1UPT, although a nice idea, has proven to be quite hard to be implemented, both for the AI, and even for the human play (sometimes, even a human player has a hard time to find his way across a carpet of doom!).
Whatsmore, in my opinion, Civ is NOT a war game, and it's not worth to become an extremely detailed war simulation with complex warfare tactics and strategies. War should be implemented, in a relatively easy (and realistic) way.
However, apart of my opinion, I'd proceed with analyzing the past solutions (1UPT and SOD) to find good and bad aspects, and find what can be improved.
I beg other forum posters to avoid too generic wishes like "I'd like a better AI". Of course we would, but it is like a Chimera. A well done AI requires a tremendous effort, perhaps out of the range of a computer game. So, if we are thinking of how our game should be done, we'b better NOT rely too much on this point.
Stack of doom characteristics:
- Very easy to implement and to play with (no unit jams)
- AI very effective to play with
- Tends to be abused (a well formed stack of doom is quite impossible to counter)
- Tends to vanify the rock-scissors-paper mechanics of the game, with specialized attackers NEVER fighting with their best opponents (eg., an attacking axeman almost never fights against spearmen, but archers instead).
- Dull combat: two stacks just keep attacking each other. Not too much to decide, apart of the sequence of the attackers.
1UPT characteristics:
- Strategic combat: each unit attacks just the unit he intends to fight
- Importance of deployment: units must be deployed in a strategic way (eg., melee on first line, archers on second line, cavalries to chase a particular dangerous enemy unit).
- Tends to form carpets of doom, hard to move around with.
- Extremely difficult to be handled by AI (and easy to abuse by a human)
- Forces some other game mechanics to limit some weakness of this solution (eg., units movement of 2 to be able to move across carpet of dooms, archer range of 2 to be able to shoot from the backline, high unit maintenance to limit the number of units in the game and limit the carpet of doom, etc...)
I think that 1UPT has a single feature that is QUITE hard to be solved: AI!
If Firaxis decides that CIV will become exclusively a multiplayer game, this could be the final solution. However, i DO hope that single-player will remain the core of the game, so I don't think that 1UPT (as it is in Civ5) could be a valid solution.
Maybe, the easiest solution could start from the Stack of Doom.
Most of the bad aspects of SOD could be resolved by encouraging player to SPREAD his units, with some mechanics working in a way that, although SOD is possible, a spreaded army is MORE effective. How? A possible way is a recall of Civ1, where a losing stack is COMPLETELY destroyed. Maybe it's too extreme and powerful, but the basic concept is valid: SPREAD your army, and your risks will be lower. For example, a losing stack could be damaged, so if you want to lower the damage suffered by your units, you'd better deploy them on different tiles. Another way could be a game mechanic so that bigger and more crowded stacks are less effective in combat, or heal at a lower rate, or suffer higher collateral damage, or have a higher maintenance cost (for logistics and supply). There are plenty of ways to encourage spreading an army.
The other problem of SOD is the vanification of the rock-scissors-paper concept. It is valid when a single unit fights another single unit, but with stacks it almost never happens: the defender will always HIDE and PROTECT the target that the attacker wish to fight against, eg., if the attacker is an axeman (and he of course wants to fight spearmen), the defending stack will always protect spearmen and send archers to counter the attack.
It's difficult to find a way (or even an abstract idea) to solve this problem of SOD. Maybe, there could be some sort of combat modificator that computes the strength and the composition of attacking and defending stacks: so, back to the previous example, an attacking axeman (in a stack with lot of cavalry) will be more effective against a stack with spearmen and archers: he will still fight against the archer, but he will have some combat bonus for the presence of cavalry in his stack (since cavalry counter archers quite well). How you can see, it risks to be a complex mechanic, also hard to be well understood by the player; not too sure whether it could work.
Annyway, in my opinion, if we want to determine WHICH solution is closest to our ideal, SOD is closer than 1UPT, since SOD limits are somehow manageable, while 1UPT has a MAJOR limitation (AI) with UNRELIABLE solutions (improve the AI!!).