Dont understand solver's flop bet sizing here - any thoughts? (1 Viewer)

boltonguy

Flush
Joined
Sep 5, 2017
Messages
1,986
Reaction score
1,748
Location
Boston, MA USA
Hero in HJ RFI 2.2BB (standard), CO calls, folds around - heads up to the flop.

1654642686205.png


Hero flops 2 pair on a very wet board. Hero bets 3/4 pot and V folds.
I was curious what the solver would do here and was surprised (see below).
I find myself in situations here sometimes torn between betting 1/2 pot and betting bigger to protect my hand and charge the draws.
Often when called V gets there and I have to give up = but I am way ahead here and V should not get there often.

1654642714307.png


With this exact hand (or Kh9h which is functionally equivalent) the solver is X 88% of the time here even though it has 71% equity.
When betting, the solver chooses 1/4 pot the most often.
Sure we are crushed by JT (16 combos) and KQ (6 combos) and 99 (3 combos) for a total of 25 combos but the balance of V's range is full of draws, diamonds, gutters and combos.

1654642794431.png


If Hero X, this is what V's range looks like.
Straight = 3.67%
Set = 2.88%
2 Pair (that beats us) = 5%
For a total of about 11%.

I guess I do think that I will get called by worse here or is my size just too big such that when getting called I am behind (only called by better at 3/4 pot)?
I could see betting smaller to keep V's calling range wider but X doesnt make sense, giving free cards to so many draws.
Thoughts?

1654643140953-png.926123
1654643359944.png
 
Snowie agrees with the solver = 88% X and when betting using 1/4 pot.
Someone help me here as I dont understand.

We are ahead here and can get worse hands to call. I dont understand the X.
Looking at later streets, looks like we're just check/calling down.
Maybe it's all those combos that beat us that drive the X?

1654643621040.png
 
I think the solver might be thinking like me here. The straight just calls (lets you trap yourself) and so does the flush draw (not bad odds to draw a diamond) while AK, KJ, and K10 like to come over the top of that trying to bet out other draws. Hands better than you here just call the solver bet while many worse hands try to raise here.
 
I know there are merits to playing small ball OOP, deep, on dynamic boards, but that’s a little ridiculous, I’ve been overplaying this spot egregiously

Interesting for sure
 
If I had to guess, it's because you unblock the nuts, and the villain ranges folds this flop a lot, or has hands that beat you, without a ton in between. You block TP, so the best continue hands that don't beat you are just pairs plus straight draws. Those hands can put you in tough spots and potentially blow you off your hand. So basically, all continues have you beat or have a lot of equity. And you basically can't get any value from under pairs and misses no matter what.

So I believe this is reflected by a check to protect your range and not get blown off the pot. I think there are probably way more spots where you should be checking as the PFR from OOP than you think.
 
Last edited:
If I had to guess, it's because you unblock the nuts, and the villain ranges folds this flop a lot, or has hands that beat you, without a ton in between. You block TP, so the best continue hands that don't beat you are just pairs plus straight draws. Those hands can put you in tough spots and potentially blow you off your hand. So basically, all continues have you beat or have a lot of equity. And you basically can't get any value from under pairs and misses no matter what.

So I believe this is reflected by a check to protect your range and not get blown off the pot. I think there are probably way more spots where you should be checking as the PFR from OOP than you think.
I read the question more of why does it check with 88% of range and the 11% it does bet goes for a small size. It’s definitely peculiar, small bets usually higher freq, lower freq usually favor large sizes. And OOP has big range equity advantage which should bump up the freq/size. I think it must be how dynamic the board is
 
I read the question more of why does it check with 88% of range and the 11% it does bet goes for a small size. It’s definitely peculiar, small bets usually higher freq, lower freq usually favor large sizes. And OOP has big range equity advantage which should bump up the freq/size. I think it must be how dynamic the board is
I would just discount the 11% bet small from the standpoint of how to actually play this. It still leans on the defensive given what the likely hands villain actually continues with are.
 
The ev difference of checking vs betting in the pic you posted is zero so it’s not like it strongly prefers it or anything.

Why you do so much checking here - you don’t have much of a range advantage if at all. You have the big sets but Broadways are more of his range, and you have many more hands that are complete misses.

Why you check this specific hand (and especially don’t go big), this is not a hand you’re wanting to go for 3 streets of value with large sizing. It would trim down too many of the weak hands by the river. You’d still get called by some worse on favourable runouts but (depending on sizing) maybe not enough to be pounding it this early.
 
This is not the type of hand you want to play a big pot with. And the types of hands you want to charge aren't going anywhere, and they will often bet when checked to anyhow. Pot control is wise here, and so is allowing your oppenent to try to bluff the pot. There are also a ton of scare cards that can fall on the turn or river. I see a lot of players overplaying two pair hands. Don't be one of them.
 
Tom Boshoff (aka tombos21) with GTO Wizard has a nice video dissecting why (to the best of his understanding) solvers recommend different bet sizes. Bear in mind that any rules or heuristics that humans create based on solver analysis is at best a crude approximation to what solvers actually do; solver logic is inhumanly fine-grained and comprehensive, perfectly covering every tiny corner case to exact percentages and is not determined by rules but rather by actual specific results for every possible situation.

Anyway... here's what the GTO Wizard people think is going on wrt post-flop cbet bet sizing:

First off: the most sophisticated but still simple-rules-based heuristic is: bet small on a dry board, bet big on a wet board, bet small on a very wet board.

1654692946476.png


Which gives rise to the "Wetness Parabola":

1654693100299.png


This is a reasonable approximation of what solvers recommend, but we can dig deeper into the details to get a better understanding of what the solvers are doing and to try to generate better heuristics that more closely approximate the solver solutions. If those heuristics "make sense" to us then we can say that we've discovered rules which "explain" the solver's behavior, i.e. we've answered "why" (to our satisfaction, to our best understanding) the solver chooses the way it does.

And the GTO Wizard people assert that the solver's reasoning looks like this:

1654693373381.png


The rest of the video explains what each of those factors means and why each factor contributes to making bet sizes larger or smaller, using charts and analysis from a solver (GTO Wizard, specifically) to make the case clear.

Here's the video:

 
Tom Boshoff (aka tombos21) with GTO Wizard has a nice video dissecting why (to the best of his understanding) solvers recommend different bet sizes. Bear in mind that any rules or heuristics that humans create based on solver analysis is at best a crude approximation to what solvers actually do; solver logic is inhumanly fine-grained and comprehensive, perfectly covering every tiny corner case to exact percentages and is not determined by rules but rather by actual specific results for every possible situation.

Anyway... here's what the GTO Wizard people think is going on wrt post-flop cbet bet sizing:

First off: the most sophisticated but still simple-rules-based heuristic is: bet small on a dry board, bet big on a wet board, bet small on a very wet board.

View attachment 926458

Which gives rise to the "Wetness Parabola":

View attachment 926462

This is a reasonable approximation of what solvers recommend, but we can dig deeper into the details to get a better understanding of what the solvers are doing and to try to generate better heuristics that more closely approximate the solver solutions. If those heuristics "make sense" to us then we can say that we've discovered rules which "explain" the solver's behavior, i.e. we've answered "why" (to our satisfaction, to our best understanding) the solver chooses the way it does.

And the GTO Wizard people assert that the solver's reasoning looks like this:

View attachment 926463

The rest of the video explains what each of those factors means and why each factor contributes to making bet sizes larger or smaller, using charts and analysis from a solver (GTO Wizard, specifically) to make the case clear.

Here's the video:

I like his videos and his explanations on why when and how
 

Create an account or login to comment

You must be a member in order to leave a comment

Create account

Create an account and join our community. It's easy!

Log in

Already have an account? Log in here.

Back
Top Bottom