What Manchuria establishes is part of the inconsistency, which had more to do with what was going on with the Japanese themselves than with their adversaries.
Saying it is "unknown" how the Japanese would fight is a copout. By that same standard, nothing is certain in life "except death and taxes", but the Americans had very good and reasonable expectation of how the Japanese would fight. It would have been criminal negligence of the highest order to assume that Japanese would fight more like the way they did in Manchuria rather than like they did in Okinawa. Your assertion that the Japanese fighting performance was totally random, like a flip of a coin, is just complete and utter nonsense.
It was a safe and justified assumption to asdume that the Japanese would do their best when fighting on their own home ground, and any commander who acted on any other assumption would have been negligent and incompetent.
The scenarios about the bombs + invasion tend to rely on assumptions about consistency of Japanese behaviour that don't necessarily hold up all that well. It depends on how the Japanese would have reacted, and Japanese reactions varied. Which Manchuria, among other things, shows. (The argument is less that "It would have been like Manchuria", and rather that Manchuria shows that it is perfectly possible the Japanese would have responded in some unexpected way.)
I know I've said it before, but the dropping of the bombs literally made the outcome of those other possibilities unknowable.
But that's also why these discussions never go away. The scenarios have been proposed since WWII are intended to close the debate, but are by nature post-fact justifications of the US choice. That choice once done can't be undone. But the justifications tend to treat things as fact that are not possible to actually establish as such. It never happened, they are fundamentally counter-factual. So the debate re-spawns.