• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
Nov 15, 2003
674
0
www.sanguo-online.com
Crucial Minor AI's

Although i'm currently working on my own project (Victory X mod), i think it will someday crossroads with VIP.
It would be good to know which minor AI's are needed(besides german minors) for VIP project and starting another thread for it, somewhere else seems pointless.
 
Yup. Greece always go for minimal gains and signs peace with the Ottomans accepting white peace or getting one province, even if all they have in front of them are empty provinces surrounded from the North, East and West by me (Russia).
Cheers
 
unknown-X-,

As the only one with AI experience in this forums, I created a revolter and would like it to:

1)Go for 2 claimed provinces, and dug in;
2)Dug in;
3)Pray;
4)Keep under the sand.
 
Report your findings

This sub-thread is kind of dead.

So no-one mess with the AI? Or is it the lack of a better tutorial? Let's make this thread a thread to report our findings, in special the front, the invasion sub-sets.
 
pimparel said:
This sub-thread is kind of dead.

So no-one mess with the AI? Or is it the lack of a better tutorial? Let's make this thread a thread to report our findings, in special the front, the invasion sub-sets.

I have yet to try and tinker with any of the AI settings, but I have some general observations from several GC games so far:

1. German Unification. I have yet to see this happen in any game, where I wasn't playing Prussia. Prussian AI needs to be tweaked, and probably needs several variants for different situations, with event switches. All too often I have seen the Prussian get too aggressive, and then beaten into a pulp early and often. this skews everything in europe.

2. Italian unification. Again, as with Germany, I have yet to see this happen. As with Prussia above, probably need several AI and switches for Sardinia-Piedmont, Papal states, and the Two Sicilies.

3. Russian colonialism. Way too high. In every game so far, Russia gets massive holdings in Africa, and the S. Pacific. They grab any thing and every thing they can, especially areas that historically went to France, Britain, and Germany. the settings in general are too high, as there is no way a united Germany could ever get in on the whole colony thing in 1880's. But Russia is teh worst offender in this.

4. Britain in India. right now this gets static after the first few years. No NW Frontier, no consolidation of the Raj, no adventures in Afghanistan, etc. Once the initial wars are settled, Britain becomes passive.

5. French in Annam. This has already been mentioned in other threads. but, French colonial gains against uncivilized nations is far too passive. They rarely do anything in Indochina, and rarely take on anything in africa.

6. Dutch are too militaristic. They usually have a fleet to rival Britain, and one of the largest armies in Europe quite early on.

I'm sure I'll have some more observations...
 
Well I was thinking in lines of AI values.

like:
Code:
invasion = {
	base = 2.0	
	random = 1.0	
	coastal = 3.0
	beach = 3.0
	distance = 0.5
	pocket = 2.0
	island = 1.5
	enemy = 10.0
	adjacentenemy = 5.0
	}
What the 10 in enemy REALLy means? That they won't attack province with enemies, or if each value are comparative with other or an absolute value.

But we may use this thread to report what the AI should be doing. Like France in Indochina.
 
pimparel said:
Well I was thinking in lines of AI values.


What the 10 in enemy REALLy means? That they won't attack province with enemies, or if each value are comparative with other or an absolute value.

But we may use this thread to report what the AI should be doing. Like France in Indochina.

Well, IMO, the two go hand in hand. ;) Probably from waaayyyyy too many test games of CORE. I'm hoping to be able to use that knowledge for Victoria, and reduce my learning curve. that all said, step 1 is to identify the problems, step 2 is to declare colonial war, and make the see the light. :D
 
My observations

United States -- way to easy to get the USA to join alliances. It should be virtually impossible for the US AI to join any alliance, especially with European powers, as the US was extremely averse to any involvement in continental affairs at the time. Only very late in the game, such as after 1905 or 1910, should there be any kind of possibility the US will ally with European powers. Currently, the alliance value of the base US 1836 and 1881 AI is 0, which is normal tendency to form alliance. It should be kicked up to 120 or 150, extreme isolationist level, and only come down by event.

I also suggested in the North American VIP thread that the USA needs several switching AIs to handle its shifting priorities during the game. My thoughts are in general:

Base 1836 AI -- extreme isolationism (150), Monroe Doctrine protect priorities in Western Hemisphere, research concentrated on commerce and industry, railroad buiding in the Northern US, small military builds. Industry AI focuses on fabric, glass, paper, furniture, and clothes factories.

Mexican American War AI (activated by MexAm War event chain, or other war with Mex)-- extreme isolationism, emphasis on destroying Mexico set as high as possible (100), AI targets Mexico City province for capture to increase likelihood of triggering VIP Guadalupe Hidalgo event, increased military builds. Tech goals switch to more military oriented. AI deactivated by treaty of Guadalupe Hidalgo and successful capture of American Southwest/California

Antebellum AI (switched to after MexAm War) -- extreme isolationism (150), research priorities return to commerce and industry. Factory and railroad building more of a priority, and increased weight on building steel, machine parts, heavy industrial factories.

Civil War AI (switched to by Civil War event chain) -- lessened isolationism, but Monroe Doctrine protect priorities slip as attention devoted to domestic fighting. Emphasis on destroying CSA/fighting to bitter end set high (100), and AI province targets set to the four provinces that trigger end of Civil War and reunification -- Richmond, Nashville, Atlanta, New Orleans. Military builds greatly increased, as is research. Military builds tweaked -- no more European style units like dragoons, just cavalry. Target countries to befriend include ENG, FRA, PRU, and RUS -- the first two to head off CSA diplomatic efforts for allies, and the second to secure potential allies if needed.

Reconstruction AI (triggered by end of civil war, lasts until 1890s) -- return to extreme isolationism. Military size reduced, commerce, industry, and now culture techs become priorities. Tweak industry and railroads to simulate building of historical industries and transcontinental railroad routes.

Expansionist AI (comes around 1890) -- the beginnning of American involement in the world. Still very hestitant to form alliances, probably set at 110 now, but there is some chance. Enemy target countries set to SPA, and target provinces include Cuba, Puerto Rico, Phillipines, Hawaii, etc. Also increased interest in China. Naval research and builds increased.

World War I AI (activates 1913, like the others) Sets priorities for WWI and increases chance US will join an alliance, reducing the alliance setting to probably 50 or less. The allies will still have to work it, but eventually the US will come.

Anyway, these are just some quick thoughts on how I think this key country's AI should be handled. I saw in another thread that there is a group that is looking at making these multiple AIs, and I would like to volunteer to work on the USA/CSA set once 1.03 comes out and we see what changes have been made.
 
A better distrubution of AI factories

I have changed the 1.02 ai factory file and it has really improved the ai factory choise. Some numbers may look abnorm but they are actually correct. You have to go into the aifiles of all nations you feel like changing and replace/add the factory section.

My improved file look like this:
factory = {
aeroplane_factory = 0.000
ammunition_factory = 4.000
artillery_factory = -60.000
automobile_factory = 0.000
barrel_factory = 0.000
canned_food_factory = 2000.000
cattle_ranch = 0.000
cement_factory = 10.000
clipper_shipyard = 1.500
coal_mine = 0.000
coffee_plantation = 0.000
cotton_plantation = 0.000
dye_factory = 0.000
dye_plantation = 0.000
electric_gear_factory = 1.000
explosives_factory = 2.500
fabric_factory = 12.000
fertilizer_factory = 15.000
fishing_wharf = 0.000
fuel_refinery = 0.000
furniture_factory = 5.000
glass_factory = 7.000
grain_farm = 0.000
iron_mine = 0.000
liquor_distillery = 40.000
lumber_mill = 1000.000
luxury_clothes_factory = 0.800
luxury_furniture_factory = 0.800
machine_parts_factory = 5.000
oil_rig = 0.000
opium_plantation = 0.000
orchard = 0.000
paper_mill = 2.000
precious_metal_mine = 0.000
regular_clothes_factory = 2.500
rubber_lodge = 0.000
sheep_ranch = 0.000
silkworm_ranch = 0.000
small_arms_factory = -25.000
steamer_shipyard = 1.000
steel_factory = 2000.000
sulphur_mine = 0.000
tea_plantation = 0.000
telephone_factory = 0.000
timber_lodge = 0.000
tobacco_plantation = 0.000
tropical_wood_lodge = 0.000
winery = 40.000
}
}

I have tried it out with france as my main testnation.
I haven't tried it in the electronic era so i have no number for theese factories. It is primarily done for 1860 effeciency levels.
Please try it and tell me if you find any oddities.

If i could now only get the AI to man the factories better.....
 
what is the point of using so large numbers and so small numbers? This setting is good for ai file that is used only for 1-5 years max or as an trigger ai that lasts only few years, because later in the game it is hard to know for sure which factories ai needs, and well...this might look like a good idea in the beginning but really it ain't.
NB!Only to be used for shortterm AI!
 
Have you tested them?
In that case, with what nation and which factories do you find ending up in to big numbers with this setting?

The trouble i had was that the Ai never produced steel/lumber/canned food and always lacked sulphur/explosives/ammunition/fertilizers with seminormal numbers. Even with 2000 only 5 of 60 french factories are steel mills.
 
Well I am messing with economical stats of the AI too. So far I made BRZ (3), PRG, URU and U00.

And I am making as a rule, that the AI file should be based in it's RGO:

URU only have Wool and Cattle. So I made an file which focus on fabric and canned_food.

PRG has a lot of wood, directing its AI to furniture, lumber. Making sure that as a land-locked country I removed steamer and clipper!!!

BRZ has more options, however historical in the 1st half of the century it should stay in its RGO's, so coffee, tobacco and later furniture, clipper, and etc.

A good way to improve the economical AI too is to set Railroad preferences in the file too. Big results to mine.

In the link in my signature, there is a .zip, which contains the modification mentioned above.
 
As far as colonialism, some states need to be tweaked to be more aggressive (ie Belgium) or less aggressive (Ottoman Empire, Russia, Netherlands).
For military building, I'm curious if the the AI can be hardcoded for certain nations to limit their military building?
For instance:
a) Britain is tweaked so that it builds 90% fleet, and concentrates only on land military if it gets involved in a Great War. Normally, the AI should only build a few land divisions (maybe 12...15 in UK, 15 in India, rather than the titanic amounts it now builds)
b) Multi-ethnic nations (Austria, Turkey) should probably build more divisions of various nationalities, but this problem may be due to problems with the manpower pool.
However, I'm no good at tweaking AI, so these are only small observations made by someone who can't change anything.
 
I have been doing some AI tweaking as part of my toying with the pace of colonization (see thread in main VIP section if curious). The things that I have been focusing on are the colonization path (if an AI has no colonization path at all then it will not colonize at all even if it can afford it, and once it begins to colonize it doesn't stop after its primary & secondary lists are finished. Rather it becomes opportunistic.) I've also done some tweaking on relative factory/RGO builds - the more nations that have these in their AIs the better the world market becomes as nations that produce specific goods at the start of the GC are now told to focus on those goods (be they raw materials or factory construction). I'm not overly familiar with the military aspects of the AI but one thing to focus on is the target section of invasion which allows for the creaton of more historical wars (I got a lot of practice with this for my overhaul of Vichy I scripted for CORE :D ).
 
I've only been able to play a few years so far with OHgamers Ai's, but they seem to put the colonization thing more in balance. Not perfect yet, but getting there. Once CORE 0.7 is ready, I think we can kick this up a notch. :D
 
Military Build Tweaking for AI

Indeed, after some cursory examination, it appears as if one can modify the AI to build only a set number of divisions. According to the ai_commands .txt file included with Victoria:

military = {
# May specify a certain number of divisions. -1 is dynamic scheme.
desired_army_size = -1
# May specify a certain number of divisions. -1 is dynamic scheme.
desired_navy_size = -1

mob_inf_perc = 75
# Percentage of infantry divisions that should be built to the mob pool.
mob_inf_perc = 75.000

Essentially, one should be able to program the AI of certain nations to build only up to the limit. The dynamic scheme, I assume, implies the AI will build as much as its resources will allow. Normally, this is no problem, but for AI Britain, it's a must until the manpower 'problem' is solved....ideally, one could limit Britannia to building only a small army, at all times.
Also, one can limit the percentage of reserves an enemy will have, very useful for countries like Britain and France which historically resisted the use of extensive conscription and armies of reservists, unlike say, Germany.
However, I have not played long enough in a new game to see if this feature actually works....here's hoping!
 
unknown-X- said:
oo, god, seems that this forum will never-ever spark up. I think i will write a tutorial for beginners on AI modding, hope this brings out more AI modders, because from what i have understood there are alot of event writers for VIP, but almost no AI modders.

This will be super, cause I still having trouble with the garrison values. At least my economy runs fine.
 
V0.2 AI changes

This thread is for me(and you) to monitor ai changes done/to be done:

Done:

nothing done, as v0.11 isn't released, yet

To be done:

- AI for Texas(suggested by Theodotus1)
- 3-way AI for England(for war, for peace, for unity)
- 3-way AI for Russia(for war, for peace, for total conquest)
- 2-way AI for France(for war, for peace)
- 3-way AI for Ottoman Empire(for war, for peace, for conquest)
- 2-way AI for Austria(for war, for peace)
- 1-way AI for Italy(for peace and war - 2in1)
- 2-way AI for USA(for peace, for war)
- Improved AI's for minor countries(for peace and war - 2in1)