Quoting The_Biz, reply 29
I think it needs a lot of refinement and balancing
strategy game quality = proportional to how frequent, difficult, and interesting the decisions are
FE/LH have a setup that allows for this, but it feels like the game can be very formulaic because there's no challenge that requires the player to adapt or think
maybe the player can set up his own challenges (eg. win without using X, win without abusing Y, with with giving the AI bonuses of Z%), but that's too artificial.
the players need to actually be players, not designers
I'm sure the game can be modded to be extremely good even without AI (just monsters), but I don't know whether that'll happen
In what way do you think this could be improved upon? More random behavior from the AI?
What games do you think handle this appropriately?
I think the fundamental problem is the game's reliance on symmetrical progression, like Civilization
this works in a multiplayer setting
But when bad AI's try to simulate that multiplayer setting, it falls apart
Even if the AI is better than the rest of the genre, it will still fail
The empire-building / diplomacy / warmongering will never succeed if it's just a contest vs. AIs
even if the AI's get more variation, they will still fail at providing a challenge
Realistically, the only chance they have at making the game great is to focus on the aspects that don't rely on strategic AI.
The environment with its monsters and asymmetrical progression and randomness has a lot of strategy to it. And the problem of tactical AI for a single battle is much easier to solve.
Rogue-like games and turn-based tactical games (RPGs / any games with a party + battles) have been successful at providing challenge in this area.
