The whole point of this message is to address too much variability of unhappiness:
Do you really think it would too swingy even with a=0.1?
Do you folks think happiness is to harsh this version 2.7?
- Thread starter josh4
- Start date
What do you mean?
My understanding is that happiness calcs are done by # of yields / # of citizens, right? Growing should always make that number go down because you work the strongest tiles first. You also divide the yields from buildings among a larger group. Needs scale linearly with population, but yields don't.
To take an example, I have a game as Carthage. On turn 200, one of my cities has 15 population. It has 90 culture, 102 science, 80 gold, 40 food, and 100 hammers (these are rounded). So per pop that is roughly 6 culture, 7 science, 5 gold, 3 food, and 7 hammers. In order for my 16th citizen to keep yields/population even, it needs to provide 28 yields per turn!
The best tiles available would be a farm (4 food 1 hammer), coast (4 food 2 gold), mine (6 production), or specialist (5 science for a scientist). I also have a few small bonuses: 0.25 production for the watermill, 0.25 culture if I had a museum (but I don't). I effectively get about 1/3 faith and 1/2 science from my religion too. So I might optimistically reach 10 yields? But there's no way to reach.
That isn't to say growing here is bad overall, but it will almost certainly be bad in terms of happiness. And if I was unhappy in this situation, locking growth would be a good move.
I actually think the median is calculated by the collection of cities at each pop.
Aka its not 1 ultimate global median. Its ..... the median of all pop 3 cities, the median of all pop 4 cities, the median of all pop 5 cities, etc etc. So when you grow, you effectively enter a new bracket of competition.
don't forget the military penalties.
Humankind uses this kind of system and I don't think it's a good idea in practice.
For instance I can look at a city and note to myself "go back here in 10 turns to build this because I predict in 10 turns it will be like this and that", and 10 turns later when I check back it can be completely different because somethings else happened behind the scene (median changes because of another city for example). You can't prevent/predict those scenarios thus you have to keep checking back every single turn, and it's a huge deal breaker.
So we actually tried this as well, we had a buffer system that preventing the changes from happening too quickly.
What this led to was the "frog on a hot skillet" scenario. Happiness would just keep dropping, slowly....but surely. And suddenly by the time the player noticed, they had a mountain of unhappiness to address, which just continued to get worse and worse (because of the buffer) while they threw everything they could to fix it.
Why in 10 turns? If you have unhappiness now the you deal with it now, because it's hard to predict future unhappiness anyway. The only thing that 'a' does, it makes the happiness changes more smooth, less swingy and that depends on the parameter.
Because that system doesn't allow me to instantly solve the unhappiness now, as it only changes 'a' amount every turn and I need happiness to be at 'x' amount to do specific thing.
And while I was waiting/doing something else that 'a' can change direction due to a city far away finished a wonder and raised the median or some city got razed and drop the median, so I can't even be sure it would be as predicted or anywhere close to it when I check back (thus the need to check every turn)
And while I was waiting/doing something else that 'a' can change direction due to a city far away finished a wonder and raised the median or some city got razed and drop the median, so I can't even be sure it would be as predicted or anywhere close to it when I check back (thus the need to check every turn)
Ok, that makes sense too. So for too big 'a', like 0.9 unhappiness would be too swingy and for too small 'a', like 0.1 it would be too much like a "boiling frog". However, there is an infinite number of values between 0.1 and 0.9, so one of them has to fit.
Rekk
Deity
- Joined
- Dec 9, 2017
- Messages
- 2,725
Two factors that you have neglected is that researching technology increases needs modifiers, and the act of growing reduces needs modifiers. As you traverse the technology tree, the happiness cost of your city growing can only increase and your city will fall further and further behind in being able to produce the infrastructure required to alleviate it.
You can only delay happiness problems for as long as you've locked growth, and your city will become more and more irrelevant if it doesn't keep growing.
It's the first time I've built public works or ticked the stop growth box in over a year, so it's a marked improvement in making happiness relevant. I don't really follow needs requirements themselves because there's usually either no options to satisfy them at the moment or the numbers inflate too much as the game advances that public works are your only answer. I'm sure they're influenced by civs and cities with strong yield biases but it's not something I engage with as a player.
You can't create enough buffer to compensate for sudden war, else you would be gimping yourself for having a big chunk of extra happiness for no reason (or spending extra money just to keep a buffer)
This is a losing battle, man.
The growth modifier is -0.5%, so instead of 6 culture I would need 5.97. I don't think that changes the math. Tech increases make it harder to achieve, not easier.
In practice your city's unhappiness is just capped at # of citizens most of the time, so yes growing absolutely hurts your happiness and locking growth absolutely will help. It also pushes your citizens away from food and onto other yields as much as it can, which itself will help happiness, or help you to get buildings/social policies that will.
Even in theory, the math formula is # of yields / # of citizens.
The top increases less than linearly because you have fixed sources (a monument always give 2 culture, never more), while the bottom increases linearly. In order for the top to be better you would need the new citizen to work a tile stronger than what your old citizens did, which doesn't make sense.
I get your point that a bigger city can build better infrastructure, but realistically that won't offset. Even if the 21st citizen in a city builds a museum completely by himself, he'll probably just break even in boredom while still dragging down the other 3. Loading savegames and checking its actually pretty rare for another citizen to break even in a single one of the happiness categories, nevermind all 4 of them.
I did not know this. This would explain why the median fluctuates like crazy, many of those pools aren't very large. Also I imagine cities would sometimes be their own median? Which means they are probably unhappy because needs modifiers usually have more increases than deductions.
Rekk
Deity
- Joined
- Dec 9, 2017
- Messages
- 2,725
It's not. It retrieves the average yield per turn in each city, sorts them in one big array, then chooses (currently) the 55th percentile.
The "base need requirement" for the turn is 45% of last turn's "base need requirement" and 65% of the "median" found this turn.
Last edited:
Yes, you can. That is simply playing safe. It's a good decision when unhappiness occur and bad otherwise. Just like you build defensive structure just in case someone attacks you. It's like an insurance policy that has it's cost.
This means that small maps are likely getting screwed a bit more on happiness, as the fewer the cities in the world, the more the "capitals" will dominate the top slots, and so the lower your regular cities will become. Though on the other hand the global needs increase per city should be less, so perhaps that balances out.
