AI Build Cities
Stop AI (NPCs) from constantly building only castles by changing AI’s behavior of building new holdings including castles, cities and temples to be more intelligent based on terrains, cultures, dynasty and AI personality.
In case you don’t know, due to current paradox logic of needing to have all domain limit filled to be able to build cities, and only theocracies will build temples, most of the time AI build castles. I changed the logic to take terrain and government type and culture into consideration. I have tested myself, and now the holding building for the AI are much better and smarter.
Note: Since castles are fewer, baron feudal lords would be fewer, replaced by baron mayors who has no courtiers or family (if temple and theocratic then even fewer), so this change potentially could increase some long term performance. However this is a hypothesis, not validated.
– Changed AI tendencies to build castles.
- Theocracy are unlikely but still might build castles after having one.
- Republics now are unlikely to build castles if one already there.
- Admin realms are a bit less likely to build castles.
- Much more likely to build castles on mountains, desert mountains and a little likely on hills.
- Much more likely to build castles on choke points like beside rivers or strait or coast.
- Less likely to build castles on semi-good farm lands like terraces hills or wetlands.
- Cultures like Castle Keepers would give massive likeliness to build castles. Pacifist reduce likeliness.
- Ethos including Bellicose and Stoic.
- AI’s personality, education trait, and most importantly dynasty legacy (e.g. warrior) matter.
- Integrated Liege directive mechanics of the normal buildings.
- Less likely to build castles, if a county already having 2.
– Changed AI tendencies to build cities.
- Theocracy are unlikely but still might build cities after having one.
- Republics now are much more likely to build cities.
- Admin realms are a bit more likely to build cities.
- Much more likely to build cities on good farm lands, like farmlands, plains, drylands, terraced hills, wetlands, etc.
- Much more likely to build cities beside rivers or strait or coast, or super good farmlands like farmlands, floodplains, oasis, terraced hills, and wetlands if you are able to build paddy fields.
- Cultures traditions like City keepers, Parochialism/Republican Legacy would give massive likeliness to build cities.
- Ethos including Bureaucratic and Egalitarian.
- AI’s personality, education trait, and most importantly dynasty legacy (e.g. metropolitan) matter.
- Integrated Liege directive mechanics of the normal buildings.
– Changed AI values to build temples.
- Theocracy are much more likely to build temples.
- More likely to build temples in forests, bc logging camps’ bonus for temples.
- Now more likely to build temples if cultures have related cultural traditions like Fervent Temple Builders, Monastic Communities, Religious Patronage.
- Ethos including Spiritual.
- AI’s personality, education trait, and most importantly dynasty legacy (e.g. metropolitan) matter.
- Integrated Liege directive mechanics of the normal buildings.
- Less likely to build temples if already having one.
For exact values, check the files yourself and edit as you like. AI personality are in use, however in practice if you have too much weight on their personality, there would be a lot of randomness, so this was weighed much less than I first envisioned (terrain, culture, dynasty are the main factors). Now if you find some weird counties, you could check history and see if there were a warmonger who ruled the place for a long time.
A good way to observe that the mod is working properly, is to look at Southern China or Anatolia after some years depending on your start date (the later the faster, as in later dates everyone is richer), they are rich enough and have lands that have multiple slots (but they are admin, so more cities, greeks might have some temples bc of religious patronage). Note that since everything is still calculated using chances (see Complains), there might still be some weird cases, so check multiple, as long as there are lands with two or more cities/temples (some lands come with these so take a look at game start), the mod is working.
– For total conversion mods like AGOT, EK2, LotR, PoD etc. Sometimes this mod would be fully compatible like with PoD (likely), but for others, putting this mod lower than them would overwrite their art, and sometimes have other issues. I won’t be dealing with all total conversions and test them, but I play a lot of them, so compatibility patch would be coming out slowly, and I will list them here.
– I only changed the AI values of the first ranks of these holdings, so as long as other mods don’t touch these, it would be compatible. But if they add assets like models or art, then no, you have to edit the art into the file or paste the AI values of mine to their files.
– Parameters instead of culture traditions themselves are used. If you don’t understand, it means some modded traditions would also behave accordingly with this mod.
– Preferably later in the list, after other mods that touch buildings to make this mod functional. However that would mean this mod would overwrite the other mods that touch the same buildings.
No translations are needed, as it changes only the buildings AI values.
I would also want to express how stupid it is to use probability instead of queues for different types of rulers. (YES, pointing at you paradox) And it would be a huge effort to change all the ai values of all kinds of buildings, which would also be a great mod to make ai build smarter. I’m only doing the holding for the moment, bc by using mods like RVB you could swap buildings but swapping holdings is much more difficult and labor heavy.
And the building system it’s sort of broken as you always has an optimum solution based on your playstyle. A nice post[forum.paradoxplaza.com] in the forum discussed this.
