Castle Building

[SonofWinter]

Will the AI build other stuff, rather than waiting for the medium, large and huge castle in the new patch? The AI can never afford to buy these, unless they get help, so has this been changed?

 

[unmerged(23876)]

AI doesn't seem to save a penny. even the King of Germany is penniless most times. nonetheless they build castles according to my observation. personnaly i build just until small castles except some important provinces like Essex but i see many medium castles in the AI controlled provinces if i play the game long enough.

 

[Riddermark]

new order of ai building province improvements is needed

 

[unmerged(21937)]

Ah right, I faintly remember thinking about taking a look at the issue in early september, but during my following absence I forgot all about it.

 

[binTravkin]

Maybe make AIs build orders resemble human ones:
Mine(something I definitely build first if I have an option)
Library
Road
if(thieves or warlike)Court Of Justice
if(out of demesne limit)Royal Post
if(warlike)Training grounds
etc.

 

[unmerged(21937)]

Problem is that the "if" part can't be there for AI. At least not without Johan-work. As of latest patches, I've noted that AI no longer builds Court of Justice as first building -> When my AI vassal is infested with thieves, I can't get him to build CoJ without having him build at least Forestry first.

 

[binTravkin]

Problem is that the "if" part can't be there for AI. At least not without Johan-work.

Well,

if(out of demesne limit)Royal Post

Already works at more or less efficency, I've seen muslims with their uberdemesnes rampantly building those at several occassions.

 

[unmerged(21937)]

True, it does, but it's rather an expection than the rule.

 

[G. Costantini]

AI doesn't seem to save a penny. even the King of Germany is penniless most times.

That's because the AIs, in war, just raises all their demense armies, no matter the cost (what do they care after all? Once their gold hits 0 they don't go in negative and don't suffer for teh spoliations from creditors...).

It's just one of those helping hands to the flaky AI. The bad side is, usually they don't have money to build anything above the less costly structures as they rarely reach a 500 or 1000+ money level before distributing the wealth to the troops.

G

 

[Riddermark]

the ai needs to leanr not to build in time of Ilnesses etc

 

[Lurken]

the ai needs to leanr not to build in time of Ilnesses etc

Agreed. But can it be done without precious Johan-time?

 

[binTravkin]

Doubt.
Unless there is a way to cancel a building by event scripting.
Then one oculd make an event which checks if AI province has ilness and cancel the building.

 

[Lurken]

Doubt.
Unless there is a way to cancel a building by event scripting.
Then one oculd make an event which checks if AI province has ilness and cancel the building.

Good point, now that should be possible, if there is a trigger to check if a building is under progress. If no such trigger is present, we can't do jack.

 

[unmerged(6777)]

Actually, I think that the build priority is not really the issue here. The AI isn't failing to build because it has the wrong priorities, it's usually failing to build because it doesn't have the gold required to initiate it. It spends all its gold on frequent wars, etc, and is rarely able to scrape together the funds for the larger projects (or hasn't built the prerequisites or doesn't have to net province income prerequisites).

The solution (IMO) is *not* to change the build priorities unless there's an obvious error in them somewhere. Instead, I'd look at two event sets that would help the AI to overcome its problems:

1. AI Free Gold Cheat
Script events that will periodically give the AI free gold to play with. The amount of gold would depend on the ruler tier, the current year (less free gold early in the game, more free gold later in the game when it's needed for the more expensive improvements), and wouldn't occur unless the AI ruler was at peace. This would (hopefully) be spent partially to finance new province structures...although a suddenly cash-rich AI might just decide to go to war instead.

2. Free Improvement Cheat
Script a set of events that will give the AI free province improvements. These would be scripted to occur only when the AI is at peace, when the AI has the prerequisite technology, and would be heavily influenced by the current year. So, for instance, a province event could be as follows:

province_event = { #I am an AI province that could use some help from a freebie improvement event to give me a huge castle
	ID = xxxxx
	picture = "event_default"

	trigger = {
		condition = { type = year value = 1250 }					#eariest can happen in 1275
		condition = { type = AI }							#only AI gets this event
		condition = { type = not value = { type = has_improvement value = { type = huge_castle value = yes } } }	#don't try to give a second one ;)
		condition = { type = not value = { type = atwar } }				#cannot be at war
		condition = { type = has_advance value = { type = huge_castle value = yes } }	#has necessary tech in province
		#maybe other limiting conditions? probably prevent event from placing improvment on terrain that it isn't allowed to be built on, for example
	}
	mean_tiome_to_happen = {
		months = 9600									#very unlikely to happen spontaneously

		modifier = {
			condition = { type = has_improvement value = { type = large_castle vlue = yes } }
			factor = 0.25								#far more likely if has prerequisite stucture
		}
		# maybe add some other improvement modifiers to make more likely if has tile factory, war academy, etc.

		#make it more likely to happen later in game (note that these are cumulative modifiers)
		modifier = {
			condition = { type = year value = 1300 }
			factor = 0.9
		}
		modifier = {
			condition = { type = year value = 1325}
			factor = 0.9
		}
		modifier = {
			condition = { type = year value = 1350 }
			factor = 0.9
		}
		modifier = {
			condition = { type = year value = 1375 }
			factor = 0.9
		}
		modifier = {
			condition = { type = year value = 1400 }
			factor = 0.9
		}
		modifier = {
			condition = { type = year value = 1425 }
			factor = 0.9
		}


		#ruler stats to have impact...
		modifier = {
			condition = { type = owner
				condition = { type = county }
			}
			factor = 3						#counts aren't nearly as likely to be building huge castles
		}
		modifier = {
			condition = { type = owner
				condition = { type = duchy }
			}
			factor = 1.5						#dukes are more likely to be building huge castles
		}
		modifier = { 
			condition = { type = ruler_stewardship value = 10 }
			factor = 0.9
		}
		#...etc to increase chance with high stewardship and decrease for crap stewardships...possibly also modifiers for marshal and traits....


		#province effects should hamper/aid building
		modifier = {
			condition = { type = has_province_effect value = { type = struggling value = yes } }
			factor = 5
		}
		modifier = {
			condition = { type = has_province_effect value = { type = poor value = yes } }
			factor = 2
		}
		modifier = {
			condition = { type = has_province_effect value = { type = prosperous value = yes } }
			factor = 0.75
		}
		modifier = {
			condition = { type = has_province_effect value = { type = rich value = yes } }
			factor = 0.5
		}
		modifier = {
			condition = { type = has_province_effect value = { type = struggling value = yes } }
			factor = 5
		}
		modifier = {
			condition = { type = has_province_effect value = { type = looted value = yes } }
			factor = 10
		}
		modifier = {
			condition = { type = has_province_effect value = { type = bubonic_plague value = yes } }
			factor = 50
		}
		modifier = {
			condition = { type = has_province_effect value = { type = pneumonic_plague value = yes } }
			factor = 50
		}
		modifier = {
			condition = { type = has_province_effect value = { type = malaria value = yes } }
			factor = 10
		}
		modifier = {
			condition = { type = has_province_effect value = { type = dysentery value = yes } }
			factor = 10
		}
		modifier = {
			condition = { type = has_province_effect value = { type = small_pox value = yes } }
			factor = 10
		}
		modifier = {
			condition = { type = has_province_effect value = { type = typhoid_fever value = yes } }
			factor = 10
		}
		modifier = {
			condition = { type = has_province_effect value = { type = revolt value = yes } }
			factor = 50
		}
		modifier = {
			condition = { type = capital value = yes }
			factor = 0.5
		}

		#possibly add modifier if in certain areas and/or provinces that traditionally had this improvement
		modifier = {
			condition = { type = province value = 496 } #Byzantion
			factor = 0.5
		}
		#etc...

		#other stuff...?
	}

	action_a = { #wOOt!
		effect = { type = add_improvement value = huge_castle }
		effect = { type = ruler_prestige value = 100 }
	}
}

Of course we'd have to script this sort of event for pretty much each possible improvement so it would take a while to do it and it would (understandably) increase CPU load.


All in all, though, I would prefer method #2 since it will certainly result in the improvements being build (instead of the money going to other stup[id stuff, or not getting used at all).

Either way, this would be something to discuss for 1.06.

 

[Riddermark]

I beg to differ here.

Ill give you an example - I have several counts. I can clikc in their provinve and see which buildings he can build. They are in red of course since i cant build them myself as if they were in my demense but no matter - still i can see what can be build. Then i give him cash for example 1000. As soon as a small castle is discovered it builds it first - no biggie its only 400 gold. But then.. when medium one is discovered the AI ignores all other things it can build for it.

AI free gold cheat is good though since the it rarely gets a good income except in the ricehst provinces

 

[binTravkin]

Well, recalling my latest Castile game and then one of Byzantium, I can say that AI do indeed build lots if there is some peace.
I once had great trouble getting my lands back from an archbishop who had the lands around Cordoba (the richest in Spain) and had built L4 castles everywhere..

 

[Grosshaus]

If the kind of event is made as MrT suggested, then it could have like 1% chance of erecting any of the buildings at a given time. One, or perhaps couple for building gategories, could be handled by one event.