Quote:
Originally Posted by robert_utk
If you build a balanced strategy for one bet size, what purpose is served by a second bet size?
What is the purpose of your bet? One size or the other will accomplish your intended result more often than the other.
There's a modicum of additional EV to be gained by having multiple bet-sizes in each spot, since some combos would do better with one size than another.
A common example comes up in short-stack situations in SNGs/tourneys. There are spots where it might be profitable to open jam over 50% of hands, expecting to get a fold, but if you have aces or kings you don't want to get a fold. You'd generate more EV with your best hands by limping or minraising to induce a shove. If you're minraising some monsters to try and get action, you'd also need to minraise-fold some hands, while others (e.g. small pairs) might do best by open jamming. Hence the best short-stack strategy is one that features open shoves, limps
and minraises. (3 different bet-sizes). Cash games with deep stacks are much more complicated, because the shoving option isn't there, but it stands to reason that some hands play better in bloated pots, while others do better in medium or small pots, so there may be
slight EV gains by c-betting 4.5bb with some hands and 2bb with others.
The problem is that each bet-sizing range needs to be "balanced" appropriately, which means some/many combos might need to be in every bet-size range, in order that each bucket isn't face up and exploitable. e.g. If you always went big with your nutted hands on Q72, then villain would know you are capped/weak when you bet small, so to avoid exploitation you'd have to use that smaller size for stronger hands sometimes too. If you're using multiple sizes, and mixed strategies, it's incredibly difficult to compute - let alone remember - the optimal frequencies.
I mean, hero in CO has about 300 combos on Q72tt. I'd be more concerned with working out which combos should
generally be bets and which should generally be check backs, instead of trying to calculate how often to bet large and how often to bet small with each of those 300 combos. (I can't store 600 numbers in my brain for every possible flop!). Giving yourself two sizing options doubles the immediate complexity of the game tree and presumably doubles your chance of making a mistake. :/
I could be wrong, as I haven't looked at solvers to see how much the sizing can alter the EV of each combo, but I don't think it's really necessary to study range-splitting in depth unless you're playing in extremely tough games, where every fraction of a bb is important.