Late Game Conquest

Thalassicus

Bytes and Nibblers
Joined
Nov 9, 2005
Messages
11,057
Location
Texas
Shifting a conversation about late-game conquest over here. :)
Spoiler :
I also think Autocracy needs to be a little stronger, simply because you get it so late, and if you ARE going for a Conquest victory - which I did last game - you're pretty much finished by the time it opens up to you. I don't know how to fix this really, since Conquest victories are always done and dusted by the Industrial era, especially on the slower speeds.
I know what you mean. Lately I've been trying to focus on what can make the end game better - that is, more competitive - as opposed to just longer. I want to get a few games under my belt with all the new changes, then try to come up with something.

The problem is conquest becomes too easy once we've conquered a lot. The tricky part is how to make late conquest harder for people pursuing a conquest victory, but still possible if we start warfare in the late game to support a different victory. In other words, it's not exactly a late game problem. It's a steamroller effect. Once we have enough power nothing can stop us.

This is why one of the first things I did is eliminate the +50% experience policy from the Honor tree. A similar idea is late game experience buildings (like the Military Academy) could provide a larger bonus than early buildings. This would let late players "catch up" with the more-experienced units of players who started conquest early. Our old units would face tougher enemies.

We could also try an economic approach. For example, I could re-introduce the maintenance cost on courthouses. The problem with that is although it does reduce our power, it's not fun to pay such high maintenance for a building that doesn't have any visible per-turn benefit. Hmm... this could tie in with an idea I believe EmperorFool made a few weeks ago. I could make the courthouse show its actual happiness benefit on the tooltip. That would give it a more tangible value.

I think one fundamental issue is the balance between puppets, occupied, and annexed cities. I feel the bonuses on them right now are too complicated and don't slow down conquest enough.

Cities look like this right now:

All Cities
  • +60 :c5production: cost for all national wonders.
Self-Founded
  • 4 :c5angry: anger per city
  • 4 :c5angry: anger per 4 :c5citizen: population
  • Adds to policy costs.
Courthouse (annexed)
  • 4 :c5angry:
  • 4 :c5angry: per 4 :c5citizen:
  • 3 :c5happy: from Police State policy (from courthouse).
  • Adds to policy costs.
  • Can be razed.
:c5occupied: Occupied
  • 5 :c5angry:
  • 5 :c5angry: per 4 :c5citizen:
  • 0 :c5science: per 1 :c5citizen:
  • -25% :c5production::c5gold::c5culture:
  • -25% :c5food: surplus
  • Adds to policy costs.
  • Can be razed.
:c5razing: Razing
  • Identical to occupied. A "razing" city is just an occupied city losing 1 population per turn.
:c5puppet: Puppet
  • 3 :c5angry:
  • 4 :c5angry: per 4 :c5citizen:
  • 0 :c5science: per 1 :c5citizen:
  • -25% :c5production::c5gold::c5culture:
  • -25% :c5food: surplus
  • No policy cost increase.
  • Cannot control production.
  • Cannot purchase tiles.
(continued below due to limit of 30 icons per post)
 
I'd like to simplify that somehow, while also making it harder. Here's one possibility for how to change it:

Courthouse (annexed)

  • Add -3 :c5gold:/turn upkeep.
:c5occupied: Occupied

  • Remove the % yield modifiers.
:c5puppet: Puppet

  • Remove the % yield modifiers.
  • Raise the unhappiness per population.
This would keep the theme of puppets are better for small cities, annexed better for large ones. I feel experimenting with -25% yield mods for puppets did not improve the game much. It made it more complicated without giving a big impact. It's the same problem that existed with the old style of "corruption" in earlier Civ games. Even if our newly-conquered cities generat no yields, it doesn't matter in the late game, because we're just rolling through them on the way to victory. There needs to be some sort of empire-wide disadvantage instead of local disadvantages.

I think the ultimate problem is the removal of city maintenance in Civ 5. It was much easier to balance things in Civ 4 where each city had an inherent cost to the empire.
 
I think a couple of those should say 1:c5science: per 1:c5citizen:, no? :shifty:

If I'm understanding you correctly, I think eliminating or drastically reducing :c5science:Science output from puppets would greatly help balance runaway/military superpower Civs.
 
Puppets have had the science penalty since 2010; that part's the same as before. The changes I'm proposing are to the happiness and % yield modifiers. :)
 
Oh, okay... I had no idea! I'm usually not much of a warmonger.

I'd think changing the Police State policy to something like "-1:c5angry: from :c5puppet:Puppets" would make more sense than "+3:c5happy: from Courthouses," which makes annexed cities better than self-founded ones.

Is the problem that annexing and plopping down Courthouses isn't difficult to do, that mass Puppeting is too easy, or neither of these? If one of the first two, it seems to me that upping those :c5angry: values would be the simplest way to do the trick.

For the record, I liked the -25% Puppet yields from a thematic perspective. If they don't work in actual gameplay, though, I suppose they must go.
 
The problem is conquest becomes easier as we do more conquest. It eventually crosses a threshold where we roll over everyone without opposition. We need something to counter that. Changing the Autocracy tree would not help, because the success of a conquest game is usually determined in the ancient through renaissance eras. By the Industrial era we're just running over people.

I haven't figured out a specific reason why this problem exists, or a specific solution, so I want to get new viewpoints. :)
 
I see – in that case, I think I should leave it to more experienced warmongers to provide their input! :D
 
Is it possible to script the AI to more actively defend if threatened? The main problem with easy conquest for me is unchanged from vanilla, once the AI's army is defeated it's a cakewalk for the rest of their empire. I was looking at the XML recently and noticed VEM doesn't increase the defensive units per city form vanilla (they're both one) - what if the AI was scripted to build two defensive units per city once it loses a city or something similar?
 
One way to remedy this problem would be by adding rebellions through events. Occupied/annexed/puppet cities have higher chance for rebellion. The rebel chance would also depend upon the size of ur empire.
Rebels would get good terrain promos & maybe city attack bonus. If u loose the city, the old civ would be revived again if possible or they'll simply plunder it like barbs (depends upon the modding tools :)).
The result would be that u may have to keep some experienced units at homeland to protect the conquered cities, slowing down the steamrolling effect. So what do u think about it ? :)
 
A conquest win is the only VC the AI absolutely cannot achieve. Given its lousy combat abilities, the only way it can delay a human conquest victory is via science and production output. The higher the output, the more likely the AI will find itself in an inherently stronger position (more cutting-edge units).

To achieve this, we could once again increase the AI production rate by era, with a sharper rise in the late game. This is a very simple adjustment that raises the odds of the big civ or two on the other continent putting up a long enough fight to take the human conqueror into the modern era. (There may be no need to increase its science rate with the production boost, but we could experiment with raising that as well in the late game.)
 
Is a +1 range building feasible? It would fix a LOT of issues, especially in steamrolling the AI with a tank and fteen thousand artillery units.
 
Make the AI more eager to garrison cities with RANGED units if possible.
Give the AI 25 % combat bonus ONLY versus human player :D ( probably not possible i know )
Or/and make puppet cities loose all and be unable to build new defensive buildings - harder to defend. That should be easy to code by making conquered city loose the def building and governor would be set to never build new one - its off the reality but at this point who cares ? :p

Not on topic but : When u raze conquered city u get a notice of it getting "new" citizen instead of lost one. Its just notification visual bug, but confusing/annoying.
 
Seems to me that the AI's wait way too long before it recognizes the fact that the player is actually trying to conquer the world. Thus allowing the player to take on the other Civ's one by one right up until the end.

Oh sure they get angry but isn't it pretty clear that if you backstab a Civ and then either wipe them out completely or wipe out every miliatary unit they have that you are warmongering?

AI's should be able to recognize this and form Alliances against the aggressor.Then recognizing that in order to fight a battle you must get your troops to a point where they can actually fight said battle. Declaring war then sitting back until the enemy is at your border doesn't work.

Plus adding a bit of actual Tactical Intelligence to the AI would help.
 
Puppeted Cities:
0) I would prefer the current yield reduction be kept.
1) Reduce science output (if possible) to zero, not just zero per population; if necessary make every science building conquestprobability=0 so they at least would have to be rebuilt.
2) Reduce their trade route income by at least half (if cannot accomplish via trade route per the city, reduce the city income by using a similar formula, even if still revolting, perhaps even if not connected). Alternatively, a simple -1 gold per population.

Annexed (courthoused) Cities:
1) PROHIBIT purchase of courthouses (I have always purchased them), or at least double or triple their current purchase price. I wouldn't know about building cost structure as I've never played at such a difficulty level as to need to build them; I would like to not have an alternative.
2) Raise unhappy to 5 for the city, keep the 1.0/pop.
3) Instead of (or in addition to) a piddly maintenance cost (annexing is for LARGE cities), courthouses get -10% city gold.


I usually go on a conquest spree in late game, just to increase my endgame score. One of the greed factors in this is world wonders & their score value (25 each).
So, please: reduce their score value to 0 or 1, and reduce their conquestprobability to less than 100 (e.g., 60-80).

Give partisan units experience per age similar to military City States (or greater) to assist in resisting the steamroller. Remember, the poor schmuck has had his unit supply capacity reduced so he needs all the help he can get with quality.
 
Puppeted Cities:
0) I would prefer the current yield reduction be kept.
1) Reduce science output (if possible) to zero, not just zero per population; if necessary make every science building conquestprobability=0 so they at least would have to be rebuilt.
2) Reduce their trade route income by at least half (if cannot accomplish via trade route per the city, reduce the city income by using a similar formula, even if still revolting, perhaps even if not connected). Alternatively, a simple -1 gold per population.

Annexed (courthoused) Cities:
1) PROHIBIT purchase of courthouses (I have always purchased them), or at least double or triple their current purchase price. I wouldn't know about building cost structure as I've never played at such a difficulty level as to need to build them; I would like to not have an alternative.
2) Raise unhappy to 5 for the city, keep the 1.0/pop.
3) Instead of (or in addition to) a piddly maintenance cost (annexing is for LARGE cities), courthouses get -10% city gold.

I usually go on a conquest spree in late game, just to increase my endgame score. One of the greed factors in this is world wonders & their score value (25 each).
So, please: reduce their score value to 0 or 1, and reduce their conquestprobability to less than 100 (e.g., 60-80).

I don't play for score, because it's too easy to game - the way you do, and generally by dragging out. But that's the scoring system, and lots of players like it.

The puppet and courthouse nerfs could slow conquest down slightly, but don't address the core issue: that the snowball starts rolling downhill too soon.
 
Lots of great ideas here! I've put an arrow next to ones that look feasible. :)
What if the AI was scripted to build two defensive units per city once it loses a city or something similar?
This is a great idea, but most of the AI variables affect the whole game. I don't remember seeing anything that might give us fine-tune control like that.
we could once again increase the AI production rate by era, with a sharper rise in the late game.
arrow-right-icon.png
This would be easy to do.
Is a +1 range building feasible?
One variable set before the game starts controls the range of all cities, so it's not possible for a building to alter that. For over a year I've wanted late game defensive structures to increase the range and hitpoints of cities.
Make the AI more eager to garrison cities with RANGED units if possible.
This is part of the tactical AI, which we have very little control over.
Or/and make puppet cities loose all and be unable to build new defensive buildings - harder to defend.
arrow-right-icon.png
It would be simple to make puppet governors less likely to rebuild defensive structures. (Those buildings are destroyed in all captured cities in vem.)
AI's should be able to recognize this and form Alliances against the aggressor.
I'd like this too, but I haven't seen any diplomatic AI variables which might encourage it. I've looked through the variables many times, but there's literally hundreds of them, so I might have overlooked something.
1) Reduce science output (if possible) to zero, not just zero per population; if necessary make every science building conquestprobability=0 so they at least would have to be rebuilt.
2) Reduce their trade route income by at least half (if cannot accomplish via trade route per the city, reduce the city income by using a similar formula, even if still revolting, perhaps even if not connected). Alternatively, a simple -1 gold per population.
1) is already in the game. Libraries are destroyed, universities have no effect, and I gave puppet AIs an aversion to rebuilding them.
2) I reduced trade income significantly in v131.1.
1) PROHIBIT purchase of courthouses (I have always purchased them), or at least double or triple their current purchase price. I wouldn't know about building cost structure as I've never played at such a difficulty level as to need to build them; I would like to not have an alternative.
2) Raise unhappy to 5 for the city, keep the 1.0/pop.
3) Instead of (or in addition to) a piddly maintenance cost (annexing is for LARGE cities), courthouses get -10% city gold.
1)
arrow-right-icon.png
would be easy to do.
2)
arrow-right-icon.png
I'm okay with.
3) I think you've got the logic reversed?

Small city with 6:c5gold: income
3:c5gold: cost = 50%
10% cost = 0.6:c5gold:

Large city with 30:c5gold: income
3:c5gold: cost = 10%
10% cost = 3:c5gold:

One of the greed factors in this is world wonders & their score value (25 each). So, please: reduce their score value to 0 or 1, and reduce their conquest probability to less than 100 (e.g., 60-80).
arrow-right-icon.png
I think this is possible, though I haven't researched score variables so I can't confirm it for sure.

Give partisan units experience per age similar to military City States (or greater) to assist in resisting the steamroller.[/I]
They should already get experience (for the AI).
 
The puppet and courthouse nerfs could slow conquest down slightly, but don't address the core issue: that the snowball starts rolling downhill too soon.

What do you think of significantly restricting unit supply: e.g., puppet cities contribute zip, lower support from population?

Remember that supply applies only to military land units; not sea/air/support.
 
What do you think of significantly restricting unit supply: e.g., puppet cities contribute zip, lower support from population?

Remember that supply applies only to military land units; not sea/air/support.

I think it would help, but not make a major difference. (I often conquer with next to no puppet cities, just selling what I don't occupy.)

My sense is that the human typically conquers due to an unbeatable tactical advantage combined with a likely tech edge and an increasing numerical edge. He doesn't need all three to blow away the AI. That's why I try to focus on how to give the AI a scientific/production advantage, with some combination of human nerfs and AI boosts.
 
  • Change The Forbidden Palace to not reduce population unhappines from puppets (if possible, i feel bad for making you work on that granary :sad:)
  • Change the autocracy opener to not include puppets. I think the opener is just a bit to strong for puppet empires. Or reduce the effect to 0.5 happieness.
 
Back
Top Bottom