City specialisation and combat: the problems of 1UPT

[void_genesis]

Continuing these thoughts- I would love to see more consideration for a conditional stacking of units in the game in a way that helps the AI to wage war and makes the game deeper for the human. I had been having the same thoughts myself but recently found another post discussing the idea.

http://forums.civfanatics.com/showthread.php?t=513495

The key idea is that any units can stack, but that their individual strengths decrease when they do so. This retains the incentive to spread out your troops for maximum effectiveness but solves the log-jam issues with them getting in each others way.

Stacking two units would give each 60% of unstacked strength (to give a slight total increase per hex to 120%) but a decrease relative to the total damage output of 200% if they were not stacked (hence an advantage if they were spread out).

This calculation could extend until beyond a certain point no extra total damage per stacked hex would be added by stacking more units. The coding would simply apply % penalties based on the number of stacked units.

The idea could be extended to make the damage dealt to a stacked hex be distributed across the units in a way that increases the total damage dealt (or something more complex- maybe ranged attacks are more effective on stacked units and melee are less effective for extra strategy).

The idea would also potentially help balance aerial warfare in the late game and add complexity to the stacking of aircraft in cities.

The effect is realistic in that disease as a result of crowding of troops has been a major historic effect (often more troops died of disease than wounds).

The mechanic would also open up some potential for new effects (UA/religion/wonder) for warmongers that change the stacking calculations.

 

[Dunkah]

This sounds pretty exciting to me. The questions are...

1. Will the AI understand how to spread out their units for maximum effect?
2. Will the AI even use the stacking or are they hard coded not to consider moves where other units are?
3. What is the maximum stacking? I can imagine stacks of 10 archers on a hill. Stacks of doom.
4. What happens when a stacked unit gets attacked? Is it like 2 separate combats with lesser effects? So my unit may not survive an attack against a 3 stack? Even if each attack is diminished.
5. Does each unit take damage or is there a stacking/attacking order? Fight the top unit first then so on down the stack, like Civ IV.
6. Guessing this would have to be done with a DLL mod.

Whoward just released a DLL mod as part of his Pick n Mix, that allows limited stacking, (2 in Forts and Citadels and 3 in Cities). He already has a 2 per tile, 3 per tile and unlimited stacking mods. (possibly others).

 

[stackpointer]

Any shift from 1UPT will require DLL modding which is not anticipated for this mod at this point. However, see my post below for a stacking mod:

Gedemon (author of many great mods) has a Combat and Stacking Overhaul mod allows unit stacking. The mod should be compatible with CEP.

I'm sure he would appreciate feedback for his mod.

 

[void_genesis]

I think it is worth starting this conversation about stacking in the CEP community given that even the head designer for CiV has admitted that 1UPT was a major flaw in the design of the game. I know the philosophy of CEP is to make minimalistic improvements to the core game, but I suspect that a carefully thought out change would improve the game enormously for the human and the AI.

Answering the questions:

1. Will the AI understand how to spread out their units for maximum effect?
The current problem is the AI gets log-jammed when getting their troops into position to deal out damage, allowing the human to strategise and pick them off one by one. Perhaps the code used to escort settlers could be adapted to use melee/horse troops to escort siege/archer units. Presumably it can be coded to move toward an open position before attacking to maximise damage output.

2. Will the AI even use the stacking or are they hard coded not to consider moves where other units are?
My guess is the AI path tracking will consider all available hexes, so it should allow troops to smoothly flow into position around a target city then attack.

3. What is the maximum stacking? I can imagine stacks of 10 archers on a hill. Stacks of doom.
A stack of 5 or more archers would only have the damage output <3 unstacked archers- beyond a certain point adding more units should not increase total damage output any further (it could even start making the units more vulnerable to ranged damage).

4. What happens when a stacked unit gets attacked? Is it like 2 separate combats with lesser effects? So my unit may not survive an attack against a 3 stack? Even if each attack is diminished. 5. Does each unit take damage or is there a stacking/attacking order? Fight the top unit first then so on down the stack, like Civ IV.

I think it makes sense for a ranged attack to be spread across all the stacked troops, making it more effective against stacked units, while a melee attack should target the strongest unit that is fortified.

6. Guessing this would have to be done with a DLL mod
I'm not a programmer, so I would like to see the downsides of this modding strategy outlined.

 

[chumchu]

Limited stacking is an interesting concept but feels like an unnecessary addition. 1 Upt is elegant design because it is simple and intuitive but leads to complex maneuver. It meshes well with the tile based economy in civ.

However, the integration has been lacking in some respects. As I've said maneuvre and formation could be better. Units have low movement values and there is a lot of terrain restricting their movements and few roads enabling it. Cities are everywhere with strong attacks and ZOC from units inside. Because of the importance of elimination, concentration of force is king and wounded units need to retreat somewhere. This is an advantage for ranged units and mobile units which is compunded by the cramped maneuver. The flank attack bonus is low and discipline is specific for honor. My point is a change in one or some of these things could lead to a major improvement and therefore I do not see the need of getting rid of 1upt. Just lowering city attack strength or paying gold for healing would go a long way.

 

[void_genesis]

I have just started a new game with Assyria using the 3upt mod from the steam workshop. I'm not using the current CEP mod but it seems to be fine with infoaddict and a few minor other mods. I am going for domination so war everywhere and it is really interesting to see the changes to combat. Combining units gives an extra layer of strategy, and limited stacking in cities makes them more challenging to take. The AI is using the stacking very well- more often than not sending stacked units out that present more of a challenge to whittle down. The AI seems much better at defending cities and retreating injured troops as well.

 

[chumchu]

@void_genesis: Cool. I will have to try it some time soon.

 

[void_genesis]

Just replayed on a great plains map, bumped up to level 6, since my first map was all awkward islands and I couldn't wait to get a navy.

Oh god....the hordes of free flowing troops of the AI! They flow right into position around my cities and hammer them mercilessly. No more picking their advancing troops off one by one with an archer or two before they even get in range to do any damage. After taking a city state two equally strong neighbours DOWed me and took a city each, so I was toast. In 1UPT mode I would have been able to hold them off.

I would recommend other people give the 3UPT mod a go and see how it changes warfare. I found it made things much easier for me as well-less tedious shuffling of my troops into position as they trip over each other. The only current downside is needing to open the military overview menu regularly to select one of the stacked troops if they are idle.

 

[mitsho]

That sounds theoretically interesting, but... doesn't 3upt just mean you should fill up every tile with 1 melee meatshield + 2 ranged/siege units. Seems to mean much less strategy for me, not more. Unless of course it's further restricted to unit type (i.e. 3 archers on one tile, but not mixed troops), but then that means strategic units are worth much less (since you'd need 3 of them as well to counter 3 normal units). Anyways, doesn't it make most sense with 3 units to fill up all the tiles anyways...

Lastly, does the AI understand the princples above (many more ranged units, massed).

I guess production costs are not adjusted for units so war would seem to become mostly a result of production power (and neglecting buildings et al.) and economic power (upkeep)? That last point may be a good restrictor for armies I guess. But not for the AI which would simplify the higher level strategy to whether you can "tower defend" against the AI's zerg strategy?

That last one is a question btw., haven't played it and I guess I will try it out, but these are big question marks for me!

 

[ExpiredReign]

I haven't played a game with 3UPT either and to be honest, I don't think I will.
The concept seems to change the core system to such an extent that Civ sounds more like a war game than a civ game.
Do we really want hordes of armies decimating the nations around? I don't.
What about the unit promotions like "flank", do they take into account stacked units?
Core mechanics of economy, are they too tweaked to account for the greater army numbers?

Too many times I have seen modders tweak an idea so much that other elements seem to fall apart, like pulling on a loose thread on your shirt and your sleeve falls off.
I could be wrong, it might be a brilliant idea. I can't see it at the moment.

 

[void_genesis]

Just replayed Assyria on level5, great plains and had a similar odd experience of Japan DOWing me around turn 50 and arriving with around 8-14 troops (hard to count through the stacking). This seems excessive to me- maybe something has gone a bit weird, though I am only running very minor mods other than 3UPT.

I think this debate centres on the core issue: the AI cannot handle the strategy under 1UTP anywhere near as well as the human player. Adding some more well thought out complexity to the effect of stacking different units in terms of damage output and damage dealt then it might be possible to make combat manageable for the AI while still being interesting for the human.

 

[mitsho]

Actually no, I think the "are ranged units now overpowered and melee ones useless" much more relevant than the AI who you can easily 'teach' the zerg-strategy to provide a challenge for the human player (though not a strategic/tactical) one.

The ability of the AI to better move their units around basically tells me that the AI will do fine (but probably not get the ideal 'tile composition' like the human does).

EDIT: So I now played one game (not to the end, but you get it ) and I must agree, the AI is much better at fighting. This is mainly due to not having to painfully move its units around until it can attack. Chokepoints matter much less (but not nothing) and even more impressingly, naval invasions work much better (since it's hard to take out all 3 units on a tile...).

... BUT ... the "invincible" ranged units is a real no go for me. It makes you spam units more which increases micromanagent and decreases strategy. That's unfortunate. It does give me hope for the next civ game (if it ever gets made ), but for now, I rather stay with the 1upt rule.