I would need an expert on GTO or someone from one of these solver companies tell me whether a 100 times larger player strategy size and 4 times larger game tree to walk is feasible for a heads-up solver to converge fast enough.

All these examples are similar to examples from the normal game, except they are conditional on one of your cards being shown.

In the non-reveal game, the more likely I am to go all-in with a set, the more likely I am to go all in with other hands.

The more likely I am to have a set in my range the more likely I am to call/float a weaker hand.

The more likely I can represent a medium hand in my range that I would call the flop with the more likely I am to slowplay a set.

Revealing my card changes which hands I will group together, and my opponent revealing their card changes my perception of their range of cards. In the reveal game it's much more important exactly which hands I pair together, but at the same time it may be more obvious which are the correct ones, at least for a human (and I think if it's more obvious for a human then the solver is likely to converge to it faster since the EV benefit will show fairly quickly). In non-reveal it's not all that important whether I bluff with 9c6d on this board vs 9h6d. It's a much bigger difference in the reveal version, but isn't it also more obvious that I would favor the 9h6d because I can represent a flush more often with the 9h than I can with the 9c?

I think there's a lot of similarities between this game and stud. For example, if the reveal happens preflop, then you need a different response strategy for each possible revealed card, and that's also true of the up-card in stud. In stud there's four reveals for each player so the player strategy sizes should be unimaginably enormous, but in practice it doesn't impede top players from finding strong strategies does it? I'm not sure if anyone has ever tried to find a GTO 7-card stud solution; I can't find anything from a quick search. I remember they used to say 7-card stud was Phil Ivey's best game. Maybe because the skill ceiling is higher than other games?

Stud doesn't have the five card board so that reduces it a lot. So basically you're combining Stud and Hold 'em to get a much more complex game. For the goal of boggling the bot it doesn't matter if the player chooses the card to reveal or not, it's just the combinatorics of the combination of having up cards and a board that make things blow up.

I don't think that it would be much more difficult for a bot to solve this, as the exposed cards severely decrease the range sizes of the players.

Also, for real players, I think this game would be a bit boring and very simple, as the ranges ott/otr are completely face up and super polarized.

I'm a bit uneducated so take this with a grain of salt but with AI isn't computational power a non factor anymore?

AI will be able to estimate strategies way better than humans within a very small amount of time trained on this game (whether it's close to being solved or not), or any game for that matter. Maybe it would break the current NLH bots for a little bit but I don't think this solves the bot problem.

Sounds like a fun game though

Won't the second equilibrium computed after "motivated card removal" be precisely the same if both cards were exposed completely randomly?

In an equilibrium both players know each others ranges and strategies perfectly and neither can unilaterally deviate from their strategy to exploit the other. To me, exposing two hole cards simply removes a large % of combinations either player can have (and the new game tree), and a new equilibrium is computed using the unknown information of their remaining card/range. It doesn't seem to me that the motivation behind the card chosen to expose matters at all. Both "new" equilibriums (random and motivated) should be precisely the same, in my mind.

Never mind I'm totally wrong and you can remove my above post.