Late game happiness fluctuation.

Wumper

Chieftain
Joined
Mar 10, 2016
Messages
48
The new versions mostly fixed the happiness issues for the early and mid game. Once you get to the late game though, things get crazy. My happines fluctuated between the low 10's to the high 30's from one turn to the other, seemingly without reason (ideological pressure wasn't present). It was bearable as it only effected golden ages but it's getting crazier and crazier, in just two turn I went from having more than 20 happiness to -7 and then to -60! I did get -16 from ideological pressure but that wouldn't explain the rest.

I'm playing a very tall Netherlands game and for most of the match, happiness was manageable but at this point there's nothing to do. Building and improvement only get you so far. My city is mostly working polders with 8 food, yet my people still complain about lack of food, have 150 culture in a single city yet still have 6 unhappiness from boredom. I don't know if it's by design or just an unwanted consequence but it's simply impossible to manage this, there's literally nothing that I can do to alleviate the situation in my bigger cities short of nuking them to decrease the population.

Sorry, if this came out a bit rambly, it's just a bit frustrating, the game was the best I've ever had up until this point. I can send the savegame if someone wants to analyze it, maybe shed some light on this.
 
Not sure how to explain those sporadic jumps with the latest version, but if you've built mostly everything already, try changing city focus to the area they are most unhappy with or using processes.
 
It helps a bit but by switching all my cities to the most problematic thing (science, production/food etc.) only got me 3 happiness total, praying that it jumps back again in a few turns did the trick. I got fudged anyway in the end. For some reason my long time ally switched ideologies, started hating me and sanctioned me so I'm fudged.
 
The happiness is influenced by your tech count (compared to the median of all civs). Did you discover new techs the turn of happiness drop? Does a lot of other players discover new techs the turns of happiness buff?

And finally, how many cities do you have? Happiness being a per-city (and per-yield-type) integer, you should expect to lose or win a total of happiness approximatively equal to your number of cities each turn due to rounding going up and down.
 
I didn't discover any techs during these fluctuations. I'm not sure about other civs but it seems a bit chaotic to be caused solely by other civs tech. I Only have 8 cities.
 
The new versions mostly fixed the happiness issues for the early and mid game. Once you get to the late game though, things get crazy. My happines fluctuated between the low 10's to the high 30's from one turn to the other, seemingly without reason (ideological pressure wasn't present). It was bearable as it only effected golden ages but it's getting crazier and crazier, in just two turn I went from having more than 20 happiness to -7 and then to -60! I did get -16 from ideological pressure but that wouldn't explain the rest.

I'm playing a very tall Netherlands game and for most of the match, happiness was manageable but at this point there's nothing to do. Building and improvement only get you so far. My city is mostly working polders with 8 food, yet my people still complain about lack of food, have 150 culture in a single city yet still have 6 unhappiness from boredom. I don't know if it's by design or just an unwanted consequence but it's simply impossible to manage this, there's literally nothing that I can do to alleviate the situation in my bigger cities short of nuking them to decrease the population.

Sorry, if this came out a bit rambly, it's just a bit frustrating, the game was the best I've ever had up until this point. I can send the savegame if someone wants to analyze it, maybe shed some light on this.
Could you add when exactly did those jumps occur? Era or turn. It might be something related to techs like those that add extra food to farms, agribusiness and such. (a reason to prefer growth over raw food).
 
The jumps started to occur in the atomic era and got really bad around turn 360. Most of the unhappiness comes from lack of food and poverty. The rising poverty could be explained by the high maintenance of late game buildings but food is simply impossible to keep up with (more than 100 hundred deficit in some cases). In cities that are not focused on culture or science boredom and illiteracy are the same while poverty is a problem everywhere. My guess is that the needs don't scale too well with huge cities, most of the really problematic ones are above 30 pops. Maybe the decreased yields from unhappiness lead to even more unhappiness? That would at least partially explain the jump, going from +10 to -10 drops 20% in all yields which of course leads to a lot more unhappiness. I'm not sure when is that modifier applied though so it might not have any actual effect.
 
The jumps started to occur in the atomic era and got really bad around turn 360. Most of the unhappiness comes from lack of food and poverty. The rising poverty could be explained by the high maintenance of late game buildings but food is simply impossible to keep up with (more than 100 hundred deficit in some cases). In cities that are not focused on culture or science boredom and illiteracy are the same while poverty is a problem everywhere. My guess is that the needs don't scale too well with huge cities, most of the really problematic ones are above 30 pops. Maybe the decreased yields from unhappiness lead to even more unhappiness? That would at least partially explain the jump, going from +10 to -10 drops 20% in all yields which of course leads to a lot more unhappiness. I'm not sure when is that modifier applied though so it might not have any actual effect.
Building maintenance has no direct influence on happiness. Same for happiness malus.
(Yields for happiness are counted before all of that)

However, less gold mean less investment, so less infrastructure, so more unhappiness. So there is indirect effects.
 
Fully agree with op. Happiness is like rollercoaster. One turn you like 20+ and the next one like -20. If you lux trading partner dow you or ideology pressure rises, you are in deep trouble - and in the most cases you can do nothing about it. Most of the time it's insta retire.
 
I wouldn't mind some kind of mechanic that goes like this

You happiness cannot be more than 10 lower what is was your previous turn, regardless of how it calculates.

IDK how feasible it is though. It probably needs to be more complex to account for capturing cities.
 
The best thing you all can do is post your happiness logs to G on Githhub. That lets him see just where the swings are occurring, to what degree, and what factors may be influencing it.
 
Maybe the solution would just be having different values for the late game? At some point the needs simply become impossible to satisfy unless a city is specifically geared towards science, cutlure, production or gold. Even if a city is highly specialized unhappiness from the other needs will be overwhelming. In my experience cities start producing too much unhappiness after they reach 20 pops, usually around the industrial era.
 
I experienced something similar to OP.
During the Industrial Era I had fluctuations of up to 96 happiness from one turn to the next (I think I had something like 25 cities).
It made the game kind of annoying.
(I was playing with the 2018.06.14 build)

The happiness is influenced by your tech count (compared to the median of all civs). Did you discover new techs the turn of happiness drop? Does a lot of other players discover new techs the turns of happiness buff?
Does this mean that being above the global tech median will actually increase the amount of unhappiness you get?
I remember reading somewhere that unhappiness would be influenced by the global tech median but I assumed that being above the median would reduce unhappiness to offset the effective happiness penalty you got from having a high Science/Production ratio.
If it's the exact opposite I'll need to change my policy on vassalization.
Previously I would accept capitulation offers when the AI had essentially just one or two cities left but if being above the tech median really is disadvantegous I'll probably just wipe out as many weak AI as I can.

Edit:
Would something like the following be effective at reducing fluctuations in unhappiness:
Instead of calculating unhappiness based on current tech level, calculate it based on the average tech level of the last 20 turns.
Since it should take at most ~10 turns to research a technology it can be expected that during that time the peaks and dips in unhappiness will equal out.
This wouldn't fix the issue of rounding errors though.
 
Last edited:
I wouldn't mind some kind of mechanic that goes like this

You happiness cannot be more than 10 lower what is was your previous turn, regardless of how it calculates.

IDK how feasible it is though. It probably needs to be more complex to account for capturing cities.

Agreed, and I think this is historically more accurate. It's true that sometimes people get enraged overnight, but more often than not the populace realizes that stuff is going wrong and gets disgruntled with time and then angry when things don't change.
 
I experienced something similar to OP.
During the Industrial Era I had fluctuations of up to 96 happiness from one turn to the next (I think I had something like 25 cities).
It made the game kind of annoying.
(I was playing with the 2018.06.14 build)


Does this mean that being above the global tech median will actually increase the amount of unhappiness you get?
I remember reading somewhere that unhappiness would be influenced by the global tech median but I assumed that being above the median would reduce unhappiness to offset the effective happiness penalty you got from having a high Science/Production ratio.
If it's the exact opposite I'll need to change my policy on vassalization.
Previously I would accept capitulation offers when the AI had essentially just one or two cities left but if being above the tech median really is disadvantegous I'll probably just wipe out as many weak AI as I can.

Edit:
Would something like the following be effective at reducing fluctuations in unhappiness:
Instead of calculating unhappiness based on current tech level, calculate it based on the average tech level of the last 20 turns.
Since it should take at most ~10 turns to research a technology it can be expected that during that time the peaks and dips in unhappiness will equal out.
This wouldn't fix the issue of rounding errors though.
Tech leaders have more unhappiness. (The reverse would be kind of sadic, and increase the likelihood of the unfun "unhappiness spiral")

Your suggestions is already implemented but not exactly as you suggested it : the tech median has a delay when increasing, and should change progressively.
 
Tech leaders have more unhappiness. (The reverse would be kind of sadic, and increase the likelihood of the unfun "unhappiness spiral")
Right.
I didn't view it like that, but that is a good point.

the tech median has a delay when increasing, and should change progressively.
What do you mean by that?
The problem that I suggested a solution for is that tech scores change very sharply on the turn you research a technology.
Since (assuming the ranking stays the same) the global tech median depends only on the research of a single civilization the global tech median will simply be the tech score of a single civilization and fluctuate just as strongly.
 
What do you mean by that?
The problem that I suggested a solution for is that tech scores change very sharply on the turn you research a technology.
Since (assuming the ranking stays the same) the global tech median depends only on the research of a single civilization the global tech median will simply be the tech score of a single civilization and fluctuate just as strongly.

From memory, the game has a "pseudo-median". Each turn, if the pseudo-median is smaller than the median, it increases. (And if it is bigger, it decrease).

Which mean that when the median increase, the pseudo-median will also increase, but more progressively.
 
After finishing a game with a few large-ish swings, I'd be in favour of CrazyG's proposal of not allowing your happiness to decrease for more than 10 (or some other number) in a turn.
 
Top Bottom