Armies

[Txurce]

There has to be a solution for affecting puppets only.

I think we've all said this now. Nerfing all of a puppet's contributions is a very simple, far-reaching fix. Balancing the numbers is just process. The real question seems to be whether there's a mechanic that can accomplish this.

 

[Thalassicus]

@Ahriman
Ahhh okay, I think I get your reasoning now: you feel ICS is no longer a concern at this point?

I've been keeping it in mind for all my decisions for several months now, finding little ways here and there to affect broad vs tall empires. If it's sufficiently balanced now in the mods that means I can stop worrying about it.

The reason for looking at "baseline" benefits (happiness per pop, research per pop, happiness per city, etc) is I'd seen several threads in the past discussing ICS, and how a primary factor is too many basic bonuses.

It seems like we all thought of the building-from-lua idea simultaneously this morning. I've been thinking it over for the past few hours and I believe I might be able to put something together without re-inventing the wheel... in particular, I know how to check puppet status (from fiddling with the emigration mod) and do checks once per turn (learned from doing promotion swaps) Combining these two... all I'll need to find is how to add and remove buildings, which shouldn't be hard. I think I can actually put it together in a rather short timeframe. I'll think about it more over lunch.

 

[Txurce]

The reason for looking at "baseline" benefits (happiness per pop, research per pop, happiness per city, etc) is I'd seen several threads in the past discussing ICS, and how a primary factor is too many basic bonuses.

I think the general opinion is that ICS is no longer head and shoulders the optimal strategy. It's still a very viable one, and probably should be. But keeping an eye on it to make sure it stays in its present, quite comfortable box is probably a good idea.

 

[Ahriman]

Ahhh okay, I think I get your reasoning now: you feel ICS is no longer a concern at this point?

I think its no longer a huge concern, I think your boosts to % yield buildings have already shifted things significantly towards favoring large cities.
I don't think the library is underpowered, and I don't think research is too high without libraries, except from puppets.

The one "baseline" benefit I would look at is trade route income. I've lost track of whether you've changed it, but at 1.25*pop it feels too high and it favors ICS a bit too much.
Alpaca changed it to (1*pop)-1. I'm not sure if those are the right values, but nerfing the trade income contribution of small cities seems a sensible thing to do. Alpaca's values are a good start point, and help reduce overall gold inflation, and mean that you have to "work" for gold more from tiles or merchants, rather than just getting it accumulate automatically from pop.

For comparison:

Size	Vanilla Alpaca
1	1.25	 0
2	2.5	 1
3	3.75	 2
4	5	 3
5	6.25	 4
6	7.5	 5
7	8.75	 6
8	10	 7
9	11.25	 8
10	12.5	 9
11	13.75	 10
12	15	 11
13	16.25	 12
14	17.5	 13
15	18.75	 14
16	20	 15
17	21.25	 16
18	22.5	 17
19	23.75	 18
20	25	 19

 

[Tomice]

I think the general opinion is that ICS is no longer head and shoulders the optimal strategy. It's still a very viable one, and probably should be. But keeping an eye on it to make sure it stays in its present, quite comfortable box is probably a good idea.

With normal number of civs on large /huge terra maps, I often find it quite hard to settle all the available terrain. While expanding is still possible, it's not really easy, even when using "avoid growth", building colloseums quickly and grabbing all happiness I can find. I tend to put more civs on those maps now.

--> I agree ICS is nerfed enough, vertical and horizontal growth both seem equally viable.

 

[Seek]

If the across the board puppet output nerf (which I agree would be ideal) turns out to not be viable, I'd suggest a basic rising city maintenance for puppets - ie, each one obtained costs more gold per turn. Positive arguments:

• A) Prevent rapid expansion from conquest for all sized empires
• B) Provide incentive to annex
• C) Limit the "two city while puppeting the world for a culture victory" option since the small empire would go bankrupt trying to support it's puppets
• D) Puppets' automatic gold focus would represent their assimilation, thus seeming more realistic

Negatively:
May create too much incentive to raze.

Though Ahriman argued that more incentive to raze wouldn't necessarily be a negative, I think that some sort of lasting penalty for razing should be added regardless. Perhaps a permanent happiness hit would be justifiable.

 

[Txurce]

If the across the board puppet output nerf (which I agree would be ideal) turns out to not be viable, I'd suggest a basic rising city maintenance for puppets - ie, each one obtained costs more gold per turn. Positive arguments:

• A) Prevent rapid expansion from conquest for all sized empires
• B) Provide incentive to annex
• C) Limit the "two city while puppeting the world for a culture victory" option since the small empire would go bankrupt trying to support it's puppets
• D) Puppets' automatic gold focus would represent their assimilation, thus seeming more realistic

Negatively:
May create too much incentive to raze.

Though Ahriman argued that more incentive to raze wouldn't necessarily be a negative, I think that some sort of lasting penalty for razing should be added regardless. Perhaps a permanent happiness hit would be justifiable.

If this is an easier approach to nerfing puppets, then all the better.

I don't understand a happiness hit for razing, since your own people aren't affected by it. The only applicable hit would be diplomatic, and the WWGD mod shows it's doable. The downside would then be worse trade relations and an increased likelihood of war.

 

[Ahriman]

I'm not sure a permanent penalty for razing is needed, at least in the current build where the AI ICS's like crazy, including building on 1 tile islands. You have to have some way to burn out their infestation.

Hopefully there's some way to reduce the AI ICS tendency and get it to focus on fewer, larger cities well. The problem is that ICS isn't a bad solution for the AI on the highest difficulty levels where its happiness bonuses are just so huge. Though I'm nervous about anything that makes the game easier, I'd lean towards toning those down.

I think: ideally, we wouldn't see so many weak location AI cities, and AI cities would be more spaced, and then we wouldn't see so much razing beyond the early-earlymid game, except occasionally for a conqueror who is expanding beyond what they can hold (but even then, a happiness penalty from razing should be a big hit, and a check on expansion).

If this is an easier approach to nerfing puppets, then all the better.

I'd definitely prefer to see yield modifiers than maintenance costs, which would be a new mechanic entirely, rather than just tweaking income values. With the income tweak, more puppets are still always desirable if you have the happiness to support them. I think happienss should be what governs expansion, not some hybrid of happiness-gold.

I don't understand a happiness hit for razing, since your own people aren't affected by it.

I don't think its too hard to imagine your own civilian population being horrified by genocide being committed in their name, particularly of "civilized" foes in more modern eras, also razing an enemy city would realistically generate a massive response; guerilla warfare, sabotage, mass unrest, etc. which would be destabilizing in general, and happiness is a catchall mechanic representing unrest, stability and administrative burden.
Much easier to puppet.

 

[Txurce]

I'm not sure a permanent penalty for razing is needed, at least in the current build where the AI ICS's like crazy, including building on 1 tile islands. You have to have some way to burn out their infestation.

Hopefully there's some way to reduce the AI ICS tendency and get it to focus on fewer, larger cities well. The problem is that ICS isn't a bad solution for the AI on the highest difficulty levels where its happiness bonuses are just so huge. Though I'm nervous about anything that makes the game easier, I'd lean towards toning those down.

I think: ideally, we wouldn't see so many weak location AI cities, and AI cities would be more spaced, and then we wouldn't see so much razing beyond the early-earlymid game, except occasionally for a conqueror who is expanding beyond what they can hold (but even then, a happiness penalty from razing should be a big hit, and a check on expansion).

I've noticed that in every game not every successful AI civ employs ICS - some go for tall growth. And because I agree that ICS is a good strategy for the AI, I think the human having to live with potential razing diplomatic penalties is acceptable. If I were going to balance this, I would tilt toward lowering the penalty (only after x razings) rather than limiting the AI's ICS approach. Adding a very conservative inter-city tile limit like Tomice once did in Civ 4 might be a happy medium.

 

[Seek]

I don't understand a happiness hit for razing, since your own people aren't affected by it.

Um, common decency? Murder enough people and the populace starts to become ashamed of the government's actions and/or uneasy about their government's ruthlessness. Of course there are examples against this argument in history, that a population would become more happy as a result of city-razing, but in general they are only for relatively brief periods. Also, if you consider unhappiness to come from all subsections of your empire (slaves as well as aristocrats) I think it's perfectly understandable.

But the diplo hit seems like a decent option as well.

 

[Polycrates]

I'm not sure a permanent penalty for razing is needed, at least in the current build where the AI ICS's like crazy, including building on 1 tile islands. You have to have some way to burn out their infestation.

Hopefully there's some way to reduce the AI ICS tendency and get it to focus on fewer, larger cities well. The problem is that ICS isn't a bad solution for the AI on the highest difficulty levels where its happiness bonuses are just so huge. Though I'm nervous about anything that makes the game easier, I'd lean towards toning those down.

Someone else's idea, but in my game I've changed the minimum settling distance from 2 to 3. Far from ideal, clearly, but it does have a pretty strong effect on preventing AI ICS. They still build in some really dumb spots (and it doesn't help with the 1-tile island problem), but overall it's a lot better.
Very ugly, blunt-instrument fix, but it does work.

I don't think its too hard to imagine your own civilian population being horrified by genocide being committed in their name, particularly of "civilized" foes in more modern eras, also razing an enemy city would realistically generate a massive response; guerilla warfare, sabotage, mass unrest, etc. which would be destabilizing in general, and happiness is a catchall mechanic representing unrest, stability and administrative burden.
Much easier to puppet.

Totally agreed. And mechanically, if you increase temporary unhappiness from razing (my suggestion was to double it), the main effect is that it's very likely to plunge an already-stretched conqueror into the deadly -33% combat penalty. It puts your troops in danger and means that you're really forced to slow down your conquest.
I don't think a permanent penalty is justified (except a permanent diplo penalty with the owner - I for one would burn with hatred for you for eternity if you razed one of my cities ), but it does seem like it should be less of a consequenceless, go-to option.
Caveat: especially with the AI as it is, I'd put a size limit (eg 4) on any serious penalties so you can burn down those really rubbish AI cities.

As far as nerfing puppets, my main suggestion would be:
an increasing +1 unhappiness for each additional puppet (i.e. +1 for the second, +2 for the third, etc)
and/or an additional increasing +1 unhappiness for each puppet in excess of your own cities
If happiness approximates stability, an empire of puppets far in excess of your own people is going to be very unstable. Mechanically, it means you can have a few powerful puppets who still give you truckloads of gold, but it's harder to do the "three core cities and an empire of puppets" exploit and at some point you have to start thinking about integrating captured cities into your empire proper.

Side thought: is it possible to move puppet unhappiness from regular unhappiness to unhappiness? Just to stop the already-super-powerful Theocracy from being so useful for a massive puppet empire, but also empowering the Autocracy one (I think it's Police State)?

 

[Buckets]

Hey Thal,

I found this rooting through CIV5AICityStrategies.xml:

	<Row>
<AICityStrategyType>AICITYSTRATEGY_IS_PUPPET</AICityStrategyType>
		<FlavorType>FLAVOR_MILITARY_TRAINING</FlavorType>
		<Flavor>-100</Flavor>
	</Row>
    <Row>
      <AICityStrategyType>AICITYSTRATEGY_IS_PUPPET</AICityStrategyType>
      <FlavorType>FLAVOR_GOLD</FlavorType>
      <Flavor>50</Flavor>
    </Row>

And it gave me the idea: What if a negative flavor for culture was added here? If it works, that means puppets would not contribute to culture. And since they don't contribute to policy costs, that seems like a fair approach. Then imagine you could give puppet ai more direction, a distaste for other things as you see fit. I do wonder however, if the puppets would resort to 'wealth' if they ran out of 'good flavor' buildings.

 

[Ahriman]

Judging from what I know about Civ4 flavors and guesswork, if you set the flavor value = -100 then they would never build any structure that provided culture. Ever.
Is that really what we want?

I don't think it should be the case that no puppet could ever expand its borders, ever. I'd rather just make them much less efficient at doing so.

Also, this would be a big buff for France, as the only civ whose puppets could expand their borders.

Its an interesting idea, but I think we're better off by making puppets less efficient at what they do, rather than trying to limit what they can do.

 

[Txurce]

Judging from what I know about Civ4 flavors and guesswork, if you set the flavor value = -100 then they would never build any structure that provided culture...

Its an interesting idea, but I think we're better off by making puppets less efficient at what they do, rather than trying to limit what they can do.

It's a great idea, with the balance probably lying somewhere south of 100, as you note.

The French advantage is one monument more than everyone else. It's definitely a buff, but I think acceptable.

 

[Ahriman]

It's a great idea, with the balance probably lying somewhere south of 100, as you note.

No, that's not what I noted.

These flavor values purely govern which structures the puppet-AI-governor will choose to build.

It doesn't affect the value of the output of the buildings.
Even with a flavor value of -100, if I capture and puppet a city that already has a culture producer in it (eg the Stonehenge wonder) then it will still pump out that full culture value; it just won't build any more buildings that have any positive culture:flavor value.

Puppets should be balanced by affecting their yields, not by messing with their AI to make them build sub-optimally.

[Another point that this reminds me of; we should go and look through the various flavor values on the buildings. For example, if there is a happiness flavor, and Opera House now gives positive happiness, then that building should have a positive flavor value.]

At least in Civ4, flavor values were solely used to determine what decisions an AI makes. They have no other purpose.

The French advantage is one monument more than everyone else

Which is a rather large buff, it no-one else's puppets can build *any* monuments.

 

[Txurce]

No, that's not what I noted.

Puppets should be balanced by affecting their yields, not by messing with their AI to make them build sub-optimally.


Which is a rather large buff, it no-one else's puppets can build *any* monuments.

1. That's what it read like.

2. I disagree, and like the proposal.

3. The buff to France is +2. What anyone else can build is irrelevant.

 

[Ahriman]

Sorry, I think I've been unclear here.
What I was trying to say is that I don't think the solution is some "intermediate" flavor value, which is very difficult to interpret without knowing a lot of details about the AI code. If you set the flavor value to -50, its also probably the case that they'll never build any culture building, unless its a 10 hammer building with no maintenance cost that gives +300 culture per turn. I don't know exactly how the AI weights "value" vs "flavor" instructions. If you set the flavor code to -5, they still probably won't build culture buildings, but they might, if it seems to be valuable enough given other construction options.

Its cool if you disagree, I just wanted to be clear.

Maybe I've misunderstood what you meant by "south of 100".

The buff to France is +2. What anyone else can build is irrelevant.

I disagree. 2 culture relative to 0 is a much bigger difference than 10 culture relative to 8.

But anyway, the key point is whether puppets should be hard-code blocked from ever building any structure that has a flavor_culture value (which is only the same as a building that actually gives culture if you make sure its that way in the xml; for example, its possible that a Wat doesn't have a flavor_culture value even though it provides culture, in which case the AI could still build it even with -100 flavor_culture).

 

[Seek]

But anyway, the key point is whether puppets should be hard-code blocked from ever building any structure that has a flavor_culture value (which is only the same as a building that actually gives culture if you make sure its that way in the xml; for example, its possible that a Wat doesn't have a flavor_culture value even though it provides culture, in which case the AI could still build it even with -100 flavor_culture).

Well if it's not possible to mod a general nerf to puppets, then this is certainly something to bear in mind.

Also adjustable is ConquestProb which controls how likely a building is to be destroyed at conquest; right now, all culture buildings get automatically destroyed (they're set at 0%), while most other buildings have a 2/3 chance of survival (set at 66%). Set culture buildings to 66% (or maybe 50%) and the fact that they will be much less likely to be constructed won't be as painful.

Alternatively, monuments could be set to 100% while all other culture buildings remain at 0%.

 

[Txurce]

Sorry, I think I've been unclear here.
What I was trying to say is that I don't think the solution is some "intermediate" flavor value, which is very difficult to interpret without knowing a lot of details about the AI code. If you set the flavor value to -50, its also probably the case that they'll never build any culture building, unless its a 10 hammer building with no maintenance cost that gives +300 culture per turn. I don't know exactly how the AI weights "value" vs "flavor" instructions. If you set the flavor code to -5, they still probably won't build culture buildings, but they might, if it seems to be valuable enough given other construction options.

Its cool if you disagree, I just wanted to be clear.

Maybe I've misunderstood what you meant by "south of 100".


I disagree. 2 culture relative to 0 is a much bigger difference than 10 culture relative to 8.

But anyway, the key point is whether puppets should be hard-code blocked from ever building any structure that has a flavor_culture value (which is only the same as a building that actually gives culture if you make sure its that way in the xml; for example, its possible that a Wat doesn't have a flavor_culture value even though it provides culture, in which case the AI could still build it even with -100 flavor_culture).

Thanks for clarifying the first point. Now I see what you mean.

You could well be right that this is not the optimal way to nerf puppets. In my mind we are looking for something that makes them second-rate compared to an annexed city. The downside to annexing is the cost of a courthouse, and that puppets don't count against culture. Courthouse costs may well be a roughly even trade for the advantage of controlling production. It's the free culture that makes them so valuable. That's why I prefer the "-100" method over nothing - it solves the bigger problem. But I can see why it's not ideal.

The reason I don't think the +2 culture matters too much is because it's not going to add up to that many meaningful tiles or extra SP's. Again, it's definitely a buff for France, but not one that puts them at the head of the pack. But the main thing is that it breaks up the culture victory exploit, even for France.

 

[Ahriman]

In my mind we are looking for something that makes them second-rate compared to an annexed city.

Well, a yield nerf does this in my mind. -30% production, gold and culture is a pretty big hit.

I take your point that stopping them constructing culture buildings isn't a bad backup option though.