Anothing this that I think need mentioning, is that you're looking at 2 different types of Demographics. Both games appeal to very different types of people, who look and want different things. You also need to understand that HOI is a lot more serious, covers a much smaller time line, and henceforth needs can't stray too far from history. It also draws in a more, serious, and maybe even older people who are very knowledgeable about the time period, and even more hardcore. You have a lot of people who know way too much about time period

. As I've said before HOI attracts a different kind of people, and that can contribute to atmosphere.
I'm not convinced by your characterisation of the demographics, here. They may very well be different, but both are serious and both have clear ties to history. Taking myself as an example, I like both, and I know far too much about the military history of WW2
and about medieval history - especially medieval economics and technology. I think both fanbases have their "grognards".
What I see in the HoI forums, though, is closer to a malady I started to understand in the tabletop roleplaying community a while back. It's still there, but at least now it's understood and there are workarounds. The problem is that:
- on the one hand, several players want a free hand to do what they think the historical players "should have done" or what they would like to explore them having done, but...
- they also, on the other hand, have vivid pictures in their mind of all those WW2 movies and stories, and they want the game to deliver that experience to them.
These are incompatible goals. It's the RPG "impossiblke thing before breakfast" - the desire for a sandbox that delivers the classic (WW2, in this case) story.
I am somewhat guilty, too, just to be clear. And there is something not too far away from this impossible dream that might be possible - a sandbox with plausible outcomes for every combination. But the problem is that this combination is incredibly hard to build as a game. The requirements of the system are subtle, deep and manifold.
For example:
Basically I outlined the result as of now after 10 months. Problem addressed to me means "solved" - the "problems" are not solved, that easy. They might have worked 3.000 hours on it, problem is still not solved.
This is a classic example. We want the AI to react to "non-standard" opposing strategies, but we want it to do so "historically" and we want it not to dither. There have not actually been that many human generals that have achieved this, let alone an AI! You say we have not seen a solution in 10 months; I would say we have not seen a (complete) solution in over 16 years. Sure, we have seen games that stop the dithering. But that is only of very limited use if "solving" the dithering only compromises every other objective, and that's what I see in every other game I have tried.
This seems to me to be a problem like that of getting an international free trade agreement. The PM of Australia has been pointed to as someone who "got a free trade deal in 6 months" where everyone else was taking 4-5-6 years to negotiate one. What a clever chap! Well, no. What he got was a
no tariff deal. Completely different thing. Sure, absence of tariffs is a necessary
part of a free trade deal - but it's only about 10% of one. In the same way, absence of dithering or "shuffling" is a necessary
part of a great Operational level AI - but it is only a relatively small part of one. And, if you set out just to "solve" the dithering, independently of everything else, you will likely (a) break everything else and (b) make it so that, when you have then fixed everything else, you will need to go back and address the dithering again.
TL;DR - "solving" the dithering might be easy, if you do it while ignoring all the other things an operational AI should do. The difficult bit is getting all the goals of the operational AI addressed all at once. Paradox are taking "so long" to do this because they are trying to solve the whole set of equations all at once. I think this is the right approach to take, even though it takes longer. Game creators have been going round in circles solving one at a time for well over 10 years. It's time to take the time to solve them simultaneously.