A campaign I'd like to run.

For various reasons, roleplaying in games like D&D has long fascinated me. My own roleplaying is interesting, but the roleplaying of people around me has a tendency to puzzle me. Even more odd is that while individual players may say their characters subscribe to particular values, their actual playing of those characters often don't reflect those values. Thus, whatever a player puts under "Alignment" on their character sheet tends not to have any bearing on how their character behaves under pressure. (Or even normally, if the player is particularly new to the game; Take the "Lawful Good" paladin that takes a liking to brigandry, for example.) In a group where the roleplaying is particularly weak, the group as a whole tends to have a stronger alignment than any individual character.

I had an idea for a campaign I wanted to run that would let me observe the aggregate alignment-related effects of a group and its characters. It would somewhat like a sandbox world, but there would be specific campaign end conditions. (No, those conditions don't include TPKs.)

Take a campaign setting where there are effectively four keystone NPCs. Each of these effective NPCs (they might be an individual NPC, or they might be an aggregate of similar related NPCs) would represent a classic D&D alignment. You could go with cardinal alignments, (LN, NG, CN, NE), or you could go with corner alignments (LE, LG, CG, CE). Each of these NPCs, intentionally or not, vie in some way for philosophical domination of the campaign setting. The ending condition is such that one of these NPC's philosophies is so incredibly dominant that there is no forseeably near event that could swing the balance away.

The party's role? Well, each of these four NPCs may hire the party for their own purposes. Each of these jobs would in some way change the balance of power withing the campaign environment. The party is free to pick up and drop employers whenever and for whatever reason they wish, but by favoring one NPC over another, or even a pair of non-directly-opposed NPCs, they're helping those NPCs become dominant. Game over when things can't reasonably tilt back the other way.

I'd like to try this particular experiment with various groups, and see what insights it might give me to a particular group's motivations and roleplaying style.