet sizing has always been based on what I put the villain on, what he puts me on and then what the largest sized bet I "think" he will call. The problem with that method is that the "I think" part wasn't quantified and therefore the EV really isn't being maximized. Sometimes I'd get high, sometimes low--and I didn't know better either way.

[/ QUOTE ]

Well some of that is easy. For example, a busted flush draw won't call much of anything, so you can disregard that. All of the various marginal hands... it's not as if you can add up the various amounts they'd call for each possible holding, and just average it out into the perfect bet amount. In this case you're selling your hand, and again the likelihood of getting a call is often not proportionate to bet amount. A large bet might look more like a bluff and actually increase your chances of getting a call, for example.

Bet sizing has more relevance on earlier streets in terms of implied and reverse implied odds, overcharging for draws or to protect hands, etc.

[/ QUOTE ]

I agree that the expanded EV calculations aren't workable at the table, but I'm sure the general process can be simplified. The right bet is all situational, right? If a guy is a habitual bluffer, check-calling a winner would be higher EV than leading out and making him fold his crap; if he's tight, betting the right amount that can help him make bad calls would be higher EV, etc.