ty very much

Yeah, I'd like to see the numbers re-run for a more realistic range. Even 22+,A2s+,K8s+,Q8s+,J8s+,T8s+,97s+,86s+,75s+,64s+,5 4s,A2o+,K9o+,Q9o+,J9o+,T9o might be too loose with a 20bb in the BB. That's 36% of hands.

In fact, isn't this entire scenario totally "solvable"? Seems like whatever you did could be iterated over all reasonably possible open ranges and all reasonably possible open sizes.

SB affects things making it alot harder. But yes, assuming SB always folds and BB never flats makes it easily solvable.

Ok guys here is one scenario for 20bb.

For now I do it semi-manually which is why I can' arrive to exact optimal strategy but I am pretty close and small mistakes of one side doesn't matter anyway.
So here for opening range of : 22+, A2+, KT+, K9s-K2s, QT+, Q9s-Q8s, JT, J9s-J8s, T8s+, 97s+, 86s+, 75s+, 64s+, 53s+, 43s which is 35.1%

BB pushes :


BTN calls :




KJoff calls 1/3 of the time.

BB strategy is best strategy vs BTN.
BTN makes mistakes by calling KJoff 1/3 of the time and calling A4s always. Unfortunately it gets tricky when you call KJoff less often and I couldn't arrive at better approximation. Anyway sum of mistakes is so small that it's practically equilibrium.

It reall should be done automatically and then it would be 100% accurate but for now it's not possible.

EDIT: BTN strategy contains sizeable (by equilibrium standards, it's 0.06BB and only when you actually get this KJ) mistake (calling with KJ) but it's only 1/3 of the time. Maybe in real equilibrium KJ calls only about 1/5 - 1/6 of the time Idunno. By playing this chart you guarantee yourself almost exactly as good EV as equilibrium would guarantee and this strategy is better against looser pushers.

"Whole thing" could be for sure approximately solved by those ICM programs if they are twisted a bit to allow raise/reraise but they work with approximation of ranges instead of exact ranges (they assume order of hands and if given hand is calling all "better" hands always call too; no such order exist in real poker). I am pretty sure 3hande raise/3bet game is easily solvable with today computing power but someone has to actually code it, spend some time calculating and then give his/her results for free. Not gonna happen anytime soon :-)

Very awesome. Thanks a lot. I find that this is also incredibly useful for MTT endgame.

In Mtt's ppl usually steal less and there is survivor factor which should make you want to avoid confrontations (pushing less/calling less).

What's buttons raise size?

why is BB defending 12ish%

surely vs a 45 ish % button 2.5x open he needs to be defending more like 17% of hands