Tuesday, January 10, 2012

Bet sizing in NLHE

quick thought of the day based on this exchange on 2p2:

original post:
Stingers thoughts on value betting/bluffing released a gem of information, but i am having trouble processing it.
Can anyone provide some examples of what stinger was talking about with his thoughts here:

"If theres a spot where im going to be bluffing a lot, or at least my opponent perceives me to be bluffing a lot, you should be betting bigger. Thats why the over-bets are so good. Over-bets gives your opponent worse odds to call, which allows you to bluff more often, profitably...

...If its going to be a spot where you're going to be value betting often, and your opponent knows that, then you should be betting less to give your opponent better odds, tempt him into calling. The key there is all about what your opponent perceives. Its not about how much you are value betting or bluffing, Its how much your opponent expects you to be value betting or bluffing. You want them to be calling at a frequency that is profitable for you."


a response:
his reasoning is completely wrong, you could even say backwards, in both cases

a small bet should be made by a higher fraction of value hands than a big bet, yes. but that does not mean that you should bet big when you want to bluff a lot, or that you should bet small when you have a lot of value hands. not at all.


my response:
I agree that he hasn't full explained the concept but I wouldn't say he has it backwards. Basically there are 4 variables to consider:

1) your holding and it's relative strength to the board.
2) your opponents range and it's relative strength to the board.
3) your perceived range.
4) what you'd like your opponent to do with his hand.

If you know the answers to all of those questions (or can at least approximate them) then you can determine the best (or close to it) bet-size in relation to the pot to attempt to manipulate your opponent into the desired action.

I guess at it's heart it's basic theory and math of poker. But it is certainly always good to think about it and remember why we bet certain sizes in certain spots.


further thoughts:

When determining bet-sizing in NLHE first decide what you want your opponent to do (aka are you value betting or bluffing). Then determine what your perceived range on the board will likely be to your opponent (does he think you're likely to be bluffing often or do you have a very strong range that will nearly always be betting for value even if it's somewhat thin). Then size your betting appropriately to manipulate your opponent into the action you desire with your given holding on the board.

Example 1: You have the nuts on the river in a hand where your range is perceived as full of bluffs (aka it is a board you would frequently barrel). You want your opponent to call off as much money as possible. Size your river bet sufficiently large as your opponent will likely call on this river against your perceived range. If your opponent is particularly suspicious an overbet might have the most value.

Example 2: You have a weak holding on a board where your range contains mostly very strong holdings and your opponent's range is likely weaker than your perceived range. Something like you were the pfr in MP with 22 and your opponent CC in the SB and the flop came down AKKr. You'd like your opponent to fold and he likely will with the vast majority of his range. You should bet as small as possible to get the likely fold so that you'll save the most possible money the times he has a Kx or Ax hand and decides not to fold. If villain is the kind to call once and fold to future barrels you also create a situation where you are actually creating value by inducing single street calls planning on barreling most turns to get folds from his medium pair or QJ type hands.

I'm not going to edit the above and it certainly isn't a complete or exhaustive treatise on the subject. But I do think there is some value in what I wrote and if you're a new or struggling player I'd suggest reading it and digesting the thoughts and thinking of how it might apply in some further examples.

No comments: