How do you play KK/QQ on the flop and turn?

If we do enough random things, no one will ever hand read us. That's for sure. Just to be clear, our plan is to flat our best hands and take the bottom of our range and cap "for marginal value plus trickiness"? You really think that people are giving tons of credit to unknowns, 3 betting them light, and then folding easily once they play back?

You're just making the game harder for no reason. We don't have to give a good player the best hand, position, and a bloated pot as a gift, and then plan on trying to outplay him. The actual hand here doesn't matter. The huge leak is in approaching poker this way.

I haven't commented back, but while I like the original play, I do fully agree with the other posters. If we're 4-betting OOP w/ 98s preflop, then we should be 4-betting our big hands (JJ+,AK) as well, and this flop is a perfect illustration why. We get a flop that should be good for our range, and our opponent raises anyway. What does that say about his holding?

If he's willing to give an unknown's 4-bet range action on this board, then we would be able to win a big pot when we do have *it* (it being AA, maybe 99 here).

You're probably right...as a basic strategy, this is the definition of fancy play syndrome. It's easier to just call the 3-bet with your whole range if you want to disguise the strength of your hand. And hey, lookie there, you get to save a SB (or two if you get 5-bet) with your drawing hand.

You're allowed to build a balanced calling and 4 betting range, too. You can actually figure out the right proportions and build a range. My issue is that he's playing backwards on purpose and hand waving that it is clearly a good way to play. Even the deception may be gone based on Jon_locke's comment about the Roy Cooke article -- we expect an unknown to play any way but well.

In an uncapped HU pot OOP, what does a balanced 4-bet range look like against an assumed solid player (based solely on stereotype of visual appearance)?

I can tell you in NL.

Sometimes it seems that GTO should be thought of as a model...of Pontiac. Easy to misapply.

Assumption #1: Raising 89s PF UTG 6max is +EV. debateable, but not outrageous.

Once villain 3b's us, there are three prevailing theories:
1) Raise with the strong portion of your range (JJ+, AK) - the old school philosophy.
2) Flat everything - the common 2+2 philosophy
3) Raise a 'balanced' range (some weak suited hands, some strong hands).

Assumption #2: Theory 3 can be +EV. Maybe neutral EV, at worst just slightly -EV. This comes from my observation of bots at 3max, who do have a reraising range OOP. It's possible due to tighter ranges in 6max, that it is never justifiable to 4b. I find this overly narrow thinking.

Assumption #3: 89s is would be part of the balanced 4b range. No, I don't want to construct a full 4b range, and a calling range. My assumption is based off of what bots reraise with: strong hands, and smaller suited connectors, which aren't dominated, and benefit from extra FE. Yes, this is hand-waving, but it is based off of meaningful heuristics.

Doug suggests raising 89s here is me being 'random' for deception. Lilly asks why I don't 4b 55-77, or TQ. Choosing to 4b 89s isn't random, and my heuristics deem 55-77 and TQ as poor candidates to 4b. Yes, they can deceive, but they suffer worse -EV consequences as a result of the 4b pre. 55-77 are showdownable; TQ is often dominated, meaning we may commit ourselves to put more money in bad. Again, I don't think these are outrageous assumptions.

The above assumptions go to show that a balanced 4b range can include 89s and premiums. The next step is the 'opposite man' step. This is where the GTO disintegrates, and I switch to an exploitative adjustment.

Assumption #4: Let X be a hand in range Y; The profitability of X doesn't depend on our actual range Y, but merely on what villain perceives our range to be (say Z). As a basic example, say villain thinks our PFR range is exactly {AA}. Then clearly we can raise whatever junk we want, and villain will fold. It doesn't matter what our actual range is, only what is perceived.

We gain deception (and therefore EV) if villain overestimates or underestimates our ranges. I posit that villains will generally overestimate our range after we 4b pre. Yes, they may recognise that I am a rando and I could have a rando spewy 4b range. But overall, I expect villains to play more passively: maybe folding more, maybe bluffing less, maybe calling more where raises are warranted. I am trying to force villain into more mistakes, per the Fundamental Theorem of Poker. Importantly, the cost of creating this added deception is at the cost of very little EV, if any. I see it as neutral EV tool to manipulate villains' perceived ranges.


Upon further consideration, I think I should want to 4b AA. But only at the beginning of a session, just once, hoping to show it down. This will reinforce to villain that my 4b range is straight-forward and strong. Afterwards, I can 4b 89s, or smooth call AA, hoping villain has misinterpreted my range. Likewise, if the first hand I 4b is 89s, and I have to show it down, I might shift and raise an entirely strong range.

Manipulating ranges only works as an unknown against good players. Once villains catch on that my 4b's can include strong and weaker hands, I lose the deception element, and would prefer to employ the simplified stratagem of flat everything.

its pretty obvious that the best way to have a balanced range is to actually have a completely polarized range.

Please stop using "GTO" when you mean something else entirely. GTO strategies (as a shorthand for Nash equilibrium solutions/strategies), where they exist, definitionally CANNOT "break down" (or they are not Nash equillibria strategies). "GTO" strategies also cannot be exploited by your opponents when they know what you are doing, again by definition.

It's interesting to me that you want to include 89s in some sort of putatively undefined pf capping range as part of an exploitative tactic, but want to do so in part because "that's what the bots do 3 handed". It's interesting primarily for a few reasons: 1) Because the bots supposedly have a tendency to evolve towards 'GTO' solutions when playing other bots 2) It suggests that there IS such a thing as a GTO strategy that includes capping suited connectors oop. 3) It is unclear from the information provided whether the 3h bots cap these hands only in raised pots where there is a coldcall, or also do so heads up?

How can you reasonably think that what 3 handed bots do in super wide range spots has anything to do with this quite narrow range spot?

Exploitation starts on the margins of your range. If your opponent plays pretty close to equilibrium your adjustments should be subtle. The further he or she strays from that ideal, the more extreme your adjustments should be.

You've made an extreme (and frankly ludicrous) adjustment against someone who appears to play at least okay. His default strategy of playing okay is going to crush you because you have strayed so far off the gto reservation.

Just ask the co-4 bots, IMO.

This has already been sufficiently addressed: "Extrapolating to 6 handed UTG & CO isn't that much of a stretch"

Thanks for returning to this conversation. No snark.
You're saying 4b'ing 89s is an extreme deviation from good play. What about this alteration: I have a 4b range of {89s, AQs, AK, JJ+}, but there is only 1 combo of 89s for every 99 combos from {AQs, AK, JJ+}. Better?

Not sure what you mean by 1 combo for every 99 combos. There's 44 total combos of {AQs, AK, JJ+}. If you pick just 1 suit of 98s to 4bet with then you'll be 4betting with 98s 1/45 times, i.e. 1 combo for every 44.

So the better formulation would be:

Even if you literally meant 1 out of 100, i.e. 1 for every 99, in some theoretical sense, that would be too taxing on the human brain (given how dumb humans are) to implement well in practice. Better to pose it as "1 combo of 98s for every 44 combos of the other hands" since that would be super easy to implement--just pick your favorite suit of 89s and run with it.