The AI does seem to be very poor, especially when compared to VR design's previous games. The biggest problem is that they went a bit too ambitious with the 4x options, in particular building roads and creating new zones. I doubt the AI is "not fixable" - while I haven't seen the code, the game does appear to be at least based on similar wargames VR have made, such as Decisive Campaigns series, and in those games the AI wasn't terrible - but the games scopes were smaller, so the AI only really had to deal with logistics and warfare, not also zone boundaries and road construction (and logistics were nowhere at the level of SE).
I don't own the previous games, but I suspect the AI in Shadow Empires is more developed than the AI in the previous games.
In previous games, it was the responsibility of the scenario designer to balance the opening conditions and objectives of each scenario. Each scenario was then repeatedly run, tested and tweaked to provide a convincing experience.
In Shadow Empires, the opening conditions (OOBs, competing forces, goals, etc.) are ambiguous. That's why the regime objectives (i.e. victory conditions) are so simple (i.e. land grab) and fixed.
Using the information we know about the DC Community Project and extrapolating from what we see in Shadow Empires, it's somewhat obvious the game's structure does not support a robust AI.
There appears to be some very limited agent system use (mostly at the unit level), but nothing substantial or multi-level. We know this because important concerns were not separated by the designer.
The changing zones and zone flipping by moving units is a core requirement for the existing AI to understand the game's state and make decisions.
< Message edited by Malevolence -- 9/16/2020 1:36:09 AM >
Nicht kleckern, sondern klotzen!
*Please remember all posts are made by a malevolent, autocratic despot whose rule is marked by unjust severity and arbitrary behavior. Your experiences may vary.