HoI 3 had so many problems at release. In the end, after the expansions, the game morphed so many times that you could hardly guess it was the same game.
Yes, I'm exaggerating ... slightly. But the point is that with a game of the scope and scale of HoI, the testing has to be EXTREMELY thorough.
The key will be to seeing if Paradox can get the A.I. to play the game effectively.
Alright - you asked for it - here it goes:
The A.I. has to be able to micro-manage the coordination of every base unit (division) along a front. I found that I had to strictly control how many divisions APPROACHED an attack province in order to ensure that ONLY enough divisions were involved in an attack along a single province border.
This means that all divisions that will take part in all attacks from multiple provinces have to be carefully managed in their movements and properly supported from neighboring provinces.
When attacking a province with good defenses, across a series of river fronts, and/or with many supporting neighboring provinces, the most important consideration is to attack with ONLY enough divisions to cover the entire front, and ensure that there are enough reserve divisions available. It is important that the RIGHT types of divisions are included in the attack, and that there are sufficient numbers of those divisions to maintain the attack.
As a long battle evolves, divisions that have taken significant organization loss must be REPLACED with fresh divisions from reserve.
Breakthroughs have to be treated more like a wedge than a long snake. The follow-through with mobilized infantry and foot-infantry support must keep that corridor open at all cost. Armor dies very quickly when cut-off. I've seen the A.I. throw away good units by not properly following them through a hole in the lines. This must be PLANNED.
Air support and air cover also have to be micro-managed. Mega-stacks of interceptors quickly determine who has air superiority in a specific theater, and once lost, can't be regained in a timely manner. Its better to keep the planes on the ground and use them only when it really matters, and then only for the first burst of a battle.
Naval factors must not patrol. Intermittent air patrols are more effective at guarding coasts than ships. The A.I. must learn to be MUCH more frugal with the use of ships.
There are a few tools missing from Paradox games that would be really convenient. One is a simple player calender where the player to plot alarms for specific times and dates, with custom pop-ups. Say a player is really concerned about getting a specific tech advance in a tree by a certain date. My memory isn't good enough to keep a whole lot of different tasks in the front of my mind at the same time, so it would be very handy to have a calendar that I could pull up and punch in an alarm to go off with a short message letting me know when the right time has arrived.
This would also save having to turn on pop-ups for an entire category when I only need it part of the time.
This leads me to a request for a new feature in the Message Settings. I would like a button that allows me to turn ALL OPTIONS to Pop-Up And Pause with a single click. With most Paradox games, I turn on all pop-up and pause, then change them through my first play-through. Better to know what all the options are first and select what to turn OFF than the other way around.
Another really basic feature that would be really handy is a pop-up calculator, or small and simple spreadsheet. Something I could bring up while playing and move around the screen. Its amazing how important the basic math is in Paradox games. There's a lot of number-crunching.
So, by my calculations, I figure the game should be ready by Q2 ... 2027.