GTO Mixed Strategies Frequencies

[buzzbuzz19]

I've recently started using more solvers to study and I realized I don't know why there are differences in how often you call vs raise (or fold) with a given hand. For example, I'm looking at GTOWizard free preflop solutions (MTT 40bb, CO open) and the BTN is 3Betting KQo 22% of the time and calling with it 78% of the time. The EV is the same for raising and calling, so wouldn't you want to mix 50/50? I'm sure I'm missing something here, just not sure what.

 

[LetterRip]

I've recently started using more solvers to study and I realized I don't know why there are differences in how often you call vs raise (or fold) with a given hand. For example, I'm looking at GTOWizard free preflop solutions (MTT 40bb, CO open) and the BTN is 3Betting KQo 22% of the time and calling with it 78% of the time. The EV is the same for raising and calling, so wouldn't you want to mix 50/50? I'm sure I'm missing something here, just not sure what.

The reason you mix at the given frequency rather than a different frequency, is that it changes your board coverage and strength of your ranges for 3betting vs flatting. If you raise more or call more, than your range after raising/calling will have different frequencies of having KQo in them. They can then exploit you on runouts that relate to hold KQo - ie boards that complete straights with high cards or that pair the K or Q. Say you 3bet KQo 100%, then they can exploit by folding more on Kxx and Qxx boards when you 3bet, and call wider on Kxx Qxx boards when you flat.

Also if you do this across multiple hands, you might end up with too many hands in your 3bet range, which allows them to 4bet bluff more, and trap more.

In practice as long as you are moving hands to simplified strategies as long as you balance it you probably won't be exploited by good players. Also good players are so rare, and population tendencies are to overfold and overcall and don't 4bet bluff, it likely won't hurt you to simply strategies to more aggressive 3betting.

 

[buzzbuzz19]

The reason you mix at the given frequency rather than a different frequency, is that it changes your board coverage and strength of your ranges for 3betting vs flatting. If you raise more or call more, than your range after raising/calling will have different frequencies of having KQo in them. They can then exploit you on runouts that relate to hold KQo - ie boards that complete straights with high cards or that pair the K or Q. Say you 3bet KQo 100%, then they can exploit by folding more on Kxx and Qxx boards when you 3bet, and call wider on Kxx Qxx boards when you flat.

Also if you do this across multiple hands, you might end up with too many hands in your 3bet range, which allows them to 4bet bluff more, and trap more.

In practice as long as you are moving hands to simplified strategies as long as you balance it you probably won't be exploited by good players. Also good players are so rare, and population tendencies are to overfold and overcall and don't 4bet bluff, it likely won't hurt you to simply strategies to more aggressive 3betting.

Ok that definitely makes sense, thanks. And in practice, especially at the micro/low stakes I'm playing, I would be very surprised if an opponent was able to pick up on me being a few percentage points off. I think my strategy at least for now is to split hands that mix to mostly 50/50 for simplicity and get more complex as I get better.

 

[LetterRip]

You don't really need to use GTO at low stakes at all. Start with a population exploit strategy for a table that is too passive (either tight passive or loose passive), then switch to a player specific exploit strategy.

 

[LetterRip]

Note a quick clarification - that when I said 'do this across multiple hands' - I meant hand combinations in your range, not multiple times you play KQo.

 

[alien666dj]

If the opponents also played according to GTO, then this could make sense otherwise it makes sense to adapt to a specific opponent when making a decision about the balance.

 

[NWPatriot]

... I would be very surprised if an opponent was able to pick up on me being a few percentage points off....

I agree with this completely. GTO solvers create un-implementable solutions. Their value is in seeing what type of strategies are un-exploitable, not in creating instructions for a specific hand.

I think we have taken balance too far. Say we take 100 hands and play a perfectly balanced strategy. If we get to showdown 5 times, our opponent will only "know" 5% of our strategy. There will be frequency information on 100 hands, but actual equity/hand strength information on 5 hands. We may have sacrificed value, in the name of balance for 95% of the hands.

All I am saying is, it takes quite a bit of time for a specific player to be able to understand what it is we are doing. We get clues about their understanding, when they start playing in a manner that goes against what we want them to do.

Play for "immediate EV" until we have been shown that we need to shift to "strategy EV".

Good luck and God bless.