No announcement yet.

Postflop Raise Sizing

  • Filter
  • Time
  • Show
Clear All
new posts

  • Postflop Raise Sizing

    Hello all,

    I'm working on developing a strategy that is based on GTO principles but simplified to a pure or binary strategy whenever it is possible to do so without losing too much EV. Think of finding spots where a solver plays a mixed bet sizing strategy on the flop, but the EV of playing a 100% 1/3 pot c-bet strategy is only a few basis points worse and, thus, better for most of us mere mortals.

    This is easy enough for many flop situations (hint: 1/3 pot bets crush!). But what about postflop raises? I'd love to identify an "optimal" raise size that might not be ideal for every single situation, but is usually good and never that bad.

    Does anyone have a better feel for this than I do? I've been running PIO sims with 60% pot raises, but it seems that this results in smaller raises than I normally see in game.

    Of course, the answer might be that it isn't possible to use a single postflop raise size without giving up substantial EV. But if this is the case, do any of you know how to quantify the EV loss?

  • #2
    Have you watched JL's master class? In this, he explains very thurl.


    • #3
      I have, and I understand that having different raise sizes in different situations is ideal. I guess what I'm wondering is (1) if you had to choose only one raise size to use in every situation for the rest of your life, what would it be; and (2) how much would sticking to one size in all spots reduce your EV?

      The reduction in EV would presumably be somewhat offset by being able to better implement a simpler strategy. (This is what I'm really after).


      • #4
        sounds awfully advanced. I don't even have PIO.

        about that simpler strategy, u sure it will ever be simple ?

        why would anyone want one size fits all ?


        • CrazyEddie Reloaded
          CrazyEddie Reloaded commented
          Editing a comment
          do u mainly play cash or mtt ? I thought pio is about different sizing & frequency... let's say u got down the sizing part, what of frequency ?

      • #5
        I play mainly live cash, but some MTT. You're totally right -- I'd still need to memorize/conceptualize frequency, and it will never be that simple. Poker is a hard game!

        There are two main reasons why I'm seeking a single raise size. First, I want to be able to run PIO sims that don't take overnight to solve. Even if I'm studying a flop spot, I need to provide the sim with raise sizes for all streets, and the more raise sizes I enter, the longer the sim takes to solve. So I'd like to be able to enter one raise size (or two: one for IP and one for OOP) that will be "good enough."

        Second, even though I am a GTO nerd and love studying the stuff (although I'm a novice compared to many), I'm a firm believer that it's easy to get carried away with the theory and forget about what's best in practice. Jonathan talks about this a lot too. GTO solvers will prefer a mixed strategy in most spots, but frequently the EV gain from playing such a complex strategy will be outweighed by the EV loss of not implementing said strategy perfectly. I believe that for the vast majority of players, the optimal strategy will be grounded in GTO principles, but also heavily simplified to eliminate options that only increase EV slightly but, cumulatively, increase complexity greatly. Add in exploitative adjustments, and that's the strategy I'm going for! (Easier said than done, unfortunately).

        In summary: I think you're probably right, Crazy Eddie. But I wanted to wrestle with the question and hear other people's thoughts, so thanks for your response!


        • #6
          If i had to choose one Size with more than 40 bb effectively :

          In Position 3x
          Out of Position 3.5x

          All raises less than 3x did not win out After solvers became more and more popular

          There was at least in online mtt a time for each of the following sizes to be standard/default where

          a) clicking it back or
          b) 2x or

          But nowadays 3x ip and 3.5x oop is preferrable
          Last edited by Guido; 01-05-2020, 03:26 PM.


          • #7
            couldn't agree more with " best in practice " ..

            I'm GTO illiterate .. curious what raise size you use here / you probably have good approximation.

            Hero open from utg 2.25x with JTo ( 98BB ) CO flats with 100BB, Both blinds call. 32BB , 98BB

            Flop: J T 2 ( rainbow )

            Big blind donk 2/3 pot.

            This is MTT hand. Notice there is a short stack in SB, and maybe a side pot brewing.

            Would u have one raise size if there was flush draw on flop ? Would PIO ?


            • #8
              One limitation of PIO is that it doesn't simulate multi-way spots. But I still think it's a good tool for analyzing these spots -- you just have to be aware of its limitations, tweak ranges accordingly, and use some poker common sense when analyzing the results. In your example, you'll want to play more conservatively postflop than you would if it were heads up, since you're against two additional ranges (and so is villain).

              I've given BB JL's BB vs BTN range from his preflop chart PDF, to reflect BB's presumably wider range due to great pot odds. I'm assuming that BB will squeeze linearly with only AQ+ and TT+. Of course, this assumption won't hold true for everyone, but it accomplishes what we need it to for the purposes of the sim: simulate BB's capped range. I gave you JL's CO RFI range.

              First thing first: BB really shouldn't have a leading range here. He has massive range and nut disadvantage, because he defended super wide and presumably doesn't have overpairs or top two sets. So right away, I'm in exploit mode. Speaking for myself (not PIO), I'm pretty content to sit tight IP and let him blast off to me. I'm assigning him a range of TPTK+ and OESDs, meaning I'm beaten by only a set of deuces. He's not giving himself a fantastic price on his draws, and his sizing is large enough that if he keeps blasting, I can get all in by the river if I want to. I can also (probably) avoid getting stacked if he does have the set, or if his draw comes in.

              But I'm glad you gave me the HH, because PIO completely disagrees with me (as it often does). I gave IP the option to raise 40%, 60%, or 80% pot (approximately 2.5x, 3x, and 3.75x), and PIO gets savage, using mostly the large sizing and raising as thin as T8 for value/protection (mixing in some folds too. Now that's thin value!). Of course, this assumes that villain has the leading range I've assigned to him, and that, after making the massive fundamental mistake of leading this flop, he will now play like a killer robot super genius for the remainder of the hand. Not a good assumption in real life!

              Another possibility is that villain has a random jack and is trying to "find out where he's at." This is another exploitative reason to slowplay.

              I'm still running the sim (multiple raise sizes will do that!), but so far it looks like PIO is raising and going to the mat with overpairs and better (but flatting many QQ and some KK). It's raise/folding a lot of top pair combos, but flatting a lot of them as well (especially TPTK, which needs less protection and doesn't block bluffs). It's protection raising an assortment of tens, and mixing call and raise with straight draws.

              My guess is that PIO might raise a bit smaller with a flush draw on the flop, because it isn't afraid of them like we humans are and wants to keep them in while extracting max value. But this depends once again on what leading range we assign to villain, since PIO would never lead the flop. If we assume that villain is leading many/most/all flush draws, I'd expect PIO to raise a medium size with a super thin and linear range. I'll run the sim when this one's done!

              Thanks for the HH. Solvers are fun!


              • jamtay317
                jamtay317 commented
                Editing a comment
                Monkersolver does do multiway

            • #9
              I like to raise those donk bets anyway: 2 pairs was just bonus.

              Another PiO question: this one also multiway , not as multiway

              MTT hand:

              Hero limp UTG + 1 with Kc 3c ( 61BB ), HJ limp behind ( 27BB ) , Big blind chk ( 15BB )

              Flop: Ac 8c Kd

              Hero bets 1/3, called by both;

              Turn: 5 h

              Pair & flush draw on flop ... We're at the turn already. How to structure turn & river bet sizing / frequency on various board run outs ?
              Last edited by CrazyEddie Reloaded; 01-10-2020, 02:08 AM.


              • #10
                Can you give me a sense of what your limping range looks like? My guess is that villains have enough Ax in their ranges that your specific hand is a check on turn and river, but I could be wrong and I'd be happy to run a sim.


                • CrazyEddie Reloaded
                  CrazyEddie Reloaded commented
                  Editing a comment
                  He does have Ax , of course...

                  No idea what my limping range looks like ... Big blind is short ... I limp pretty pretty wide so he can't go desperado

                  and it depends on who was in the BB etc etc .. no way to recall
                  Last edited by CrazyEddie Reloaded; 01-10-2020, 07:47 AM.