Bet Sizing Theory (Part 2)
Posted by Kevin Rabichow
Video Player is loading.
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Resume Video
Start from Beginning
Watch Video
10
Loading 11 Comments...
This is great! Please keep going in terms of producing this type of content.
The point about not only looking at all the bet sizes together and eyeballing what looks best aka frequency != EV is something I have not seen formulated clearly before but it does clear up a lot of confusion.
For example when I look at the PIO heatmap I look at what cards are getting more betting volume (bet size X bet frequency) and have been kind of implicitly assuming that these cards are "good" for the bettor's range. But recently I came across someone stating that bricks are good for the previous street defender (the more bluffcatchery range) rather than for the previous street aggressor (the more polar range). And in terms of EV totally true. But in terms of betting volume going back to the PIO heatmap often the opposite, which for me was a change to my thinking process.
Maybe that's a good idea for a video to look at common misapplications of the statement "this card is good for our range" in coaching materials because I do think often what the coach means is "we can bet a lot and big on this card".
Thank you for the feedback! I like the idea of extending this to the concept of "good for our range"
How do you come up with .1 in EV equates to 1bb/100?
This simulation is for a $10bb game, where the pot size ($52) is a 2.6bb open and call. When strategy EV changes by $0.1, it's changing by 1/100bb in a single hand, or 1bb/100 hands.
24:10 - Why does the solver change it's cbetting strategy for IP when we remove OOP's option to probe anything other than 30%? How can a node change based on another node that only comes later in the game tree? This is to say that the solver constructs it's flop strategy based on a read on it's opponent's turn tendencies..
Anyway, great video! The way you explain it's very clear and easy to understand. Thanks!
Yes, the solver absolutely looks into future street options! It's important to think of the bet sizes we set as restrictions for pio to optimize around, regardless at what part of the tree they occur. Thanks!
My guess would be that the solver changes its cbetting strategy to betting more frequently because its strongest hands have less incentive to check back when less money is being put into the pot OTT when OOP is restricted to using this smaller sizing...since we're 'frontloading' most of our strongest hands into a flop betting range we're going to be allowed to supplement that betting range w/ more weak hands -> our cbetting frequency increases
Yeah but how can he look into Villain's future street options? That would be the exact opposite of what a solver's function is, which is to find the equilibrium that creates a perfect defense against any possible strategy.
Sorry, I'm not totally sure this is what you are asking. Any solver I'm aware of has limitations like this, because it must be restricted to certain bet size options in order to function on a normal computer. The solver learns through iteration that its "opponent" only uses certain bet sizes, which means it can "adjust" to the set of options that have been provided.
would it be possible that the changes impacted on IP Flop cbetting frequency come from the parameter box used for "probe" being the same that will be used for OOP Turn barrel sizing after him XRaising Flop? (thus making it correct for IP to cbet more on the Flop given that after facing a Flop XR he will then only ever face a small Turn barrel). aren't we tampering with that line as well each time we look at changing "probe" parameters?
Yes, definitely. This is something that makes these experiments very tricky, and often somewhat inconclusive. A couple workarounds - we can head to that node (facing flop xr) and look at how much strategies have actually changed, or we can ensure that we're leaving a 2nd bet size option that's more appropriate for the scenario you're concerned about.
Be the first to add a comment
You must upgrade your account to leave a comment.