New Beta Version - September 25th (9-25)

[Rafs]

+10%/-20% is the real cap.
If you consider "+10%" as the normal situation, it leads to a relative loss of 27% of yields when you are at "-20%".

You are are almost certainly right, since the last time I almost lost my game to a happiness spiral I've been paying lot of atention to how I tech and grow to avoid this.
What I think could be considered is not taking happiness malus/bonus yields into account when we calculate happiness, maybe this could cushion unhappiness a bit more?

 

[Moi Magnus]

What I think could be considered is not taking happiness malus/bonus yields into account when we calculate happiness, maybe this could cushion unhappiness a bit more?

That's already the case. Happiness is computed from local yield. Local yields are added to form a global yield, and unhappiness is applied on global yields.

The "unhappiness spiral" is not instantaneaous. It is:
Unhappiness -> Less gold -> Less investments -> Less infrastructure -> More unhappiness.

 

[Rafs]

It does affect food (right?), which I think in turn affects distress (not that much mind you, but less food means that sometimes you will be able to work less tiles with hammers or gold, so it can contribute to the bad situation, and not all towns will have big food excedents).

I know the spiral is not instantaneaous, though my past issues were more related to the fact that I did not have the techs to build the buildings to fend off this unhappiness, so no matter how much gold I had saved to splurge it served no purpose. And since I wasn't fixing my happiness instantly, all the bad things were happening like rebels pillaging connections, tiles and trade routes, and worst of all, cities that produced the most happiness/less unhappiness going into anarchy (I remember my capital doing this and losing 17 happiness at that time, since it was happiness neutral at the time but had 34 pop, that hurt so bad).
Because I can't tell what techs/social policies people that have happiness issues have taken, I can only speculate they either:
1-Skiped the techs related to specific key buildings as I did in the past, rushing to a specific tech for a UI or wonder.
2-Have low gold yield and/or can't save up much gold (this can be because of spying, or lack of instant gold yields+low gpt)
3-Grew too much population for the infrastructure they could construct and invest into.

What I personally would find helpful to the discussion is to see what a player can do to avoid this, and try to help with this behaviour instead of flipping all the game around what I personally believe is poor play (don't want to offend anyone, I'm not claiming to be the best player but I think we can all improve our play through tips/seing how others handled their problems).

 

[CrabHelmet]

IMO, biggest problem for Happiness is prioritizing growth over infrastructure. Every time I'm working a growth tile, I need to be able to justify it to myself: what am I growing for? If I go up by one Citizen, what would that Citizen work? Because if he is going to be working e.g. an unimproved Forest tile at 2 1, what was the point? 2is net neutral (one Citizen consumes 2), and 1 is rarely even minimally close to compensating for a new Citizen after, say, Classical at the latest. Suppose your City currently has 20. +1 is a mere 5% speed increase, which is not worth an additional +1. I see so many people assigning their Citizens to growth tiles when they have no specific goal in mind from said growth. You need to be constantly thinking: what tiles are good for me to work? What tiles are just going to slow me down?

This is a good thing for players to be thinking. The only problem I have with the present set-up is by the time you can realize you've overgrown, it is really hard to fix the problem - it takes a lot of turns to row back from low happiness, and if you accelerated into high-growth far enough, there can be sudden massive system shocks that come out of nowhere and hit you hard. Every-time someone contributes a save where they have a massive unhappiness swing, I predict they have overly populous cities, and, yep, correct every single time.

However, I am not all sure that using as a buffer works. Why not? Because it doesn't "teach" you what went wrong. Say you hit massive unhappiness, and you have your buffer. What are you going to do? Well, maybe beeline techs for Need reductions, but that's not really teaching players what they did wrong in the long-term and I think it creates a false equilibrium, where poor players will just use as a crutch because the actual solution is non-obvious.

I am strongly convinced @Gazebo should instead look at a Citizen Migration mechanic. Unhappy Citizens should consider moving to a) a City in the same Empire with positive Happiness, or b) if none is available, move to a City with positive Happiness is a nearby Civ. Why is this a better solution? Because it teaches players not to overprioritize growth. There's no point in growing if your Citizens just keep defecting, so what you'll do is lower growth... which is one of the best ways to fix Happiness problems. Now the 'release valve' is also a learning mechanism.

 

[Moi Magnus]

It does affect food (right?), which I think in turn affects distress (not that much mind you, but less food means that sometimes you will be able to work less tiles with hammers or gold, so it can contribute to the bad situation, and not all towns will have big food excedents).

If I remember correctly, unhappiness only affect growth (so food after consumption), so should not lead force you to change your tile affectation (unless you want to keep growing fast).

 

[Rafs]

If I remember correctly, unhappiness only affect growth (so food after consumption), so should not lead force you to change your tile affectation (unless you want to keep growing fast).

You could be right, I've never looked very deep into the distress calculations (don't know if it uses base food+hammers, food after consumption, or if hammer bonuses from GA or Train station actually come into play for example).

I still think that the loss of gold from unhappiness is a small factor (albeit unhelpful as you are already in a bad situation).
I think unhappiness is suposed to be a punishing situation from your mismanagement and to keep you in check from focusing too hard on a single thing while neglecting the rest. If players are falling into this situation very often, I think it is more helpful to look at why this happens and how they can fix this issue. I think CrabHelmet had more accurate analysis than me, and I personally think population only starts mattering more around the time you get public schools as this directly translates into a higher amount of passive science (unless you play civs like china or other civs which actively encourage/"trap" you into growing!).
Anyways, just my 2cents, and while I don't think the GA buffer is a bad idea, it would suck to see GAP generation nerfed to make place for a buffer I personally don't want to use.

 

[Stalker0]

G this feels like we are violating the no new systems rule. I get that unhappiness could use some adjustment but this feels like a hammer when a scalpel is needed.

Effectively we would have to ensure that no combination of GAP grants more than about 20 per turn, or the happy system just goes out the window. And there’s are things in the game right now that give 1000s of GAP.

This isn’t an adjustment, it’s a complete revamp of the yield.

We still want to go gold right?

 

[CrabHelmet]

I agree with @Stalker0. A limited migration system is both a) better than fundamentally reworking a yield, and b) less disruptive.

 

[Enginseer]

Migration is a concept that only promotes snowballing. If a civ stays really ahead of happiness and techs, the civ will be the envy of the world in population.

A migration modmod is better suited.

Also I fail to see why we need to rework GAP. This was never an issue until we see unhappiness problems.

 

[Moi Magnus]

I agree with @Stalker0. A limited migration system is both a) better than fundamentally reworking a yield, and b) less disruptive.

Some counter-arguments:
+ It may be difficult to train the AI to "understand" a migration system.
+ Would it really solve the problem we currently have? It would not prevent sudden drop of happiness, nor really help against. Loosing population is rarely a good thing, even if you are in negative happiness.
+ If done wrongly, it may be exploitable by humans. They would depopulate city with high food production (by not working specialists) and attract the population to low food production cities.
+ Snowballing effect, as said by Enginseer.

 

[Gazebo]

Calm down I never thought its a real bug, more a weird result of some calculations and city manager behavior.
Its a bit suspicious my happiness drops dramatically at entering a new era, 30 happiness in a 7 city empire (no large empire) is 4 happiness per city. This means 4 of 5 happiness values were close to jump over the gap with one tech?
Cause of that I would check what is directly happening, if such a jump happens.

How is: If anyone find an easy solution to this problem, he gets his Great Person?

Dude. You’ve been saying it is a bug this whole time. Don’t try and spin this. Own your argument at least, don’t flip-flop.

that could indeed make the drops less steep.
for example: i get a drop from 20 happiness to -30. that's like if i had a productivity mod from 120% drop down to 70%, that's almost halving the productivity. so it's no wonder the cities are unhappy, if they produce 42% less yields than before.


now, what if the %-bonuses/maluses from happiness scaled with city amount, instead of being linear... if thats possible at all?
some extreme examples:
in a 3 city empire, a change of +-10 happiness is rare, but because of the low amount of cities, it would have a big % impact.
meanwhile, in a 30 city empire, a change of +-10 happiness happens like every turn, but since the bonuses/maluses are reduced due to the high number of cities, the % impact is low. But if the happiness changes by +-100, the % impact is big again.

of course the numbers need to be crunched, but this would likely solve the problem, except for the remaining problem of "cities flip at -20" which is a stupid thing (and imho should be a modmod) anyway. oh, and players would need to evaluate happiness numbers according to their empire size.

Unhappiness penalties are calculated after the fact for the purposes of city yields. So they have zero impact on cities directly.

G

 

[CrabHelmet]

The whole reason the mechanic works is because it has a natural equilibrium effect. You don't need to understand it. As Citizens move from negative to positive happiness Cities, needs drop in negative happiness Cities and rise in positive happiness Cities, until you run out of positive happiness cities. Once you have no positive happiness cities, then Citizens start going elsewhere. The only training the AI needs is to pull back on growth if it has unhappiness issues, so as not to advantage neighbours - but my impression is that the AI already knows that it needs to pull back on growth when it has unhappiness problems, since this is already a good thing to do under the status quo, and is the entire reason I am suggesting this mechanic. So I don't think any real changes to the AI are needed.

It definitely does help happiness. Here's why: your yields are generated by B + Py, where B is the yields from buildings, P is your population, and y is your yield per Population. Your needs are given by Px, where x is the required yield. Unhappiness occurs when B+Py<Px. As P tends to 0, B+Py yends to B, but Px tends to 0 - which means that the city overall tends to a state of natural happiness as Citizens depart. Put in a way related to game mechanics - as my Citizens drop, I still get the yields from my Buildings because they don't disappear, so my yields per citizen are relatively higher and thus I can meet need requirements more easily. In fact, it is much better than a GAP buffer in this respect, because the GAP buffer does not have an equilibrating effect - it won't fix your happiness, it will just give you a slightly longer time to sort it out, and honestly, for newer players, might lead to the problem getting worse (spiralling).

I don't think this is that exploitable because honestly, I rarely have Cities with significant amounts of positive happiness. Honestly, fire up a game and see how many you have. It would be really hard to get them to move to Production cities. More realistically, at any givn point in time, most of my Empire's happiness usually comes from my capital or Empire-wide sources. As such, the only real exploit is that it might make capitals grow a little larger than normal. I don't think this is hugely problematic and can be addressed separately.

There is no snowball effect. When you have disturbingly high happiness, having less Citizens is helpful. It teaches players to reorient around Production, which means they finish infrastructure, which means the problem doesn't occur. Quite the opposite - this mechanic reaches an equilibrium, and should be implemented for that reason. The whole reason players have unhappiness problems is because they blindly think more citizens = more good, always. That's absolutely not right at all, and I don't think any top level player will say it is.

I think this is an enormously better idea than a GAP buffer, which I think is a very poor idea indeed. It effectively makes GAP synonymous with Happiness, creating a weird doppelganger mechanic, and doesn't teach players how to actually overcome their problems.

As a bonus, if you want, you can add some niche extra stuff for verisimilitude. E.g., unhappy Pops from conquered Cities prefer to migrate to their original Civ, Pops prefer to migrate to Civs with the same ideology, other fun stuff like that.

 

[CrabHelmet]

Just as a random example, I went and got my save from Vox Populi Challenge 7, with Portugal. The point of the challenge was a tech-race under hostile circumstances, and I went Tradition, which is normally the happiest tree. On turn 199 I have 27 Empire Happiness between 5 cities, which is quite high, I don't normally go much higher than 50 before Renaissance ends. However... my cities have 3, -3, -3, -2, -5 happiness respectively. All the rest of my Happiness is from Luxuries and Empire-wide modifiers. For people saying "the migration mechanics is abusable", no, it'd be really hard. The only place I can get people to migrate to is my Capital, as I mentioned earlier, and even then, that is a temporary reprieve because it can only take 3 more Citizens.

 

[tu_79]

Ok. Let me put it this way. You settle 4 cities in a row and your happiness drops under 10. You try to settle the fifth city and discover that you can't settle under 10 unhappiness. You learn to not produce settlers when you are near the limit. You have the means for preventing more unhappiness at this stage. The system works.

Next. You expand, and expand and everything seems right. You are always over 10 happiness, which is the maximum value you get production bonus. Suddenly, your happiness drops into revolt territory. Reason unknown. You seek happiness but you get just drops here and there. You discover the building that may save you ñis too far away in techs or in hammers. Meanwhile your cities revolt.

The way to prevent this is controlled growth, purposeful teching and puppets/vassals. But how can we let know players who do not come here for questions?
Could unhappy people in a city eat more food? Hovering over growth could say ' - 20 food lost to unhappy people in the city', or something like that, leading to an alarming 'starving' status. This way, cities that are generating more unhappiness grow less, reducing future unhappiness from such city, and players may notice the problem before it spirals down.

 

[CrabHelmet]

Unhappiness already reduces growth. The trouble is reducing growth is not enough, because while reduced growth stops needs increasing (e.g. it stops Unhappiness/the problem getting worse), it doesn't make needs decrease (e.g. make Unhappiness/the problem get better at the cost of something else, which is what is needed to break the feedback loop).

Whatever mechanic is reached will have something to do with forcibly reducing population. If it's not going to be Migration, it'll be Starvation or Plague or something along those lines. But I think Migration is better for the reasons I provided.

 

[Moi Magnus]

Unhappiness already reduces growth. The trouble is reducing growth is not enough, because while reduced growth stops needs increasing (e.g. it stops Unhappiness/the problem getting worse), it doesn't make needs decrease (e.g. make Unhappiness/the problem get better at the cost of something else, which is what is needed to break the feedback loop).

Whatever mechanic is reached will have something to do with forcibly reducing population. If it's not going to be Migration, it'll be Starvation or Plague or something along those lines. But I think Migration is better for the reasons I provided.

The thing I don't like about "migration" is that it implies a complex system of the pop actually moving from one city to another, so a city being influenced by nearby cities.

However, I have no problems with a local food (not growth) bonus/penalty indexed on local happiness, or an equivalent system, assuming:
1) It does not cycle with distress.
2) The starvation unhappiness is either removed from the game, or at least does not cycle with itself.

 

[CrabHelmet]

Also 3) it needs to be City-specific. You don't want Pops to be removed from Happy cities which have their needs met, that doesn't help address the problem. If we are going with just a food penalty, rather than migration (boo), then it should only be applied to unhappy cities in an unhappy empire, not all cities in an unhappy empire, and not to unhappy cities in a happy empire.

 

[Gazebo]

G this feels like we are violating the no new systems rule. I get that unhappiness could use some adjustment but this feels like a hammer when a scalpel is needed.

Effectively we would have to ensure that no combination of GAP grants more than about 20 per turn, or the happy system just goes out the window. And there’s are things in the game right now that give 1000s of GAP.

This isn’t an adjustment, it’s a complete revamp of the yield.

We still want to go gold right?

Rhetorical questions rankle me.

To be honest it is definitely a difficult balance, but in this case, it would be a pretty minor amount of new code. Just a GAP < 0 check.

Also, discussions don’t hurt anything - we can discuss changes but that doesn’t mean it is going to happen.

G

 

[Gazebo]

Talking about new systems, the migration system would be a _lot_ of new code. And I don’t think it would help that much.

 

[CrabHelmet]

@Gazebo, I still don't understand how your GAP fixes the problem. Suppose I am cruising along at 20 Happiness, hit a key tech, and drop to -40. What does the GAP do? Well, it means that I have a small buffer before the effects start kicking in. However, it doens't actually resolve the underlying issue - it is a band aid. How can I resolve the underlying issue? I need to either a) reduce needs, or b) increase the relevant yields. Reducing needs can mean teching to a new building, but I might not have time for that, or there might not even be one available for a few eras, depending on when my problems started. Increasing the yields is tough, because the GAP buffer means my growth keeps up, which means the yields I need to meet are also increasing. I just don't see how GAP solves the problem, at all.

The solution has to be something that screams at the player "STOP ALL GROWTH NOW". GAP does not do that.

If you're not going with Migration, fine, what about the Food penalty discussed above? It's not my favourite solution, but a) it would actually work, and b) it should only require minimal coding, in my impression.