Puppets shouldn't produce urbanization unhappiness + Warmonger discussion.

[ElliotS]

Venice is the worst AI civ and it's not particularly close. Domination is the rarest victory type among AI by far excluding time, and the bottom rungs of the AI balance are scattered with consummate Warmongers, who seem unable to translate their ability to take more cities into winning via domination or any other victory type. This is despite the AI getting much, much better at waging war over the development of the mod.

In player hands I've noticed at least one big problem holding Domination and Venice back, and it's that Puppets SUUUUUUUCK. Their yield caps make them offer fairly little output, which is fine, but they almost always produce LESS gold than they consume via upkeep, which seems wrong. Even worse is the title of the thread: Puppets are a happiness black hole. You can't control what they work, and then they start working a bunch of specialists and plunge your happiness into the ground. (Which incidentally makes it harder to fight.)

I've been back to playing a lot of VP (Mostly with friends on the 4UC modpack) and feel like Warmongers are a bit over-nerfed, and the data seems to back that up. That's not to say they can't work or can't win, I prefer warmongers and have won with them, but it's SO MUCH easier to play Arbia and spam GPs or Austria and win diplomacy and especially Inca and do whatever you want while you're basically invulnerable with nearly unassailable cities that start the game with INSANE yields.

It also feels like the AI is a bit too reactive to warmongering. I've had AIs request I declare war on someone with them, then when I do so and we take a few cities, get mad and denounce me for being a warmonger. Obviously someone running away with the game should be considered a threat, but certainly I wouldn't care as much about an AI warmonger taking a few cities off of someone if they're not in/near the lead. Maybe most warmongering opinion should be considered in context of how you're doing compared to the other civs? (Maybe it is, but AI friendlyness discussion should probably be a full new thread now that I think about it.)

I would love to hear feedback from everyone else, but it feels like some tweaks are in order. My thoughts to start would be: 1- Puppets shouldn't have their gold production reduced. 2- Puppets shouldn't produce urbanization unhappiness. 3- Friendly AI should be a little more chill to both casual and heavy warmongering.

 

[hokath]

Hey Elliot, nice to see a post from you again.

Just to add some information, there's some quite egregious AI tactical bugs atm that Kungcheops has found that makes them bad at taking cities. These should be fixed in the next version.
It seems like this might make the AI considerably better at translating fighting into city captures.

On the Gold issue, yes and I think this is also an issue for newer players who maybe do not realize that puppets have what is an effective gold maintenance cost.

 

[Anarcomu]

What about giving the puppets a building upkeep reduction equal to their yields reduction ? Things that specifically improve puppet yields, like Venice UA and one Imperialism policy, would increase the upkeep by the same amount.

And keep in mind that you can always either raze or annex puppets. Razing improves healing, spawn partisan that are easy to kill (and provides yields and war score), and directly increase your war score.

 

[youngsteve]

I have noticed in the test games that warmongering nations seem to fill the bottom of how nations perform in games, plus Venice. This didn't used to be the case & they used to perform better, but I have notced how puppets have performed in previous versions & nearly always annex them. Not sure sure how this works in present version but does sound a problem.

Regarding Venice. In fairness to the developers they have enhanced Venice quite a lot over the years, but the AI still struggles with it, though haven't played myself for a long time. When I did play, I used to get annoyed with how the puppets performed afterwhile, which I was powerless to deal with, so usually gave up. I remember there were limitation with buildings & certain policies where you needed multiple cities under control to take advantage off. Not sure how they can be made more playable.

 

[ElliotS]

What about giving the puppets a building upkeep reduction equal to their yields reduction ? Things that specifically improve puppet yields, like Venice UA and one Imperialism policy, would increase the upkeep by the same amount.

And keep in mind that you can always either raze or annex puppets. Razing improves healing, spawn partisan that are easy to kill (and provides yields and war score), and directly increase your war score.

That upkeep reduction would also solve my issue with puppets bizarrely taking gold instead of giving it. I just don't know if that would be as easy to code as what I am pretty sure would be a trivially easy change to allow puppets to produce full gold. Either one works fine, but puppets shouldn't COST you money.

The problem with Razing is that the AI just resettles the area, which is annoying. It also doesn't give many yields, and warscore isn't that hard to come by. I wouldn't mind some sort of a Razing buff, especially if there is a lot of pushback to the puppet issue.

Annexing them is an option, and one that becomes increasingly appealing when they grow large enough that their happiness becomes a problem. Sometimes it's just a lot to manage, and I don't think every puppet should essentially force you to annex it. Also Venice doesn't have that option and it sucks lol.

 

[L. Vern]

Regarding Venice in the AI test games:
important to note that at the time of the latest run, Venice was bugged in that they couldn't purchase units in their puppets, so that's certainly a major contributor to them not being able to take a single game.

On patch 3.10, they were above average with a 13% winrate, and even more recently on patch 4.5, they were the 7th highest winrate civ in the AI games with a 21% winrate, so I don't think they're inherently weak, at least in the context of emperor AI games.

 

[Anarcomu]

The problem with Razing is that the AI just resettles the area

Well, not a great example (because specific to a civ), but that is actually a good thing (early game) when playing the Aztec : you get immediate culture/science yields (like other warmongers) without empire size malus and you hit the 25% warscore quicker, letting you sue for peace and golden age and go to your next target right after. I prefer to vassalage others instead of conquer them, so I take fewer cities, raze them, and let them settle again when the war is over.

 

[ElliotS]

Regarding Venice in the AI test games:
important to note that at the time of the latest run, Venice was bugged in that they couldn't purchase units in their puppets, so that's certainly a major contributor to them not being able to take a single game.

On patch 3.10, they were above average with a 13% winrate, and even more recently on patch 4.5, they were the 7th highest winrate civ in the AI games with a 21% winrate, so I don't think they're inherently weak, at least in the context of emperor AI games.

Oh wow, I didn't know that. I checked the most recent data and saw them in dead last and confirmed my bias lol. That said I just checked your 4.5 thread and it seems like they're at 14.5% WR there. Certainly a fine place to be.

I still think that the changes I mentioned to puppets are probably a good idea, but if done Venice might need a slight nerf to keep it from being the best.

Well, not a great example (because specific to a civ), but that is actually a good thing (early game) when playing the Aztec : you get immediate culture/science yields (like other warmongers) without empire size malus and you hit the 25% warscore quicker, letting you sue for peace and golden age and go to your next target right after. I prefer to vassalage others instead of conquer them, so I take fewer cities, raze them, and let them settle again when the war is over.

Any idea what the requirements for the AI to offer vassalage are? How much do you generally need to raze to get them to capitulate? I'm always happy to learn new fun ways to commit warcrimes!

 

[Anarcomu]

I won't be able to give very precise information to you, but I can still give you some some good advises :

• 5 things gives warscore : killing units, pillaging tile, pillaging caravans, damaging/taking cities, and razing cities.
• The city warscore increase depends on the city's importance, with new, low developed cities weighting less, and the capital above all else, counting more.
• Razing cities have big benefits : it increase war score, allow faster healing, generate partisan units that can be killed for warscore and yields, and don't count towards empire size.
• Precision on empire size : they do count towards it, until the moment they are entirely razed, then they stop counting. You will want to adopt a new policy before choosing what to do with the city.
• Having a high level scout in the back-line, sitting on mountains or far from harm, and going for caravans, pillaging tiles and roads etc. will help increasing warscore in addition to contributing to weaken the enemy.

Now, more useful information :

• An AI that propose peace will generally only back down their proposition if you take a city.
• This mean that when reaching high warscore (80+) and they sue for peace, you can launch a military operation against the capital without hope of actually taking it :
  • If the situation becomes dire, just sign peace.
  • Damaging the Capital (memo : it is a very important city) might be enough to reach 90+ warscore, which is generally enough to offer vassalage to your enemy.

I'm looking at an ongoing games I have (King difficulty both) :

• As the Aztecs, I took and razed two low development cities from Germany, which had 4 total cities at the start of the war (I bullied them previously), then deal some damage to their capital. Vassalage ensued.
• I also took one low level city and one well developed city from the Chinese (didn't raze them), 6 cities at war start. Again, this was enough. In this case, those cities were close enough that I considered annex them later, hence I kept them.

As a general rule, if taken cities are close enough to my territory and well settled, I keep them, unless the next targeted city is well defended, in which case I might consider the warscore gain to be more important.

 

[ElliotS]

To be clear I am mostly wondering what the least damage you can do to an AI and still get it to capitulate are, which is drifting off topic of this thread. The fact is that I haven't heard any real discussions of puppets being a good option, only ways to avoid having puppets. I think they're too weak right now.

 

[hokath]

I think it's generally accepted that we're in a regime right now where more cities = better.
In this sort of world, puppets will always be weak, and you'll always want to find ways to have fewer puppets.
So I think this conversation also needs to be talking about the research/culture/tourism/etc. empire size scaling effect (which is the upside of puppets).

 

[rkkn]

yeah puppet yields should be a percentage of the net income. (income - upkeep) * puppet_mult
income * puppet_mult - upkeep doesn't make any sense. that latter is more like you're subsidizing them rather than them paying you tribute.

"We've conquered you. Now... do whatever you want and we'll pay for your buildings."
???

 

[ElliotS]

Well I don't think that more cities should NOT be better, otherwise there's no reward to the risk of expansion.

So the question is what is the point of puppets? As it is right now, crappy cities you don't want to burn seem to be the suggested use for puppets. Like when someone settles right on my northern border in the spot that barbarians keep spawning but also I don't really want to because it offers almost no yields.

If we want to keep it in that narrow use-case, which is fine, the tweaks of "no urbanization" and either "gold reduction is after upkeep" or "no gold reduction" works fine.

We could expand use-cases of puppets in a variety of ways, but I don't know if we need to. I just would like if they were a bit less of a noob-trap. (Especially in multiplayer where I try to mind my turn times and leaving more cities as puppets would help.)

 

[Voremonger]

I think the fundamental issue is that warmongering just inherently revolves around snowballing. If you play as e.g. the Aztecs and don't push your early advantage you're just going to get outscaled but if you manage to take over multiple Cities you're in a good position. So I feel like the balance is always on this knife's edge of warmongering being either very strong or very weak and that is before you even factor in humans vs. AI. For the most recent version I've only played on Epic speed where AIs were reasonably competent at warmongering (though this was done by non-"warmongering" Civs after they scaled up passively), on Standard speed I feel like the AI are always struggling to make progress against each other because they just grind up their Units.

There are some things though that the AI is not doing and that would make them more competitive vs. humans. The most important thing would be for the AI to just spam Roads on every tile near the border. The extra mobility is crucial either for defense or for rotating out injured Units when you go on the offense. In my opinion it would be better to just let every Unit use enemy Roads without promotions (since Roads are kind of BS as-is) but I suspect that that would be an unpopular change since it would make it much harder to gain an advantage.

The Range Promotion is similarly abusable, I think the AI should be more careful with their ranged Units since keeping them alive will pay off long-term. Again my opinion is that the promotion itself is the problem (which is why I don't use it myself) but if the AI gets better at abusing it as well maybe people will finally agree that it needs to be changed

 

[Voremonger]

This maybe crosses a bit into metagaming but I think in AI vs. AI combat specifically they're not doing a very good job at pillaging enemy Roads. Unless I'm misinformed the AI get a production boost when you kill their units so pillaging their Roads and increasing the latency which with they can replenish their front lines from their Cities in the back is really useful. As a human player I think I would be hurt less if some Roads were pillaged because there are comparatively fewer Units flowing from the back to the front (though it would be annoying). But if the AI were to try and flank their opponents more aggressively with e.g. Knights or Scouts in order to capture their Workers and pillage their Tiles I think that could maybe serve as a way to slow down human warmongering specifically (since any losses are more costly for them) without just giving a bunch of stat boosts/penalties.

 

[Legen]

I remember that puppets used to have a -70% yield penalty and, at the time, they were considered too good relative to annexing. It was increased to the current -80% penalty and since then, people have complained that they are too weak. We could experiment changing the penalty to -75%.

Partially related, there's an old puppet-related proposal that passed long ago, but is still unimplemented:
(2-08) Allow Purchases Of Religious Units, Religious Buildings and Worker Units In Puppet Cities

 

[Voremonger]

I think a major factor for balancing puppeting vs. annexing also was the map size. IIRC the cost increase for Technologies/Policies used to be 3-10%, now it's a flat 5% (Happiness modifier is still 150-60%). But honestly I would prefer touching some of the other downsides of puppeting than adjusting the yields %. I think it's very annoying that (to my knowledge) it is not possible to purchase Tiles while at the same time the rate at which new Tiles are acquired via Culture is reduced by the penalties. Before you get Imperialism there is also the concern that for strategic reasons you a lot of the time really want to rush Walls but the city governor selects some other building. Maybe you could make it so that you cannot directly determine what a Puppet should build but if you invest into one of the Buildings that is what gets prioritized?

 

[paul_soryu]

I never noticed puppets eating gold, probably because I always had tons of gold to spare. Right now playing as The Huns and at the end of Renaissance Era I have 25000 spare Gold (usually 5k-10k) and anything between +100 to +300 GPT without any international trade routes (because I am at war with everyone). I've left some bad coastal conquered cities as puppets because why not, but now with the knowledge that they eat Gold (+Urbanization) I question if they are really worth it:
- they take up tiles from good cities
- they cover only 2-3 resources out of range of good cities, which would quickly grow their borders to said resources
- they eat gold
- urbanization unhappiness creates too much unhappiness without a reasonable payoff in Science/Culture/Faith

Worst of all is I think Imperialism is underpowered for what game style it is supposed to support, because it has inferior Happiness bonuses compared to both Industry and Rationalism, a single happy face per city with a garrison, which also still consumes Unit Supply which I gotta say is pretty bad when you need all the units to actually do the conquests, at some point you're just gonna be hard limited by Happiness, which is maybe intended so that you couldn't conquer the world in the middle of the game? Rationalism has +1 from University and -2 to all Unhappiness. Industry has -2 Poverty, +1 from Workshops, and +3 per unique owned Luxury, which you will have a lot when you're conquering. Oh, and Imperialism can also buy GAdmirals to get +2 Happiness, but that means not buying someone better and +2 per Admiral is not a lot when you can have +3 per unique Luxury

Overall Industry just seems way better at supporting your conquests than Imperialism does:
- much more Happiness to work with
- +15% to Prod/Gold/Science/Culture is comparable even to very good doubled Monopoly bonuses, which can also be banned by World Congress if you're too successful in your conquests
- Imperialism has good raw Food bonuses, which is bad when you don't have enough Happiness, Industry has one little bonus (+10% during WLTKD)
- +2 Trade Routes translates into Food/Prod/Gold, with Gold helping if you have a lot of Poverty
- -20% Yields reduction in Puppets incentivizes you to have a lot of Puppets, but isn't it just better to raze/annex them anyway?

So it's like Imperialism is about conquering faster but less, and Industry is about conquering slower but more.

Maybe there should be a bonus to reduce the increase in Needs from Empire Size, like the one defensive buildings have? I think it is okay if creating puppets isn't that good when you are not warmongering agressively, you could conquer a couple of cities and either annex them or pay the Gold/Happiness price to have some Luxury or Strategic resources which you wouldn't have otherwise, but when conquering the world and creating puppets on scale is your intended playstyle, I think they shouldn't compound into what is essentially a huge debuff. You already have a lot of Luxuries and Strategics, you can kill anyone who resettles the place, you can wait for your borders to grow (Imperialism does help with that with +5 Culture) or use GGenerals which you probably have a bunch, so why creating a puppet at all?

The urbanization unhappinnes in puppets always seemed weird to me, I think it was discussed before and I think it was said that it is intended balance. I don't like that Specialists give you no GP points, 80%/60% less Yields, but still produce a full Unhappy face.

IIRC Imperialism was nerfed a couple of years ago after it was observed to be OP in AI games? I remember it was nerfed for some reason, Yields reduction in Puppets used to be super powerful.

 

[Voremonger]

My opinion is that once you are at the point in the game where you get the choice between Imperialism and Industry the tactical advantages granted by Imperialism are more important. Factories scale really with empire size well since each one you build increases the yields for all of your other ones. So I generally feel like I already have enough Production and I can fix my Happiness issues by building Public Works, I cannot as easily trade extra Production for tactical advantages if I go with Industry. Honestly once you get Civilizing Mission that is I think a big power spike since you can get quite a lot of yields by taking down one of the weaker AI. And after that, if you have like 20 Cities your biggest strength will be the ability to pump out large amounts of Units, which Imperialism gives you a multiplier to. (I usually play huge maps where my biggest priority in the late game is to kill off any AI that could cheese me with a Science Victory.)

 

[Anarcomu]

I find the food/prod bonus to farms and plantations to be very odd in the imperial policy, maybe something should be done about it ? the production/science from sea tiles can be useless on heavy land maps, but otherwise makes sense since you would want coastal cities for naval units if you have to conquer an oversea enemy.