[NFP] [discussion] Major flaws of Civ VI - part 1: City combat strength and defense

[kaspergm]

First off, a disclaimer: The point of this thread is not to bash on Civ6 in general, I think it’s a great game with many good elements. However, I also think it has some major flaws, and that if we don’t discuss these, we and the developers can’t learn from them and make Civ7 even better.

The first flaw I’d like to address is the way cities can defend themselves in Civ6 (and also could in Civ5, I could add) without any actual military presence. With walls present, this defense also doubles as an offense, which only makes the problem worse.

So why do I think this is a problem? Well, it’s well known that you can go through an entire game of Civ6 and manage even a large empire with only a handful of military units. Even in a conquest game, you can conquer entire civilizations with only a small offensive force, while you leave your entire empire empty of military units. This is both unrealistic and very bad for game balance, and it favors builder civs (who focus on technology and/or culture) over militaristic civs.

Here I’ll try to list some specific points that I find problematic, how they interfere with other aspects of the game, and how they possibly could be improved.

• The city ranged attack (in Civ6 it requires walls, which is at least a small step in the right direction) is one of my pet peeves. This feature, probably more than any other, renders defensive armies all but obsolete. The fact that cities can generally take out contemporary units (particularly siege weapons) in only a couple of shots notoriously makes the AI all but unable to capture cities. My preferred solution would be to rethink how garrison units work: A city should only be able to range attack if a ranged unit is stationed in the garrison. However this creates a possible conflict with the desire to also station melee units in a city. Therefore I think the 1UPT rule needs to be reworked for cities, to allow you to station a number of units for city defense, possible scaling with presence of military districts and wall levels. This also ties into my next point.
• The city passive defense strength in Civ6 has long been a big thorn in my eye. The fact that city strength scales with technology and, even worse, the way that city strength magically jumps all over your empire once you build a unit with higher melee strength, is just really bad. My favorite obnoxious example: Build a melee ship to boost defensive combat strength of all your cities, even those inland. Really!? Rather, I think a cities combat strength should depend one some fixed parameters: Wall level, terrain, and melee combat strength of unit(s) stationed in the city. Walls should control damage taken from ranged and siege units as well as buffing against melee/mounted units, while combat strength of garrison unit should control interaction between melee attacks into the city and the garrison health.
• Another aspect I’d like to see addressed is the invincible garrison unit. Why is it that an archer placed in the city cannot take any harm from archers stationed outside shooting into the city, not to mentioned bombardments from siege units? Whoward69 made a mod for Civ5 called “collateral damage” which caused garrison units to take some fraction of damage offered to cities. I’d like this feature to be developed and generalized: Why can’t ranged units outside the city target a ranged unit inside the city? Obviously walls should offer a big defensive bonus, but none-the-less.
• Then there’s the question of the non-directional city walls and siege support units. By this I mean: A catapult stationed on one side can remove walls on ALL parts of the city, and ONE battering ram stationed next to the city will cause melee units on ALL sides of the city to cause full damage to walls. Particularly the latter one is a pretty serious balance issue - battering rams are plainly overpowered, even if they now have a narrow time window before becoming useless. Obvious solution is to make walls work in segments (one on each hex side) as well making battering rams/siege towers only work for units actually sharing their hex.
• In a completely different perspective, there’s the question of unit costs in Civ6. Because you need relatively few units, each unit has a very high production cost. This obviously ties into a bigger question of overall production costs in Civ6 (a problem that was never really fixed after the very early IZ nerf). In a version of the game where you’d ideally need at least one unit for defense of each city, and several units for defense of border cities, I think unit costs need to be toned down. This would also help defending civs, who might actually be able to muster an army. In order to balance this and prevent quick steam-rolls, I think military units need to tie into the city population system. This could either be done through a “one population loss per unit” system, or (and this would be my preference) by making a military unit require a citizen allocation, similarly how you can assign a citizen to work in a farm, in a mine, or in a campus. This would mean you’d actually have to provide food to maintain your population if you want to maintain a large army (plus it would give a neat twist of each military unit have a “home town”).
• And finally talking of steam rolls, I think there needs to be some mechanism to halt an offensive avalanche once you break through border defense. In Civ6, once you break one city, it’s generally very easy to quickly take over a handful more, sometimes even an entire empire, because the defender has lost his army. In order to prevent this, I think the population loss when conquering a city should reflects citizens and defenders fleeing to nearby cities, where some should spawn as military units for the defending civ. Not necessarily super equipped units - having three citizens flee one city and spawn as tanks in another city obviously feels a bit strange - but exactly how that could work would probably require an overall rethinking of the unit combat system, which could be a topic for another thread.

Now I know many of the points addressed above would require major redesigns of central aspects of the game, which is why I think this discussion is mostly interesting in terms of what we could wish for in Civ7. So which suggestions can you add to improve city combat in Civ games?

 

[HalveBlue]

The combat system in Civ VI is atrocious!

The combat stats and unit types aren't properly balanced, even before unique Civ bonuses are factored in.

A lot of mechanics are badly abstracted or completely absent; promotion trees are wonky; forts are useless...

I hope for the next go around, they hire somebody from the war gaming industry (Matrix or Slitherine) to rework the combat system of the game.

 

[Oberinspektor Derrick]

While I am apparently not as negative to city combat strength as you seem to be, I still think it is a major problem as well and I appreciate the discussion.

The main issue I have in civ 6 is that citiy defence has been made too easy, and offense disproportionally harder.
This is especially obvious for the AI, in that it totally fails at any kind of domination play and snowballing from there.
Civ 5 could see AI empires snowballing out of control and taking up bigger and bigger parts of the continent's landmass, but not so much for civ 6.
And that was fun, because if you played peacefully you were always at risk of getting rolled or out-teched if you didn't prepare accordingly during the entire game, and not just gaming certain mechanics to stay invincible.
And if you played domination yourself (continents map), nothing was as exciting as having subdued your own continent, and seeing your new massive "future nemesis" in all his glory in the "new world".

I think that while many of your suggestions seem good, I personally think the civ 6 wall mechanic is the main culprit behind this.
The difference between a city going from 0 walls to Ancient walls is dramatic, so much that it can often suffice as a "get out of jail free"-card (security wise) for the majority of the game onwards.
Not only does the city gain an extra layer of defence, it also gains a ranged attack, and for melee units the hidden combat strength modifier makes attacking a walled city ithout battering rams/siege towers a suicide mission for hardly any effect.
Add in the fact that siege units got pretty nerfed in civ 6 compared to civ 5 (no longer can you upgrade your trebuchets to +1 range and indirect fire), and you have some pretty damn solid defences that really have no weaknesses (particularly against the AI).
Until you get to Flight/Advanced Flight that is.
At that point you can finally shoot with your siege units from outside the city's range, and bombers flatten any walls immediately.
I'd really like to see these dramatic differences get toned down, both the difference from going to 0 walls to Ancient walls, and from pre-Flight to getting Balloons and Bombers.
Civ 5 was much more balanced in this department - no city felt truly unconquerable if given enough time, Ancient Walls were not a "lol, you can't hurt me anymore" building, and Bombers were not the "lol, I win" unit that they are today.

Fun fact: I run the "stupidwalls" mod, that reduces wall strength by 50% (+50 hp per level iof walls instead of the current +100), back to what it used to be before GS buffed them to their current state.
Even with -50% wall strength from this mod, those walls are still damn near broken good, stops the AI right in its track and still present a formidable challenge to overcome early on.
I have no idea why Firaxis would even buff walls to their current level, I think that even the pre-buff walls should instead have been nerfed by another 20-30%.

As for the scaling of city combat strength, I think that it's a good idea that defences scale, but really dumb that they scale from the highest melee combat strength, like you say.
Why can't they scale gradually with each tech researched, and otherwise get buffed with walls, hills and policies?
The current system feels really gamey, in that securing city defences only relies on the player manipulating the right mechanics.
Get yourself Ancient walls in the particular exposed city, and then produce/buy a Horseman on the other side of the world.
Once he pops on the other side of the world, that will magically and instantly take care of all your defences across your empire.
Dumb? Yes.
Good for the game? No.

 

[hhhhhh]

Historically, cities are hard to take, as long as they are walled. And building walls at certain places are sometimes major events which got recorded in history. And lots of times, rioters or rebels have to bypasses strongly walled cities (so they can only loot the countryside).

But I get your point. If you are playing against AI, then ancient walls suffices for fending off any kind of attack, even the ones with medieval or renaissance units. That's partly because AI is bad. But I wish there's something like ancient walls are only 50% effective when facing medieval units, and 25% when facing renaissance units (and be ignored by industrial units).

Agree on city passive defense strength. I wish it's more related to the population and so on (another way to make population a little more meaningful!). It's weird that when you have advanced units you can settle 1-pop city which has higher defense than someone's capital.

----------------

It looks like you are writing a series of posts. I hope you do rant about the chopping mechanism and how districts get more and more expansive. (Late game districts can be easily 400 production, which is super annoying and one can only hope to chop those.)

 

[Mount Suribachi]

I think you need to take into account what Firaxis is trying to achieve by making city walls so strong - it's not to stop the AI attacking the player, it's to stop the player attacking the AI.

Just imagine how easy it would be to ROFLstomp the AI if they didn't have walls everywhere? They also changed something to make the AI build encampments everywhere. Ancient Walls + Xbow + Encampment makes conquest incredibly difficult and tedious till you get to bombards & frigates, or even better artillery, bombers & battleships.

When everyone complained about the AI constantly conquering city states, Firaxis didn't change the AI behaviour to make them less likely to attack city states. Rather, they gave city states Ancient Walls immediately and now they survive much better.

Chairman Yang of the Human Hive was the biggest thorn in the players side in SMAC. Why? Because he got "city walls" in every one of his cities that he spammed everywhere

 

[Oberinspektor Derrick]

I think you need to take into account what Firaxis is trying to achieve by making city walls so strong - it's not to stop the AI attacking the player, it's to stop the player attacking the AI.

I think they could still have solved that problem in other ways than making the current iteration of walls, and generally feel like they took a step back in this version of the series.

I remember one of my last Domination games in civ 5, where I was playing Zulu against the Assyrians.
I really struggled with some of their cities, especially one bastion of a city that was settled on a hill, had all levels of walls, Tradition from the policy tree, and a strong garrisoned Infantry unit in the city (I was still using Riflemen) backed up by Artillery.
The way I eventually solved it was by approaching the city in a wave of infantry, and using great generals to claim land towards it by retiring them into a Citadel, using workers to build roads on the claimed land, and setting up 3-4 siege units (Cannons or Artillery, cant remember), outside the range of the city (all with the +1 ranged upgrade and Indirect fire upgrades).
It still took me very a long time to break that city (in the ballpark of 15 turns), but overall it was an enjoyable experience as I still felt like I could do small steady progress by playing well.
The rest of his empire was also hard to take, and eventually I got me some Great War Bombers to finish him off.
It was still a real fight doing so though, since his cities could return AA fire on said bombers and their defences still took a while to crack, a lot more than cities take in civ 6 (1-2 turns is usually enough to bring a city from full to 0).

Now contrast this to civ 6, where some well placed cities like these can be near unbreakable, and often the only thing you can do at that point is to ignore it and wait for Flight/Advanced Flight, at which point you can level the city in 1-2 turns with no real counterplay from the AI.
It's not fun when things swing from one extreme to the other like they do in this version of the series.

 

[Linklite]

I feel like the OPs solutions woukd end up just being swings and roundabouts. Sure, it wouod be easier to take a single city, but when you do, the local cities get (effectively) free military units? In my games, that would effectively render them invulnerable. I'd be at more risk of losing a city, but substantially less risk of losing the game. Also, requiring population to build military units would absolutely nerf offensive play in the game. Pop increases so slowly and desired for districts etc that requiring it for military units, until your cities reach that plateau where pop becomes largely irrelevant, would cripple warfare in my opinion.

I definitely sympathise with some (and to an extent, all) of the points raised, but they flow from a lot of the systems of the game. I'm hoping that I'm not coming across as just being downer or negative or just being funny, what I'm trying to say is that this isn't just a patch job, it requires a rethink of the central conceits of the game, and, I daresay, to rebuild the game from the ground up.

I could go into some of them, but those would be long posts and further derail the thread, which is my point. The problems brought up (and they are real and valid) can only be solved at the expense of other issues because they are symptoms, not causes. You want to make cities easier to capture, but I don't want to spend 80 turns, a quarter of the game, making this great city, only to lose it in 2 turns to Genghis. So you come up with a modification, to which I say....and so forth, because all these systems that are giving you grief and you want to get rid of or marginalise were built to solve other problems that resurface.

In other words, I propose that the problems are deep rooted and require to be pulled out root and branch and start again from scratch. I also think we need to be careful with these discussions - it's easy to think of solutions, but we rarely conceive of the problems that they bring. We then wonder why the Devs never did that in the first place.

On a related note, has anyone noticed that pre-wall cities in the early game have gotten harder to defeat? I used to be able to get a couple of warriors and capture a few cities before their CS was too high. The last few games, they've become high enough to effectively make them invulnerable to my warriors before I can even take one. I haven't tracked numbers, but it is a lot harder to get those initial cities.

 

[nauberry]

First off, a disclaimer:

• Another aspect I’d like to see addressed is the invincible garrison unit. Why is it that an archer placed in the city cannot take any harm from archers stationed outside shooting into the city, not to mentioned bombardments from siege units? Whoward69 made a mod for Civ5 called “collateral damage” which caused garrison units to take some fraction of damage offered to cities. I’d like this feature to be developed and generalized: Why can’t ranged units outside the city target a ranged unit inside the city? Obviously walls should offer a big defensive bonus, but none-the-less.

Back in civ 5 days I enjoyed the collateral damage mod a lot. If I remember correctly, it also applied a chance for a city with 0 hp to take population damage on further ranged attacks on it (think of battleships bombarding a coastal city, with no units being able to capture this city). I think some of it could be incorporated into a future civ game. As it stands now, it feels very annoying to capture a walled city with a ranged unit inside, as there is nothing you can do on the ranged unit apart from capturing the city.

I think a large problem, that you also partly addressed in your post, is that armies in civ 6 are tiny. In the late game, the unit resource costs are so demanding, that having a large army is hard for even the player, and even more for the AI. If you compare it to civ 5, where every AI would have carpets of units in the late game. In the other hand, having vast armies accounts for long turn timers. I do not really have any solid ideas on how to fix it, but maybe some sort of manpower resource would help.

In paradox strategy games, all players have some sort of manpower pool, from where they can draw units. I think this could be combined with civilization rather well. For example, the manpower could be directly proportional to your cities population. Units could be trained/hired fast and cheaply as long as you have manpower. After this, you would be forced to buy units with gold as in mercenaries. For further balance, manpower could only be drawn from unoccupied cities, so that you would not benefit from newly conquered cities during a war.

The manpower would at least address another large problem of civ 6 combat and war. If you are able to launch an offensive without losing units due to them hitting 0 health, wars are not really costly for the attacker. You can always just fortify your units and let them heal to full. Having unit healing tied to manpower, large offensives could seriously hurt the player on the offensive, even though they would launch a successful war. This would leave other players/AI as a much larger threat even after a successful war, as your manpower would probably be very low.

On a completely other topic, which might not be a real issue, more of a personal distaste, is that civ 6 maps don't feel like entire planets anymore. The scale of the cities is so large, that no non-urbanized areas are left in the late game. This could be solved with larger maps, but unfortunately they are too unstable to work and also are very slow to play on in the late game. My wish is that civ 7 could somehow keep the districts and growing cities somehow, but maybe allowing multiple districts per tile to leave room for forests, farmlands etc. surrounding the cities.

 

[Noble Zarkon]

I have to say I basically totally disagree with the premise of your post, adding City Defence has been a great innovation to the series - no more city captures by a lone warrior on t5!!

I'm with @Mount Suribachi, this system is there to make things more difficult for us. I play on Emperor with the TCS Mod that gives CS free walls which helps stop the AI just razing them (why do they keep razing Auckland on Archipelago maps ).

 

[Mount Suribachi]

Now contrast this to civ 6, where some well placed cities like these can be near unbreakable, and often the only thing you can do at that point is to ignore it and wait for Flight/Advanced Flight, at which point you can level the city in 1-2 turns with no real counterplay from the AI.
It's not fun when things swing from one extreme to the other like they do in this version of the series.

Oh I agree, it's why domination is my least favourite victory type, its just so tedious. OTOH, there is some level of historical realism for the medieval period where the defensive technology dominated (castles) until the offensive technology (gunpowder) caught up. It will be interesting to see if the trebuchet does anything to adjust the mid-game conquest play.

 

[Oberinspektor Derrick]

I have to say I basically totally disagree with the premise of your post, adding City Defence has been a great innovation to the series - no more city captures by a lone warrior on t5!!

I can't really recall that being a thing in civ 5, but either way, they didn't need to add the current walls just to prevent a "lone warrior" from conquering cities.
There are countless other ways they could have prevented that, like adding a minimum combat strength so that even the weakest of cities own can hold off 1-2 weak units, but needs involvement from the defender to boost it past that (like walls used to do in civ 5, garrisoned units, placing the city on a hill etc.).

Oh I agree, it's why domination is my least favourite victory type, its just so tedious. OTOH, there is some level of historical realism for the medieval period where the defensive technology dominated (castles) until the offensive technology (gunpowder) caught up. It will be interesting to see if the trebuchet does anything to adjust the mid-game conquest play.

Domination is tedious indeed.
What I dislike the most about it is the pacing - ancient era rush as fast as you can during that short window before the AI gets walls, then continue slowly forward with siege units, until it basically halts in the Medieval/Renaissance, followied by a mindless insta-win once you hit Industrial.
I often don't finish my games once I've gotten to Advanced Flight, because it's so obvious that I'll be winning easily within the next 20-30 turns from there on, and I might as well just start a new game at that point instead of pressing "next turn" with my brain turned off until I win.
Which is a real shame, since earlier versions of civ (civ 5) still had you fighting to deserve the win at that point as the AI had plenty of sources of AA.
True story here: I've never built a single biplane/fighter/jet fighter in civ 6 ever, hardly any late game sea units, and I've played this game way beyond 1000 hours by now.
There has never been any need against the civ 6 AI, which makes the game really lose a lot of depth in the late game.
The AI overly relies on the crutch that is walls, and utterly fails to defend itself once anti-wall tools like Bombers appear.

I doubt that the Trebuchet is gonna "fix" much, even though the forums have suggested it for ages now.
The problem with walls fundamentally remains the same - even with Trebuchets, you still have to roll your siege units into counter-battery fire (the defending city), and due to a lack of range and indirect fire on your siege unit (Trebuchet or not), often right next to the city before you can even fire.
I've had games where my Catapults were getting oneshot, and even though I could upgrade them to Bombards, the Bombards were also insufficient to significantly damage those walls before they got destroyed - hence a Trebuchet would serve no purpose in those cases.
It might be good for some edge cases, but overall the walls and city strengths are way too strong for my liking (until Bombers that is, where they are way too weak).

 

[Noble Zarkon]

I can't really recall that being a thing in civ 5

Civ IV and earlier iterations.

 

[Kupe Navigator]

I can't really recall that being a thing in civ 5, but either way, they didn't need to add the current walls just to prevent a "lone warrior" from conquering cities.
There are countless other ways they could have prevented that, like adding a minimum combat strength so that even the weakest of cities own can hold off 1-2 weak units, but needs involvement from the defender to boost it past that (like walls used to do in civ 5, garrisoned units, placing the city on a hill etc.).



Domination is tedious indeed.
What I dislike the most about it is the pacing - ancient era rush as fast as you can during that short window before the AI gets walls, then continue slowly forward with siege units, until it basically halts in the Medieval/Renaissance, followied by a mindless insta-win once you hit Industrial.
I often don't finish my games once I've gotten to Advanced Flight, because it's so obvious that I'll be winning easily within the next 20-30 turns from there on, and I might as well just start a new game at that point instead of pressing "next turn" with my brain turned off until I win.
Which is a real shame, since earlier versions of civ (civ 5) still had you fighting to deserve the win at that point as the AI had plenty of sources of AA.
True story here: I've never built a single biplane/fighter/jet fighter in civ 6 ever, hardly any late game sea units, and I've played this game way beyond 1000 hours by now.
There has never been any need against the civ 6 AI, which makes the game really lose a lot of depth in the late game.
The AI overly relies on the crutch that is walls, and utterly fails to defend itself once anti-wall tools like Bombers appear.

I doubt that the Trebuchet is gonna "fix" much, even though the forums have suggested it for ages now.
The problem with walls fundamentally remains the same - even with Trebuchets, you still have to roll your siege units into counter-battery fire (the defending city), and due to a lack of range and indirect fire on your siege unit (Trebuchet or not), often right next to the city before you can even fire.
I've had games where my Catapults were getting oneshot, and even though I could upgrade them to Bombards, the Bombards were also insufficient to significantly damage those walls before they got destroyed - hence a Trebuchet would serve no purpose in those cases.
It might be good for some edge cases, but overall the walls and city strengths are way too strong for my liking (until Bombers that is, where they are way too weak).

I thought they fixed the air combat in recent patch.

 

[Gedemon]

I thought they fixed the air combat in recent patch.

AI offensive air combat is fixed.

defensive is not.

 

[ggalindo001]

I agree with the OP on issues -- I don't know that I have viable solutions in Civ VI simply because I think combat needs to be overhauled.

1UPT is a main problem here IMO. It handicaps the AI far more than the human player, and allows the human player the ability to keep a minimal defense force throughout the game. And it prevents the AI from having effective ranged units available to defeat defenses -- due to their extreme vulnerability.

I do agree with the notion that many of these items (ranged attack, increased city strength, inability to damage units within the city, separate walls strength) are all mechanisms to keep the human player from rolling over the AI. But it also prevents the AI player from rolling all over other AI and the human player. In most games in the last few years that I have played, AI captures of other AI are not military captures but loyalty captures. Loyalty has become the mechanic for civilization collapse, not economics or military (which is more historically accurate).

I have played with a mod that allows for faster movement and limited stacking of units (1 ranged, 1 melee, 1 mounted) -- and the AI seems to handle this a lot better.

I don't want to necessarily get back to the "stacks of doom" from Civ 4, however, I think some scalable stacking could be a longer term solution to combat in this genre. For example:

In the first 2-3 ages, you can stack up to 2 units of each major category into one tile. So, 2 melee, 2 ranged, 2 bombard, 2 mounted, 2 scout. Ramping up ultimately to allow for 4 of each category when you get to the atomic/future ages (including corps/armies as 1 "unit")

Introduce the concept of indirect fire and/or collateral damage -- so bombards have the ability to damage up and down the stack, and when fired against a stack/unit that contains bombard, there is equal return fire (to allow for defense within another's turn -- risk/reward for using bombard). And if the player attacks without bombard into a stack that does have bombard, there is return fire within the turn. Fortifications (on both sides) reduce impact of collateral damage, and defensive structures (such as city defenses and encampments) further reduce impact of collateral damage.

Melee attack melee first within the stack, mounted attack mounted first within the stack, bombard across the stack, ranged attack melee first, then mounted, then ranged. This would align with historical combat IMO where ranged generally were supporting attack. Scout units also start with melee.

Remove observation balloons, but keep siege towers and battering rams. Ironically, for the first time ever, I have seen the AI build an observation balloon recently, but didn't use it properly!

All of that above is spitballing, but I do believe that major overhaul of the combat mechanics is needed going into Civ VII, to go somewhere between Civ III/IV and where we are today.

 

[Gedemon]

I agree with the OP on issues -- I don't know that I have viable solutions in Civ VI simply because I think combat needs to be overhauled.

1UPT is a main problem here IMO. It handicaps the AI far more than the human player, and allows the human player the ability to keep a minimal defense force throughout the game. And it prevents the AI from having effective ranged units available to defeat defenses -- due to their extreme vulnerability.

I do agree with the notion that many of these items (ranged attack, increased city strength, inability to damage units within the city, separate walls strength) are all mechanisms to keep the human player from rolling over the AI. But it also prevents the AI player from rolling all over other AI and the human player. In most games in the last few years that I have played, AI captures of other AI are not military captures but loyalty captures. Loyalty has become the mechanic for civilization collapse, not economics or military (which is more historically accurate).

I have played with a mod that allows for faster movement and limited stacking of units (1 ranged, 1 melee, 1 mounted) -- and the AI seems to handle this a lot better.

I don't want to necessarily get back to the "stacks of doom" from Civ 4, however, I think some scalable stacking could be a longer term solution to combat in this genre. For example:

In the first 2-3 ages, you can stack up to 2 units of each major category into one tile. So, 2 melee, 2 ranged, 2 bombard, 2 mounted, 2 scout. Ramping up ultimately to allow for 4 of each category when you get to the atomic/future ages (including corps/armies as 1 "unit")

Introduce the concept of indirect fire and/or collateral damage -- so bombards have the ability to damage up and down the stack, and when fired against a stack/unit that contains bombard, there is equal return fire (to allow for defense within another's turn -- risk/reward for using bombard). And if the player attacks without bombard into a stack that does have bombard, there is return fire within the turn. Fortifications (on both sides) reduce impact of collateral damage, and defensive structures (such as city defenses and encampments) further reduce impact of collateral damage.

Melee attack melee first within the stack, mounted attack mounted first within the stack, bombard across the stack, ranged attack melee first, then mounted, then ranged. This would align with historical combat IMO where ranged generally were supporting attack. Scout units also start with melee.

Remove observation balloons, but keep siege towers and battering rams. Ironically, for the first time ever, I have seen the AI build an observation balloon recently, but didn't use it properly!

All of that above is spitballing, but I do believe that major overhaul of the combat mechanics is needed going into Civ VII, to go somewhere between Civ III/IV and where we are today.

That's quite similar to how I modded it in civ5, and yes, it helped the AI a lot, even with just 2-3 combat units per tile and counter-fire mechanisms.

Civ6 allows 1 "formation class" per tile since a relatively recent patch, but AFAIK we can't mod counter/support-fire mechanism in 6 ATM.

I'd love civ7 to use that kind of approach, I'd prefer it over 1UPT (of course), SOD, and even the "stacked for movement / unstacked for combat" mechanism)

 

[vortical42]

I agree with the OP on issues -- I don't know that I have viable solutions in Civ VI simply because I think combat needs to be overhauled.

1UPT is a main problem here IMO. It handicaps the AI far more than the human player, and allows the human player the ability to keep a minimal defense force throughout the game. And it prevents the AI from having effective ranged units available to defeat defenses -- due to their extreme vulnerability.

I do agree with the notion that many of these items (ranged attack, increased city strength, inability to damage units within the city, separate walls strength) are all mechanisms to keep the human player from rolling over the AI. But it also prevents the AI player from rolling all over other AI and the human player. In most games in the last few years that I have played, AI captures of other AI are not military captures but loyalty captures. Loyalty has become the mechanic for civilization collapse, not economics or military (which is more historically accurate).

I have played with a mod that allows for faster movement and limited stacking of units (1 ranged, 1 melee, 1 mounted) -- and the AI seems to handle this a lot better.

I don't want to necessarily get back to the "stacks of doom" from Civ 4, however, I think some scalable stacking could be a longer term solution to combat in this genre. For example:

In the first 2-3 ages, you can stack up to 2 units of each major category into one tile. So, 2 melee, 2 ranged, 2 bombard, 2 mounted, 2 scout. Ramping up ultimately to allow for 4 of each category when you get to the atomic/future ages (including corps/armies as 1 "unit")
.

I have seen other 4x games (stellaris and GalCiv 3 come to mind) where they have stacking but it is limited by a 'logistics' level that you can increase via techs or buildings. While that would help balance out walls by allowing you to bring more force into a limited space, it wouldn't do much to improve the AI. It just swaps the issues of 1 UPT with the problem of making an AI that can understand the costs/benefits of different unit groupings and how to make the most efficient use of the logistics cap.

The problem runs deeper than just 1 UPT or walls. The real issue is that the combat AI is tactical but not strategic. If it sees a target of opportunity it will pounce on it every time with no regard for how that action effects the overall goals it is trying to achieve. A garrisoned unit will happily march into the teeth of death and leave a city undefended if it thinks it can pick off an enemy or grab a civillian. Likewise, when they go on offense every unit is doing their own thing and will chase down whatever juicy targets they run across. The result is scattered forces that can easily be picked apart.

I get that good AI is hard. There is only so much it can do without making the turn calculations balloon out of control. That doesn't mean it can't be improved. Simply adjusting the weights used to determine unit behaviour could make a world of difference.

 

[Gedemon]

While that would help balance out walls by allowing you to bring more force into a limited space, it wouldn't do much to improve the AI.

there is a condition here for this to see an improvement of AI without any other changes: playing at a difficulty level where the AI has a production bonus, which translate into more AI units on the same position (the algorithm doesn't spread them out, if it can put 10 on a plot, and determine that a plot is a good position to attack a city, it will put 10 units here, then start placing the 11th on the next best plot, which open other AI issues when you use collateral damages... and then you also need to change the behavior)

 

[Noble Zarkon]

1UPT is a main problem here IMO. It handicaps the AI far more than the human player

Agreed, the Civ IV AI wasn't great but it could use it's production bonuses to build a big stack and overwhelm you (or other AIs) with it.

 

[Kaan Boztepe]

i personally always found it ridicolous that the AI would start moving its whole army to one end of the map to attack an undefended city only to move them to another city one turn later when i move units to defend the first city and leave new target city undefended. With this simple tactic you can tie up the whole army of the AI doing nothing.
on the point of walls i dont think they should be able to attack at all if there is no ranged/siege unit inside or even provide anything but very low protection unless there is at least one military unit inside. The strength of the walls/attack should be directly tied to the unit inside as well as terrain considerations etc.
also i agree that big cities should get better defences than newly created cities , capitals even more so , this would also help city states survive.
on the point of armies at the moment there is basically no limit on the number of units that a player can create , beside resources maybe but then again if you can trade for them is it really that difficult? making the army tied to your population makes sense otherwise you can have a civ with 4, 1 pop cities and 100 units which is funny. Maybe only allow a city to make a unit if it has enough population to support it like number of districts it can construct. The militaristic civs could then get bonuses to those which would make them at least matter in the game. Another idea is to have the army require food&resources per turn like your citizens.
Then again all these ideas would most likely be the death sentence for an AI that cannot even handle the basic version we have now.