So there is a difference between optimal play, with regards to a side's chance at winning the game, and optimal play for a game.
As an example, with regards to winning the game when perm. alliances exist, the optimal game-play is for the 1st place player to team up with the 2nd place player, and crush the rest of the world.
Similarly, if you are badly enough off, the optimal play for someone losing a war is to not capitulate -- only capitulate if you get enough resources that you stand a chance of breaking free later. (Unless you consider winning as a minor vassal of the game-winning empire to be better than being wiped out).
The 'roleplaying' components to the game are part of what makes the game better in the sense of better AI.
Along these lines, some thoughts...
Perm. Alliances should be used to make the mid/end game more interesting. The AI should prefer to form a perm. alliance between two parties that are relatively close, and they should seek to form a combined power/score/etc that is somewhat, but not overwealmingly, higher than the top player in the game.
The idea is to allow civilizations that have 'fallen behind' to once again become rivals for victory. At the same time, two civilizations that are 'top of the game' should resist combining -- the resulting game isn't as interesting, as all it did was entrench dominance.
I've already put forward a similar though -- the AI should prefer more recent religions, all things being equal, to older religions. As it stands, founding Islam or Christianity and changing to those religions is almost always a diplomatic disaster: the 'old' religions have had too much time to become entrenched by this point.
However, if the AI prefers to switch to new religions, then staying in the old religions becomes a diplomatic disaster as more and more civilizations start disliking you. This means that starting with and spreading the newer religions is rewarded more.
A state with a shrine to an older religion should still be somewhat stubborn about switching gods, and you should have a reluctance to switch gods to the same as civilizations you dislike, and a tendency to switch gods to religions of civilizations you like...
Another vassalization thought -- should civilizations be more open to becoming a vassal? If you are above 50% of the land area of your master civilization, the arrangement is completely voluntary on your part. You can quit whenever you want.
As an example, with regards to winning the game when perm. alliances exist, the optimal game-play is for the 1st place player to team up with the 2nd place player, and crush the rest of the world.
Similarly, if you are badly enough off, the optimal play for someone losing a war is to not capitulate -- only capitulate if you get enough resources that you stand a chance of breaking free later. (Unless you consider winning as a minor vassal of the game-winning empire to be better than being wiped out).
The 'roleplaying' components to the game are part of what makes the game better in the sense of better AI.
Along these lines, some thoughts...
Perm. Alliances should be used to make the mid/end game more interesting. The AI should prefer to form a perm. alliance between two parties that are relatively close, and they should seek to form a combined power/score/etc that is somewhat, but not overwealmingly, higher than the top player in the game.
The idea is to allow civilizations that have 'fallen behind' to once again become rivals for victory. At the same time, two civilizations that are 'top of the game' should resist combining -- the resulting game isn't as interesting, as all it did was entrench dominance.
I've already put forward a similar though -- the AI should prefer more recent religions, all things being equal, to older religions. As it stands, founding Islam or Christianity and changing to those religions is almost always a diplomatic disaster: the 'old' religions have had too much time to become entrenched by this point.
However, if the AI prefers to switch to new religions, then staying in the old religions becomes a diplomatic disaster as more and more civilizations start disliking you. This means that starting with and spreading the newer religions is rewarded more.
A state with a shrine to an older religion should still be somewhat stubborn about switching gods, and you should have a reluctance to switch gods to the same as civilizations you dislike, and a tendency to switch gods to religions of civilizations you like...
Another vassalization thought -- should civilizations be more open to becoming a vassal? If you are above 50% of the land area of your master civilization, the arrangement is completely voluntary on your part. You can quit whenever you want.