Open Side Menu Go to the Top
Register
Insane Solver Output: Mix Fold on KKJAA with Ax to 138% pot river jam, no royal possible Insane Solver Output: Mix Fold on KKJAA with Ax to 138% pot river jam, no royal possible

02-06-2023 , 12:52 PM
Hey all, thought this was insane and just wanted to share. Obviously this has no bearing on playing vs human opponents. From GTO Wizard NL50 100bb:

HJ open 2, sb raise 11

Flop (23): KhKdJs SB b4.6, HJ call

Turn (32.2): Ah SB b10.65 HJ call

River (53.5): Ad SB all in for 73.75 (138% pot)

For HJ, A9s, A8s, A5s are nearly pure folds, although obviously it has only tiny amounts of these that float flop. AT is a call 58% of time, AJ calls 82% of time, AQ calls 89% of time. KQ and KJ mix call. Obviously calling AK. HJ also has KK for quads too.

All calls are actually less than 1bb EV with Kx being higher EV than Ax, so this could be a precision error because it implies that calls are greater than 0 EV. Especially considering HJ calls only 40.6% of time when MDF is actually 42%. Unless maybe this has to do with rake, meaning marginal +EV calls are actually negative EV?

SB jams only 1.6% of the time here, 9% of the time when they have KK, almost never with a boat, and a small% of time with hands they almost never have, like 8h8d 19.2% of the time.

SB much prefers a 36% pot bet when it bets top boat. It is just so hard to get called, it likes to bet a size where Kx can call, and some Q high, QT (straight), and JJ is forced to defend sometimes, and HJ still overfolds, calling only 66% of the time when MDF is 74%. So to make sense of what the solver is saying, SB gets such a good price on its bluffs and it is so hard to get value that all in is polarized to quad K or nothing. It doesn't make sense to bet top boat so big when you won't get called and a smaller bet with bluffs still prints. Therefore, at equilibrium, the perfect response is to sometimes fold Ax (unless this is a just a precision error by the solver).
Insane Solver Output: Mix Fold on KKJAA with Ax to 138% pot river jam, no royal possible Quote
02-06-2023 , 04:29 PM
seems a precision error. Once solving to high accuracy my sim has no shoves for oop OTR after betting 1/3 on the turn and in practice I think folding an ace would probably be bad here as IP.

https://gyazo.com/2eb9beb38e5e38266d5897845f6be40a

Looking at these obscure things in solvers - especially ones that aren't fully solved - isn't that useful imo.
Insane Solver Output: Mix Fold on KKJAA with Ax to 138% pot river jam, no royal possible Quote
02-07-2023 , 11:17 AM
I only looked at it because the hand came up in the trainer mode, and I saw the gto opponent folded AT, which is a type of fold I have never seen before. It makes sense if it is a precision error. It is just interesting even if useless.

I am curious, does your sim have the same bet size options as gto wizard on all streets, same ranges, same rake structure for 50 NL?

Now I am starting to wonder, how impactful are the precision errors to the GTO Wizard output in general? This one really makes you scratch your head. But I wonder if it is more limited to obscure nodes.
Insane Solver Output: Mix Fold on KKJAA with Ax to 138% pot river jam, no royal possible Quote
02-07-2023 , 06:25 PM
Its not insane if SB just jams KK for value then Ax is just a bluff catcher
Insane Solver Output: Mix Fold on KKJAA with Ax to 138% pot river jam, no royal possible Quote

      
m