• We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.

Stellaris Dev Diary #192 : Perfectly Balanced, As All Things Should Be...

Hello!

This week we’re going to look at some more changes we're planning, as well as a review of how some of the experiments mentioned in the last few dev diaries have evolved.

Thank you for the massive amount of feedback in those threads.

Reduction in Pops

Due to the effects on performance and a desire to reduce the micromanagement burden in the mid to late game, some of the things we’ve been deeply looking into are different ways of dramatically reducing the number of pops in the galaxy.

These experiments have generally revolved around modifying the growth (or assembly required) for pops as an empire’s population grows, with some variants trying a logistic pop growth (where growth follows an S-shaped curve as planets develop, based on a carrying capacity of a planet). These experiments have reduced the end date pop count to somewhere around one half of the old numbers with the expected performance improvements.

Organic pops will follow a curve where they begin at standard population growth, increase growth as the approach a midpoint between population and the planetary carrying capacity, then slow down to zero as they reach the top of the curve. Pop Assembly, on the other hand, is generally slow but consistent. The biggest change is that producing a new pop no longer costs a static amount of pop growth - it increases as the empire population does.

A significant reduction in pops has a cascade of major implications for the overall economy, production, and other gameplay effects. As such, these also require a pass on buildings, technologies, and even seemingly minor ripple effects like what the value should be for the trade value generated by pops.

There will be a lot of patch notes.

Most buildings have been standardized to now give 2 jobs per tier rather than the old 2/5/8 progression.

1605711331057.png

Just one example of many.

We’ve also changed a few buildings to have new or additional features, such as the Spawning Pool and Clone Vats, which have had their Pop Growth modifiers replaced with the new Organic Pop Assembly. This fills the same slot on the planet as Robotic Pop Assembly, so generally you’ll want to pick one or the other. (Clone Vats also picked up a food upkeep cost to represent simple materials to break down.)

1605711370874.png
1605711378849.png

Pops is Soylent Green!

A few other jobs got minor perks added to them, like the Medical Workers from Gene Clinics making it a little easier to live on less hospitable worlds.

1605711434441.png

Doesn't normally produce exotic gas, this one happens to be a lithoid.

And a few new techs have been added to help compensate for lost productivity. One tech line increases both the job production of a planet as well as job upkeep - those fewer pops are still capable of producing the work of more on a developed planet.

Ring Worlds

As part of the balance pass, Ring Worlds have been bumped up to 10 segments from 5, and the jobs per segment have been adjusted.

1605711480292.png
1605711496833.png
1605711511728.png
1605711521188.png
1605711530973.png

The Shattered Ring origin now possesses a warning that it may be a Challenging Origin for Lithoids due to a scarcity of minerals, and now also applies the Ring World Habitability Preference to your pops. We’re considering adding a similar warning for Hives selecting the origin, since the habitability preference change puts a serious crimp in their expansion.

1605711541929.png

Put a ring on it?

Their starting blockers have also been adjusted to give a more balanced spread of jobs.

Ecumenopoleis

Like the Ring Worlds, these start with all building slots open. As mentioned before, you can now use the Arcology Project decision on a planet that has a mix of City and Industrial Districts.

Note: Empire has all technologies but no traditions active.
1605711566787.png
1605711585016.png
1605711593479.png
1605711601512.png

The ecumenopolis has a unique distinction of being able to have both the Factory and Foundry building lines on the same planet.

Habitats

The changes to Habitat modules are much smaller in scope, but here’s the list of their districts.

1605711621995.png
1605711632042.png
1605711641350.png
1605711651091.png
1605711658034.png

Void Dwellers have gotten a bit of attention as well with some tradition swaps for those that had minimal or no beneficial effects for them.

1605711683433.png

1605711691774.png


Replacing Public Works Division:
1605711706121.png


And for Void Dwellers with the Adaptability tree:
1605711724002.png


Interstellar Franchising and Imperious Architecture now also function for Habitats.

Updates to Dev Diary 190

Some of these updates may not be new to people following the forum threads, but it's easy to miss things so I figured we should go over them.

Many people requested the ability to fully specialize their foundry and factory worlds. We've modified the Forge and Industrial World planet designations to shift one pop on each Industrial District to the appropriate job if possible.

1605711738324.png
1605711745816.png


We've also upgraded the Food Processing Center, Mineral Purification Hub, and Energy Nexus to provide an extra job to each of their associated resource production districts. (The Food Processing Center will also improve Hydroponics Farms.)

1605711771358.png
1605711779670.png
1605711789149.png


One of the suggestions made in the thread was to add a civic that increases unlocked Building Slots. Sounded like a great addition to Functional Architecture.

1605711797879.png

Functionality increased!

Updates to Dev Diary 191

We’ve explored some additional options regarding the resettlement system we outlined in Dev Diary 191, and after trying a few things, and have settled on some extensive modifications to the system.

All planets with free sapient unemployed pops that are not locked down by migration controls will have a small chance every month of moving one to another planet within their empire that has jobs that they are willing and able to work, housing, and habitability of 40% or higher. This chance is increased if there are multiple unemployed pops that meet the criteria.

The system now prefers to move higher strata pops first, so rulers and specialists will move before workers, and this system also functions for gestalt empires. It will not relocate non-sapient robots or slaves. It will generally prefer to move pops to the planets with the most free jobs.

After some experimentation we’ve chosen to keep the Transit Hubs as Starbase Buildings that provide a system wide buff to the chance of auto-resettlement occurring. (Rather than being essential to have it occur in the first place.)

1605711834820.png

Doubles the chance the pops choose to resettle themselves.

Greater Than Ourselves has been rewritten to also massively increase this chance when the edict is active, with a +200% bonus.

We initially had these pops considering destinations available through Migration Pacts as well, but decided against keeping that since it introduced a new Migration Controls micromanagement element that we didn’t find desirable.

We’ve also done a minor update to the Authority bonuses that seemed a little bit weak.

1605711874350.png

1605711882524.png


Democracies now have a bonus encouraging their pops to seek their dreams, and Dictatorships have a bit of an easier time holding things together when they’re a bit overstretched.

Closing Thoughts

One other little quality of life improvement that was just added is this filter on the colonization interface.

Colonisation QoL.gif


That’s probably long enough for today. We’re looking forward to your feedback on these as well.

Next week w̷e̵'̸l̸l̴ ̴b̸e̴t̵̮̄ǎ̸͈l̷̠̈k̴͔͂i̴̞͒n̷̪͊g̸̳͗ ̸͚̎a̵͉̐b̵̤̿ȯ̴̲ṵ̵̀t̸͇͂ ҈҂▒©╛⅜

1605711927580.png
 
  • 209Like
  • 111Love
  • 24
  • 13
  • 7
  • 2Haha
Reactions:
I really would like to have a detailed explanation, why the game only using 2 cores maximum, in an era, when you can have 64 cores CPU at home?
OK? I can understand, the heavy CPU intensive function is to calculate what a POP is doing. Then why giving every 500 pop to an another CPU core? :), so then, every problem solved. No need to take away my puppets anymore! :)

This whole think is a paradox situation... :D, in 2010, if you used only 2 cores, it was OK, but now? Only company named paradox can do that! :p
First of all, as grommile already said, that is not the case. As far as I know it uses all cores, if it is feasable.
Secondly, as a softwaredeveloper, I can always facepalm reading something like this. Do you even have the faintest idea how multithreading works?

Just imagine you have Calculations A, B, C and D, which need to be done, then the results of A and B get calculated together. Same with C and D. After that the results of that are again calculated together by some formula.

You parallelize A, B, C and D probably. But the follow up calculations need to wait for the results... No matter the amount of cores. Maybe A and C are even dependent on each other which reduces that even more.
Just because you have 10 coworkers in the kitchen does not mean, they can all simultaneously work the coffee machine which magically makes the coffee brew faster.

Before making jokes about gamedevs who try to make a fun game learn coding yourself please.
 
  • 6Like
  • 4
Reactions:
The issue, of course, is that spiritualists aren't the best early-rushers (that would be militarists). They're not even the second best (authoritarian/slaver guilds) or the third (egalitarian/meritocracy). Spiritualism's weaknesses are best compensated for by early rushing, but that doesn't mean they're actually good at it.
Indeed, but they can trade weaker earlier game to set the stage to better exploit it's mid game at it full potential (inward perfectionist into divine empire for exemple). The problem is most of their strong points are yet to be correctly implemented if not at all (ex: faction, soft power...) or overshadowed by feature creep (memorialist civic) and balance issue (edicts, unity...). A reseach and unity overhaul with the introduction of actual politic is in order to correct those issue.


Reseach should not act like a rafined ressource like alloy to be spent. They should take inspiration from DW:U and combine them with their own mechanics ( ecavation's chapter progression, vic2's tech school, hoi 2's science team, sector, fleet manager, etc.)

The ascension perks should have the same dynamic that the ascension path that should be reduce to 1 perk, and not to be gatekeeped by technology but act more like the vic2's tech school. For example an empire can choose to focus it's economy developement into the void instead of the planetes or a middle ground of the 2 gaining space related bonus and unique tech option and upgrade while still accesing basic worlds's upgrade (like a spiritualist having access to robots and clones). Those option and upgrade can variate depending on the composition of the empire for example a psionically ascended materialist will focus on material application of the Shrood will spiritualist will be more abstracted and xeno(phile/phobe) will be hybride. In a rpg's sense the ascention will act like a modular class build for a hero.


The introduction of politic (similating an actor capacity to influence it environement) thus the remouval of the infamous purple mana can then allow finally the spiritualist (pacific and xenophile included) to be revelant and use it strenght at it was planned from the begining.


Tradition's tenets and unity should act as the federation and federal cohesion work at a national level.
 
  • 1
Reactions:
First of all, as grommile already said, that is not the case. As far as I know it uses all cores, if it is feasable.
Secondly, as a softwaredeveloper, I can always facepalm reading something like this. Do you even have the faintest idea how multithreading works?

Just imagine you have Calculations A, B, C and D, which need to be done, then the results of A and B get calculated together. Same with C and D. After that the results of that are again calculated together by some formula.

You parallelize A, B, C and D probably. But the follow up calculations need to wait for the results... No matter the amount of cores. Maybe A and C are even dependent on each other which reduces that even more.
Just because you have 10 coworkers in the kitchen does not mean, they can all simultaneously work the coffee machine which magically makes the coffee brew faster.

Before making jokes about gamedevs who try to make a fun game learn coding yourself please.

Are you saying that I employed 5 coffee-maker interns and my coffee isn't even coming faster? I wish one said that to me earlier
 
  • 9Haha
  • 1
Reactions:
Are you saying that I employed 5 coffee-maker interns and my coffee isn't even coming faster? I wish one said that to me earlier
Real shame, isn't it?
And here I am making my own coffee since I am the first in the whole company every morning :(
Intern made coffee is too weak anyway :D
 
  • 1Haha
Reactions:
Real shame, isn't it?
And here I am making my own coffee since I am the first in the whole company every morning :(
Intern made coffee is too weak anyway :D

That's a rough life you've here, my dude !
I dream of a world where coffee would be made when you arrive in the morning
 
  • 1Like
Reactions:
First of all, as grommile already said, that is not the case. As far as I know it uses all cores, if it is feasable.
Secondly, as a softwaredeveloper, I can always facepalm reading something like this. Do you even have the faintest idea how multithreading works?
I wish the hype about multithreading/parallel didn't make people think it's a silver bullet for computation heavy tasks.
There is no amount of CPU cores that will make up for the combination of O(n^2) and gamers insistence on pushing everything to the limit.
 
  • 5Like
  • 2
Reactions:
Guys, thumbs up for using logistic curve. I'm really happy with direction you are going as it looks like most of problems troubling Stellaris right now are really being worked on.
 
  • 3Like
Reactions:
I like the adoption of Carrying Capacity (its a great mod, and having a somewhat more 'organic' pop growth for organic empires is great).
Mixed feeling about the slower growth the larger your empire's population. It does allow 'smaller' empires a change to catch up, and encourage vassalization, so yay? But building habitats and getting more pops that way ('tall') will still cause all your planets to suffer the slower growth.

Concern about the Clone Vats.
The faction/ethic/playstyle which is most hostile to robots, and thus would benefit the most from this building, is spiritualists - who are nudged towards psionic ascension (and psionic ascension also doesn't work the greatest with robots either, Assimov's psychic robots be darned). But Clone Vats are a biological ascension path unique building.

Maybe the clone vats could be a tech unlocked building and bio ascension gets a different benefit from its tier 1 unlock? (A bonus to habitability could be flavorful, tailing your biology to a planet rather than reverse, and allow Bio Ascension 100% habitability on all planets with Robust without relying upon the Worm's Blessing or Lithoid).

Habitat Changes
Some good things! Tradition Swaps for Void Dwellers is good. Habitats scaling with some of the traditions they're meant too is also good. But habitatation districts need the bonus to housing from the technologies and traditions if they're meant to scale in comparison to planets still. 1.5x efficiency of districts won't hold if at the mid to end game the city districts are better than the habitat districts (same housing, but provide more clerk jobs).

Speaking of 1.5x efficiency, the Habitat manufactory district being only 2 jobs is bad. Considering the habitats have an alloy upkeep, making it 2 metallurgists + 1 artisan, and swap to 3/3 depending on designation, should work well.

Other Things
Medical workers providing only 1% habitability isn't going to do much. Its barely worth considering and certainly won't make them worthwhile to use from a min/maxing perspective (and changes to relative growth values won't do too much for that either, and arguably makes them worse as pop growth tapers off faster with more growth speed). Especially with only 4 medical workers with the t2 hospitals, that's 4% habitability. Not going to do much to help. 2.5% at least, but 5% habitability will probably be the sweet spot, making it feel impactful to build them where needed (effectively giving them a 7.5% pop growth per worker).

Democracy Authority bonus went from nearly useless to niche, but I guess it also makes sense with egalitarian factions demanding banning of resettlement. But at the same time could create issues if you're not wanting a pop to migrate from a particular planet (such as if you're upgrading a science lab on your capital). Maybe a bonus to leader pool could work instead? (The idea of a more meritocratic system giving more potential for leaders. Although that could work as an alternate/addendum to Meritocracy to try and make it less potent). The rewards for elections is probably a greater issue with democracies though (and space miner/researcher agendas past year 100 xD) than a weak authority bonus however.
 
  • 2Like
  • 1
Reactions:
We’ve also changed a few buildings to have new or additional features, such as the Spawning Pool and Clone Vats, which have had their Pop Growth modifiers replaced with the new Organic Pop Assembly. This fills the same slot on the planet as Robotic Pop Assembly, so generally you’ll want to pick one or the other. (Clone Vats also picked up a food upkeep cost to represent simple materials to break down.)

View attachment 653806View attachment 653808
Pops is Soylent Green!

Please add in a "clone lords" or "gene masters" civic/origin, like with Mechanist, that lets an empire start with cloning tech but have something like an atrophied natural growth rate. Would be cool if combined with Syncretic Evolution to have a civilizaiton of mad scientist (Auth+Materialist) growing their own pet labor force.
 
  • 1Like
Reactions:
Ringworlds, Ecumonopolis, Gaia Worlds. All of these can be found in the wild without needing to have the technology to make them yourself. Having an origin that starts with 1 makes sense, but should come with downsides. For Gaia worlds it's life seeded, for Ecumonopolis it's that it's a relic world and you have to restore it. With this update, ring world origin will finally have a downside.


The point to Gaia worlds like many others is that they are a rare find. It should be a reward (or bait) when you find/conquer one. I don't know if the 100% hab & happiness bonus is enough to make a player excited at the find. Aside from that, you'd be asking a player to make an investment of an entire ascension perk for 10% production on a single planet. I don't think this will make players take this perk over arcology project.


Well invasion should give you all the benifits though. If you invade a planet you conquered you get their pops too and all the benifits of Synthetic/Biological ascension because you have their modified population. Planets are no different in that regard, conquer an ecumenopolis and you get all the gain without the pain.


There should be a benefit to the perk that is inaccessible to those without it and at least 60% of the perk's benefit really should be locked behind the perk. In regards to bio ascended modified pops, it really shouldn't be an option to mod subspecies into a bio ascended subspecies. Last I checked, this was possible, you couldn't modify the template, but you could. Also pretty sure bio ascension came with a pop growth boost. I know synthetic does get a pop assembly boost.

Invading doesn't mean the invader magically figures out how that stuff works. It just means they have access to what's comprehensible with their little minds. Honestly, the game rewards invasion too much and that does devalue things. Yes, moving about 60% of the benefit for each perk behind having to have the perk won't make every perk worth taking, some need buffs, mastery of nature easily comes to mind here. World Shaper isn't another one, but IMO, I think if they did that, they'd have an easier time with ascension perk balance because suddenly instead of a bulk of it being available in the wild, as an origin or plunder of conquest, but not is locked to the perk. You don't have be like "well we'd love to X, but that would mean Y origin would be too strong, Z system would be too good, N event would be too desired and no one would want to diplomacy because you might as well take it. It's behind the perk, so if the perk is too good, you can adjust only the perk. Not saying invasion shouldn't have up sides, just that it shouldn't viable as a means to access most of the benefit behind a perk; especially, when some perk require tons of investment (no, being able to build stuff where you want isn't even worth 5% of the perk).

Though IMO, I strongly suspect ringworld and void dweller origins without ethics are less strong than they currently get made out to be. They are stronger origins, but what really breaks things is that some ethics are probably too good. Usually, when I see setups where someone is proving how broken something is it's like a 90% chance you'll see either materialist or xenophobe picked. I know we all sit here and say spiritualist and psionic ascension need some buffs, but the xenophobe and materialist ethics probably need some nerfing, which I know isn't going to be a popular thing to say. It's just when you have to ethics that provide a boost to the top two resources, pops and research, well it shouldn't be a surprise when those end up being rather powerful before you even begin to touch on the other things going for them.
 
  • 1
  • 1Like
Reactions:
Would some level of abstraction help alleviate this problem?
The easiest way is to express it in terms of a speedup effect.

But I can promise you that having game systems that depending on sorting or allocation functions that grow in polynomial time will render that moot eventually.

We don't know the specific algorithms they use under the hood, but it's usually much much more effective to either reduce the time complexity of what you're doing (going from sorting all the time, to say, a linear lookup process) or reduce the amount of stuff you're doing it to (in our case, cutting down how many pops we have.)
Only after that do you need to look at implementation specific stuff - for example, in Python, looking up values in a list vs in a dictionary.
There should be a benefit to the perk that is inaccessible to those without it and at least 60% of the perk's benefit really should be locked behind the perk.
Invading doesn't mean the invader magically figures out how that stuff works.
I have made a few civic mods and I encounter this issue. You don't want people to use a civic to get the immediate benefits, then dump it for something else. That would be more like an edict than a civic.
I've done a few different things, to use my megacorp civics as examples: Peak Production is a civic which allows you to use a planetary decision to reduce the habitability and food output of a planet in exchange for boosting industrial output. One thing i did here was have the +% bonus only apply if you have the civic - otherwise you just get a token pair of jobs and still have the pollution effect.

Another is a civic that adds a trade arcology to ecumenopoleis. This would be a prime civic to "pump and dump" once you've built up. So the trade district actually converts back into leisure arcologies if you don't have the civic (and vice versa!)

It's an interesting problem.
 
  • 2Like
Reactions:
There should be a benefit to the perk that is inaccessible to those without it and at least 60% of the perk's benefit really should be locked behind the perk. In regards to bio ascended modified pops, it really shouldn't be an option to mod subspecies into a bio ascended subspecies. Last I checked, this was possible, you couldn't modify the template, but you could. Also pretty sure bio ascension came with a pop growth boost. I know synthetic does get a pop assembly boost.

Invading doesn't mean the invader magically figures out how that stuff works. It just means they have access to what's comprehensible with their little minds. Honestly, the game rewards invasion too much and that does devalue things. Yes, moving about 60% of the benefit for each perk behind having to have the perk won't make every perk worth taking, some need buffs, mastery of nature easily comes to mind here. World Shaper isn't another one, but IMO, I think if they did that, they'd have an easier time with ascension perk balance because suddenly instead of a bulk of it being available in the wild, as an origin or plunder of conquest, but not is locked to the perk. You don't have be like "well we'd love to X, but that would mean Y origin would be too strong, Z system would be too good, N event would be too desired and no one would want to diplomacy because you might as well take it. It's behind the perk, so if the perk is too good, you can adjust only the perk. Not saying invasion shouldn't have up sides, just that it shouldn't viable as a means to access most of the benefit behind a perk; especially, when some perk require tons of investment (no, being able to build stuff where you want isn't even worth 5% of the perk).

Though IMO, I strongly suspect ringworld and void dweller origins without ethics are less strong than they currently get made out to be. They are stronger origins, but what really breaks things is that some ethics are probably too good. Usually, when I see setups where someone is proving how broken something is it's like a 90% chance you'll see either materialist or xenophobe picked. I know we all sit here and say spiritualist and psionic ascension need some buffs, but the xenophobe and materialist ethics probably need some nerfing, which I know isn't going to be a popular thing to say. It's just when you have to ethics that provide a boost to the top two resources, pops and research, well it shouldn't be a surprise when those end up being rather powerful before you even begin to touch on the other things going for them.
IMO, the strongest ethics are actually authoritarian and militarist; authoritarian gives the biggest eco-boost (What's +10% research speed when you can have +40% researcher output from slaver guilds + extended shifts + indentured servitude? What's 20% robot upkeep when you can make 40-90% of your pops have no CG upkeep with stratified economy? The influence is just gravy.) and militarist is much stronger than xenophobe as a pop growth and expansion ethic (why grow your own when you take other people's?).
 
There should be a benefit to the perk that is inaccessible to those without it and at least 60% of the perk's benefit really should be locked behind the perk.
Oh but there is. The Arcane Generator is unique to shattered ring. And the Decrepit Tunnels blockers are more of a boon then a bane. Aside from those you
  • Don't need to have the late game Mega-Engineering technology
  • Don't need to have the late game Galactic Wonders ascension perk
  • Don't have to do any of the work required for the 7 other ways to get a ringworld naturally
In regards to bio ascended modified pops, it really shouldn't be an option to mod subspecies into a bio ascended subspecies. Last I checked, this was possible, you couldn't modify the template, but you could.
Why shouldn't you? The bio ascension is more about the technology then the people themselves. Society has advanced far enough down the biological tech tree so they can now gene-mod a perfect species. There's no thematic nor gameplay reason to not allow you to mod immigrants, slaves, uplifted or servile species.

Also pretty sure bio ascension came with a pop growth boost. I know synthetic does get a pop assembly boost.
Both Engineered Evolution and Evolutionary Mastery nor the technology they unlock come with a growth boost. The only growth boost you get is from the Cloning vats which you have to build on each planet individually. Maybe you're thinking about Xeno-Compatibility, but that's only possible if you are some form of xenophile which excludes about 7/8 of the other empires.

But you know, when in doubt you could always look it up

Invading doesn't mean the invader magically figures out how that stuff works. It just means they have access to what's comprehensible with their little minds.
Of course they don't magically know how the tech works. But the genetically or synthetically modified pops also don't suddenly devolve to their previous state, which is what I was getting at before.

...mastery of nature easily comes to mind here.
So a species masters nature. They do it in such a way that they optimize their planet by managing biomes and building in smart orginized ways leaving more room on their planet. And then all this infrastructure and planning then suddenly disappears once someone conquers the planet? That logic is flawed in so many ways. Mastery of nature is like terraforming, a permanent change to the planet. It should not disappear if you do not have the ascension perk.

In closer inspection you will see that only those that have the perk can take the decision to transform the planet. If a planet that hasn't been transformed yet gets conquered, the conqueror doesn't get to keep that decision. On top of that, in a situation where both the conquered and the conqueror have Mastery of Nature, an already transformed planet can't be transformed twice (or multiple times). In that sense the Mastery of Nature function is already properly locked behind the perk itself.

You unlock the decision to do stuff, not the changes that result from it. This is true for all ascension perks, be it changes to pops in the synth/bio/psionic ascencions, in the world trough world shaper, mastery of nature and hive/machine worlds OR trough mega structures. Their decisions stay with the ascended, but their results are free for the conquering.

World Shaper isn't another one, but IMO, I think if they did that, they'd have an easier time with ascension perk balance because suddenly instead of a bulk of it being available in the wild, as an origin or plunder of conquest, but not is locked to the perk.
The design you want is counter productive. All these juicy things being there for the taking invites ire and conflict. These are the building blocks or actual reasons to go to war with another species. "I want your pops and I want your ring world, so I declare war"

You don't have be like "well we'd love to X, but that would mean Y origin would be too strong, Z system would be too good, N event would be too desired and no one would want to diplomacy because you might as well take it. It's behind the perk, so if the perk is too good, you can adjust only the perk.
All that does is shift where you make changes to adjust balance.

Though IMO, I strongly suspect ringworld and void dweller origins without ethics are less strong than they currently get made out to be. They are stronger origins, but what really breaks things is that some ethics are probably too good. Usually, when I see setups where someone is proving how broken something is it's like a 90% chance you'll see either materialist or xenophobe picked. I know we all sit here and say spiritualist and psionic ascension need some buffs, but the xenophobe and materialist ethics probably need some nerfing, which I know isn't going to be a popular thing to say. It's just when you have to ethics that provide a boost to the top two resources, pops and research, well it shouldn't be a surprise when those end up being rather powerful before you even begin to touch on the other things going for them.
This entire part is just semantics really.
  1. To prove which origins are stronger or weaker you could build exactly the same empire on all of them. In those tests you'll see specific empires be strong with a couple of empires. But irregardless of the outliers, over the broad spectrum Scion, Shattered Ring and Void dwellers will be stronger and lost colony, life seeded and post apocalyptic will be weaker. Purely based on what they do.
  2. In order to get balance you can either buff the weaker [Insert ethic, civic, origin, ascension, mechanic], or you can nerf the stronger [Insert ethic, civic, origin, ascension, mechanic]. In the end you probably want to do both to those that pop out of the scale to bring everything within a spectrum you are happy with.
 
...
The easiest way is to express it in terms of a speedup effect.

But I can promise you that having game systems that depending on sorting or allocation functions that grow in polynomial time will render that moot eventually.
...
Anyways, other than traits, we could ditch individual pops entirely. I think that you could compromise though by removing anything that affects an individual pop that isn't on a planetary level - like those event chains where one specific pop becomes unhappy or something - and putting all that crap into traits, so pops are purely differentiated at the species level.

Then our planets can store pops as a table of "i have N pops of species X" and the precomputed data table for each subspecies is centrally stored and accessed for stuff like job allocation. This would cut down the job computations from being a function of # of pops to # of subspecies on a planet, which is vastly lower.
I think this pop abstraction idea has merrit. Do you still stand by it?
 
I think this pop abstraction idea has merrit. Do you still stand by it?

I like the idea of getting rid of individual pops (or at least how they are now). But it would be a lot of work to retool pops into pop size and redo jobs. Especially if you wanted to keep class distinctions. In a new framework, seems like you'd have districts getting worked, and each unit of pop size would work one district. Buildings would have static effects on district productivity. Figuring out classes would be hard though. This would simplify things a lot and make the AI for handling pops much better.
 
  • 1Like
Reactions:
I am concerned about the first 100 years of the game feeling even slower though. Granted, I've only played machine empires since I came back to the game, so maybe it is different for other races. Right now pop growth seems very constraining.
 
I think this pop abstraction idea has merrit. Do you still stand by it?
Yes. Within their framework it is the clear option that doesn't change anything the player observes but does greatly simplify the process of computing that.
For example, job sorting used to run on every planet every day. Then they made some performance change sin 2.7 i think which cut that frequency to like, once a month ish.

Now we have two simple attack vectors here - we can further reduce the frequency, or we can reduce the complexity of the allocation.
Right now it has to check all pops individual because they might have distinctions. If we remove those edge cases, though, and there are like only a few in the whole game, then it's no longer individual pops we have to compute on, but the buckets of subspecies because they are all identical within that.

One thing that makes job computations slow is the need to compute all the job modifiers for each pop every time because they are stored at the job level - so currently you have no way of knowing that the pop you evaluated as the best fit for a miner job is still the best fit a month later. By moving all job modifiers & weights to traits (they almost all are afaik) that's when you can introduce those centralized data tables. If all subspecies have uniform job weights within them, and that only changes when I change the subspecies or their living standard, then i can just compute that array of weights once and only update it when I change either of those 2 things. This works because JOB DEFINITIONS DO NOT CHANGE ONCE THE GAME IS LOADED.

The other attack vector, reducing frequency, can slingshot on this. If job weights only change when traits or living standards change... then the job allocation will only change when 1) the number of pops on a planet changes or 2) the number of jobs changes. This means that you can flag the job allocation algorithm to only run when either of those things changes. It takes 15 - 30 months to grow a pop and about a year to build a building or district, and even accounting for the new unemployed pop migration, you can take your (now much simpler) job allocation function and employ it a fraction of the time you were before.

And all we need to do to achieve this is just remove a few edge cases.
I like the idea of getting rid of individual pops (or at least how they are now). But it would be a lot of work to retool pops into pop size and redo jobs. Especially if you wanted to keep class distinctions. In a new framework, seems like you'd have districts getting worked, and each unit of pop size would work one district. Buildings would have static effects on district productivity. Figuring out classes would be hard though. This would simplify things a lot and make the AI for handling pops much better.
Nothing you observe would change, only the way it's handled under the hood. Class distinction are not any harder to track than job distinctions. At the end of the day you are basically just taking an array of tuples of (species_id , num_of_pops) and tallying up who goes where by referencing a table of "this species_id has a weight of X for this job_id.". Hopefully I have gotten across why this is a LOT faster than taking a list of every single individual pop on the planet and then for each job on the planet computing their weighting.
 
  • 1Like
Reactions:
After reading through the diary entry twice and thinking about it, I can tell you with confidence: now resettlement is going to become even more painful.

It feels like the devs are completely misunderstanding the point of manual resettlement, and what's the optimal way to do it. Which is hard to believe, so there is probably some utter motive that I'm missing.
You resettle primarily to:
  1. Get to 5 pops ASAP (ideally, literally the same day the colony is established, ignoring habitability, employment, housing, stability etc) and drop the Robot factory. Nothing else (with a single exception of the egg event) matters about the world in question, because this is how you snowball in Stellaris. The world can have all non-world-exploding negative modifiers that exist in the game, and people will still do it, so long the world is physically capable of starting producing robots. 10 influence per pop? Whatever. I'll just include these 30 influence into the cost of grabbing the system.
  2. Getting to 10 pops just to press the button on the colony building (resetting them back afterwards is entirely acceptable; the goal here is to remove the growth penalty). Okay, now this is an impactful extra tax. Which means doing the first thing (with robots) becomes even more important. Because now your hopes to start growing normal pops on fresh worlds any time soon are dead, and spamming robot factories becomes even more critical. Also that means that if you are not getting the robot tech, you are even more screwed.
  3. After you are completely done with this phase, you can chill down and relax, because if you are still alive, you probably already won, and fixing unemployments at this point is a "win more" issue. You will outgrow all AIs by so much in a couple of years that I honestly stop bothering with resettlement at this point.
And now you are telling me there is a randomized system that will undo my efforts and force me to do all operations in one go, or else there is a risk that the month rolls over and someone leaves the planet, and if I miscalculate, let's say, minerals by a little bit, there is a chance the game will force me to pay for resettlement again.

And by the time it starts doing anything actually useful, it doesn't matter anymore.

...

This feels like a... worrisome change, to put it lightly. Also with how resettlement is often the only chance to fix the non-working auto-job-assignment algorithm (that will casually fail to fill job slots with pops in any non-trivial scenario), I'm very worried that you are not doing anything to the strongest non-bug-exploiting strategy in the game (pop explosion), but simply making it even more painful than it already is.

I would be happy if the move was in a completely opposite direction: make manual resettlement cheap, but unnecessary (as opposed to expensive and critically-important, which is exactly what this proposed change does).


Also, my understanding is that having slaves in the empire is a must now (because it completely dodges the prohibitive Influence cost). Which is fascinating, because slaves mean even more micro.

__
Btw, if you are curious: yes, even on the latest patch the auto-job-distribution doesn't work. At all. Please consider hiring a tester or something. Example setup: 2 bureaucratic jobs, 1 portal job, 1 mechanist (it's a typical 5%hab world that exists only to spit out pops and do the paperwork, because you need to put these buildings somewhere).
I have 3 meat pops and a bunch of droids who CAN work the mechanist job, and doing it on all other planets. With this setup, it's literally impossible to force the game to fill all 4 slots. One of the meat pops will always kick the droid from the mechanist job, leaving another specialist slot empty (even if none of the non-specialist jobs are prioritized, and even if you manually close all other jobs on the planet) If you focus bureaucracy, it will be the portal guy. If you focus portals, it will be the bureaucrat guy. Focusing assembly does nothing. And bugs like this happen all the time, and migration back-and-forth several times is often the only way to fix the issue, if it's possible at all; but now it will cost us a lot of influence to even try.
 
Last edited:
  • 3
  • 3
Reactions: