There´s a problem with the AI not building/upgrading buildings

[Ruck]

I just realized that after I got my empire rank (ironman game) and could field 20k troops, later in the last 100 years even 60k troops. Wherever I had to go to wars I had to split my Army like 6 times, because the supply was so low everywhere, besides in my own counties which I had most buildings at rank 6.

That my own vassals have problems with buildings was obvious to me even earlier, as I was wondering why even my vassal De-jure Kings are paying so few taxes to me. I had f.e. one small republic duchy vassal which was over the whole game the vassal which paid me the most (although it didn´t build new buildings as well), much more than kings. It´s not that the AI is not building anything, they do, but very very rarely. I also tried to help my vassal king of Pommerania f.e. and build every +gold building in his capital up to rank 3-4 only to watch him later having 800 gold on the bank and not using to build something. A little bit later that money was gone and I have no clue for what he gave it out.

After the end date I looked in some counties f.e. Paris and London which are some of the most rich counties in the game with lots of slots and rich farmland, but even those where really low developed with buildings.

Paris f.e. :

This leads to a lot of problems like very low supply in AI countries in midgame to endgame and also very poor AI characters, as their income can not compete with us anymore and low income from our vassals.

 

[Hootieleece]

I started a suggestion thread about this.....you might want to link this post to it.

I'm not sure my theory as to why is correct.

AI needs to upgrade its castles and be more interested in warring for De Jure territory...especially in 1066 start date.

I am still in my first save. I started with the tutorial. I am King of Ireland and Earldom of Dublinn is the most upgraded castle on the map and it is 1260! I have upgraded it to a level 4 castle and all buildings to level 4. I also created more...

forum.paradoxplaza.com

 

[charals]

At least the pope won't fill up Roma with churches anymore...

 

[Invader_Canuck]

This might be an issue with how the AI budgets. Currently in the AI files, the AI budgets 75% for "short term spending" and 25% for "long term spending". Neither of these terms is that clearly defined. The AI is also told to hold a certain amount of money depending on title rank, as well as a certain war chest size depending on title rank.

Looking in the file right now, barons want to hold 25 gold for short term spending and 25 for war chest. Counts want to hold 50 and 50, dukes 100 and 100, kings 200 and 200, emperors 300 and 300. In my own testing, counts and baron level title holders seem to be the only ones who reliably will upgrade their holdings. Presumably because they have less "Don't spend any money below this threshold" restrictions. I've run some hands off games for a hundred years or so with the ratio of short term vs long term changed to 50/50. Baron and Counts seem to invest more heavily. However higher tier lords seem to still be pretty reluctant to invest. The Rurikid dynasty for example had one king in one run that had 3k gold, thousands of prestige, and he didn't want to spend a nickle on upgrading buildings.

 

[Hootieleece]

I haven't gone through the game files. My observations are just from gameplay and the fact that almost every count or above character in the game world as of 1260 has less than 100 gold except me and the Pope.

 

[denmj]

Then need to write a better economic AI for rulers. It is all math. What move makes me more money? Then modify that based on traits of the ruler. Then make the choices for the ruler.

So impatient rulers will go for the more instant feedback so more likely to conquer then build.

High stewardship would help them make the better financial choice etc.

Temperate players would be builders.

 

[Ruck]

I started a suggestion thread about this.....you might want to link this post to it.

I'm not sure my theory as to why is correct.

AI needs to upgrade its castles and be more interested in warring for De Jure territory...especially in 1066 start date.

I am still in my first save. I started with the tutorial. I am King of Ireland and Earldom of Dublinn is the most upgraded castle on the map and it is 1260! I have upgraded it to a level 4 castle and all buildings to level 4. I also created more...

forum.paradoxplaza.com

This might be an issue with how the AI budgets. Currently in the AI files, the AI budgets 75% for "short term spending" and 25% for "long term spending". Neither of these terms is that clearly defined. The AI is also told to hold a certain amount of money depending on title rank, as well as a certain war chest size depending on title rank.

Looking in the file right now, barons want to hold 25 gold for short term spending and 25 for war chest. Counts want to hold 50 and 50, dukes 100 and 100, kings 200 and 200, emperors 300 and 300. In my own testing, counts and baron level title holders seem to be the only ones who reliably will upgrade their holdings. Presumably because they have less "Don't spend any money below this threshold" restrictions. I've run some hands off games for a hundred years or so with the ratio of short term vs long term changed to 50/50. Baron and Counts seem to invest more heavily. However higher tier lords seem to still be pretty reluctant to invest. The Rurikid dynasty for example had one king in one run that had 3k gold, thousands of prestige, and he didn't want to spend a nickle on upgrading buildings.

I somehow believe that it might be an hardcoded and not so easy to solve problem. I mean it´s quite unlikely that the gametesters didnt find that bug/problem out prior to release and if it would be easy to fix, then we already had it fixed by now. I hope that I´m wrong with that assumption.

 

[Jarolleon]

This might be an issue with how the AI budgets. Currently in the AI files, the AI budgets 75% for "short term spending" and 25% for "long term spending". Neither of these terms is that clearly defined. The AI is also told to hold a certain amount of money depending on title rank, as well as a certain war chest size depending on title rank.

Looking in the file right now, barons want to hold 25 gold for short term spending and 25 for war chest. Counts want to hold 50 and 50, dukes 100 and 100, kings 200 and 200, emperors 300 and 300. In my own testing, counts and baron level title holders seem to be the only ones who reliably will upgrade their holdings. Presumably because they have less "Don't spend any money below this threshold" restrictions. I've run some hands off games for a hundred years or so with the ratio of short term vs long term changed to 50/50. Baron and Counts seem to invest more heavily. However higher tier lords seem to still be pretty reluctant to invest. The Rurikid dynasty for example had one king in one run that had 3k gold, thousands of prestige, and he didn't want to spend a nickle on upgrading buildings.

That's interesting. It seems the cause is vassal Dukes/kings going to war more often, and the AI's preferred method of warfare is trying to outmaneuver the enemy through embarking. Each time it embarks, it pays the army size divided by 100 in gold. So they save up a war chest, attack someone, hire mercs, and then spend the rest on embarking. By the time they've exceeded the war chest enough to build again, they've found someone else to fight. This explains why realm priests and barons reliably build their holdings up (they can't attack people). This is compounded by the attacker and defender both having war chests, ensuring that the defender won't go down easily and might actually win depending on the quality of the mercs they hire.

 

[Invader_Canuck]

It's probably because Dukes go to war more often, and the AI's preferred method of warfare is trying to outmaneuver the enemy through embarking. And each time it embarks, it pays the army size divided by 100 in gold. So they save up a war chest, attack someone, hire mercs, and then spend the rest on embarking. By the time they've exceeded the war chest enough to build again, they've found someone else to fight.

It's an idea, could increase the cost of embarking a bit it might cause the AI to embark less, but it also might just bankrupt them even more. I'm running a hands off game with the war chest/short term requirements lowered, and it might be having an impact. But in general it seems like the higher the title tier, the less infrastructure they do. I'm 120 years into a hands off atm, and there are cities/temple holdings that have filled out every slot, and are working on tier 2 of their building chains, while Paris has a single building in it. There are baronies and counties held by guys with 1 or at most two titles, and they are 3-4 buildings deep and some are at rank 2.

 

[Jarolleon]

It's an idea, could increase the cost of embarking a bit it might cause the AI to embark less, but it also might just bankrupt them even more. I'm running a hands off game with the war chest/short term requirements lowered, and it might be having an impact. But in general it seems like the higher the title tier, the less infrastructure they do. I'm 120 years into a hands off atm, and there are cities/temple holdings that have filled out every slot, and are working on tier 2 of their building chains, while Paris has a single building in it. There are baronies and counties held by guys with 1 or at most two titles, and they are 3-4 buildings deep and some are at rank 2.

I wonder if Paradox intended this, to discourage the Medieval II playstyle of 'Just build armies and steal the goodies from your enemies" in favour of actually building things. We all know how much they want "playing tall" to be viable lately.

 

[aqvamare]

I think gifts are the problem, in the early game, gifts are a little bit higher than low level buildings.
But for higher tiers, or later in game for higher title, you need always a number for gold, which is shorter than a building or new holding.

Result, AI throws the money to other AI, and they throw it to even more other AI, a circle were nobody update anymore.

 

[Heatth]

It's an idea, could increase the cost of embarking a bit it might cause the AI to embark less, but it also might just bankrupt them even more.

In my own experience, the AI will embark all their armies even when already bankrupt and going to a place they could have gone by land (like from South England to North England). So I imagine increasing the cost would just make the situation worse.

 

[Atalvyr]

In my game most of my vassal counts/dukes build at least some buildings, while my vassal mayors and my bishop upgrade everything very quickly. So I think the theory of the higher ranked AI spending it all on a war chest is accurate.

Basically you have to build your economy from the bottom up, by making sure the baron level vassals have the first few tax buildings in place. They will then upgrade their holdings reliably and the income will let the counts and dukes earn enough to upgrade their holdings eventually.

Also, theocratic religions have a leg up, as their bishops dont formally have a “rank” and thus don’t keep a war chest. Meanwhile they have lots of income that they then spend on upgrading stuff.

 

[Ruck]

It's an idea, could increase the cost of embarking a bit it might cause the AI to embark less, but it also might just bankrupt them even more. I'm running a hands off game with the war chest/short term requirements lowered, and it might be having an impact. But in general it seems like the higher the title tier, the less infrastructure they do. I'm 120 years into a hands off atm, and there are cities/temple holdings that have filled out every slot, and are working on tier 2 of their building chains, while Paris has a single building in it. There are baronies and counties held by guys with 1 or at most two titles, and they are 3-4 buildings deep and some are at rank 2.

What happens in your hands off game if you give one of the higher AI characters (a king maybe) lets say 100.000 gold ?

 

[Karlington]

I noticed that castles were often very neglected even in the 1400s, while cities were usually really built up - usually tier 3 or 4, and tier 4 cities with fully developed buildings in all slots weren't that rare. And considering the kind of money Realm Priests have it's no wonder that virtually all the temples are fully tricked out:

Obviously this guy is going to upgrade everything to the max - it's nothing to him, and he doesn't really have expenses.

I wanted to see if things were different, so I scrolled on over to the Muslim world where there are no Realm Priests and your normal secular rulers hold temples as well.

Things were different. The temple holdings down there were as poorly built up as the castles, while the cities were still built up very well.

I don't know if this directly relates to embarking costs as some have suggested, but I think it's pretty clear that it's rooted in the AI decision-making of "normal" rulers, since the problem doesn't exist for mayors and Realm Priests.

 

[aqvamare]

I noticed that castles were often very neglected even in the 1400s, while cities were usually really built up - usually tier 3 or 4, and tier 4 cities with fully developed buildings in all slots weren't that rare. And considering the kind of money Realm Priests have it's no wonder that virtually all the temples are fully tricked out:
View attachment 621196

Obviously this guy is going to upgrade everything to the max - it's nothing to him, and he doesn't really have expenses.

I wanted to see if things were different, so I scrolled on over to the Muslim world where there are no Realm Priests and your normal secular rulers hold temples as well.

Things were different. The temple holdings down there were as poorly built up as the castles, while the cities were still built up very well.

I don't know if this directly relates to embarking costs as some have suggested, but I think it's pretty clear that it's rooted in the AI decision-making of "normal" rulers, since the problem doesn't exist for mayors and Realm Priests.

Do mayor and realm priest even send money gifts to other ruler?
Muslim ruler with temple holdings send gifts and will be low on money.

 

[Karlington]

Do mayor and realm priest even send money gifts to other ruler?
Muslim ruler with temple holdings send gifts and will be low on money.

How do you track what gifts they send and to whom? I have no idea about any of the rulers there.

 

[Hootieleece]

I just realized that after I got my empire rank (ironman game) and could field 20k troops, later in the last 100 years even 60k troops. Wherever I had to go to wars I had to split my Army like 6 times, because the supply was so low everywhere, besides in my own counties which I had most buildings at rank 6.

That my own vassals have problems with buildings was obvious to me even earlier, as I was wondering why even my vassal De-jure Kings are paying so few taxes to me. I had f.e. one small republic duchy vassal which was over the whole game the vassal which paid me the most (although it didn´t build new buildings as well), much more than kings. It´s not that the AI is not building anything, they do, but very very rarely. I also tried to help my vassal king of Pommerania f.e. and build every +gold building in his capital up to rank 3-4 only to watch him later having 800 gold on the bank and not using to build something. A little bit later that money was gone and I have no clue for what he gave it out.

After the end date I looked in some counties f.e. Paris and London which are some of the most rich counties in the game with lots of slots and rich farmland, but even those where really low developed with buildings.

Paris f.e. :

This leads to a lot of problems like very low supply in AI countries in midgame to endgame and also very poor AI characters, as their income can not compete with us anymore and low income from our vassals.

I noticed that castles were often very neglected even in the 1400s, while cities were usually really built up - usually tier 3 or 4, and tier 4 cities with fully developed buildings in all slots weren't that rare. And considering the kind of money Realm Priests have it's no wonder that virtually all the temples are fully tricked out:
View attachment 621196

Obviously this guy is going to upgrade everything to the max - it's nothing to him, and he doesn't really have expenses.

I wanted to see if things were different, so I scrolled on over to the Muslim world where there are no Realm Priests and your normal secular rulers hold temples as well.

Things were different. The temple holdings down there were as poorly built up as the castles, while the cities were still built up very well.

I don't know if this directly relates to embarking costs as some have suggested, but I think it's pretty clear that it's rooted in the AI decision-making of "normal" rulers, since the problem doesn't exist for mayors and Realm Priests.

May I link your posts to mine in the suggestions forum? to possibly give some evidence to the devs.

 

[Ruck]

May I link your posts to mine in the suggestions forum? to possibly give some evidence to the devs.

Of course you can :=)
BTW I made also an Bug report which links to this thread : https://forum.paradoxplaza.com/foru...ly-is-building-upgrading-their-lands.1419857/

 

[Karlington]

May I link your posts to mine in the suggestions forum? to possibly give some evidence to the devs.

Feel free to link mine.