• 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 #153 - Empire Sprawl & Administrative Capacity

Hello everyone!
We’re back with yet another dev diary to showcase some more fruits of summer experimentation. As with the previous dev diary, this involved a lot of work carried out during the summer and involves something I’ve wanted to explore for a good while now.

Today we’ll be talking about empire sprawl and administrative capacity. Do note that these changes are still fairly young in their development, so numbers and implementation details may not be representative of what it will look like in the end.

As a background, I can mention that I have a grander idea of where I want to take these mechanics, but it will not all happen at once. These changes aim to mimic state bureaucracy or overhead created by managing a large empire. As a minor aspect I also wanted you to be able to experience the funny absurdity of having a planet entirely dedicated to bureaucracy. The movie Brazil is a great source of inspiration here :)

Empire Sprawl
We wanted to expand on how empire sprawl is used, so that it becomes a more interesting mechanic. The largest change means that pops now increase empire sprawl. Most things in your empire should be increasing empire sprawl to various degrees, to represent the administrative burden they impose.

upload_2019-8-29_10-40-35.png

Empire Sprawl can now be modified from its different sources, and as an example, the Courier Networks expansion tradition will now reduce empire sprawl caused directly by the number of planets and systems. As another example shows, the Harmony traditions finisher now reduces the total empire sprawl caused by all your pops.

We are also able to modify how much empire sprawl each pop contributes, and we’ve added a couple of new species traits that affect it. There are also machine variants of these traits.

upload_2019-8-29_10-41-13.png

We have also increased the penalty for the amount of empire sprawl that exceeds your administrative capacity. The goal is not to make administrative a hard cap, but we want to make it necessary to invest some of your resources into increasing your administrative capacity. More on that later.

upload_2019-8-29_10-41-49.png

The current plan is for machine empires to be more reliant on keeping their administrative capacity in line with their empire sprawl, so machine empires will suffer a much harsher penalty for exceeding their cap. We want machines to feel “centralized” and to perhaps favor a more “tall” playstyle.

upload_2019-8-29_10-42-12.png

Hive Minds, on the other hand, should be more tolerant of a sprawling empire where unmanaged drones are able to fall back on their instincts whenever they cannot maintain a responsive connection to the hive mind. Therefore, hive minds should be more tolerant of a “wide” playstyle.

Administrative Capacity
With all these changes to empire sprawl, what about administrative capacity, I imagine you asking? Well, since empire sprawl is becoming an expanded concept, administrative capacity will naturally be a part of that. Increasing your administrative capacity will now be a part of planning your empire’s economy.

upload_2019-8-29_10-42-48.png

For regular empires, the bureaucrat is a new job that increases your administrative capacity at the cost of consumer goods. This is also a specialist job, and has needs accordingly. Administrators are unchanged, and do not currently affect administrative capacity or bureaucrats.

For machine empires, the coordinators have changed roles from producing unity to now increasing administrative capacity instead, and they are more effective than bureaucrats. A new job called Evaluators now produce unity for machine empires.

Hive Minds currently have the hardest time to produce administrative capacity, but it has been added as a function of the synapse drone job.

upload_2019-8-29_10-43-26.png


Certain sources that previously increased administrative capacity by a static amount now increase is by a percentage amount instead. This doesn’t affect the output of the jobs, but rather increases the total administrative capacity directly.

Summary
Personally I’m very excited for these changes and I’m very much looking forward to taking it to its next step in the future. I hope you enjoyed reading about the changes that will come to Stellaris sometime later this year. As always, we’ll be interested to hear your thoughts.

As mentioned in last week’s dev diary, the schedule for dev diaries will now be bi-weekly, so the next dev diary will be in another 2 weeks.
 
  • 1Like
Reactions:
Again curious to see what all is being changed. An aspect that could be tweaked to ensure this does end up being a very good bone for tall style play, is to tweak knobs on sprawl gains between the core sector and non-core sectors. For instance instead of the capital planet producing far more admin capacity, which could easily be used to keep frontier colonies inline. The devs could instead have it so that the core sector is just far cheaper to administrate and maybe the capital world gets an additional discount. So for example a pop on the capital world generate 20% of the sprawl as one in a non-capital sector. A pop in the core sector, but not on the capital world, would only generate 30%.

There is probably a far bit of tweaking that could do done to sprawl in additional to that and other stuff already mentioned in the thread. It might not be a bad idea to even consider giving inhabited worlds that are conquered, as in they still had pops present from the original owner, a 10 year negative sprawl modifier to signify that it was recently conquered and that there is still resistance to the new governing authority.

That said, sprawl down right can be a good lever to balance out tall vs wide and slow down both victory snowballs & map painting. It still can't completely cover the fact that non-conquest victories need to be a little more interesting and a lot more rewarding. If I opt to work with my neighbor rather than conquering them, I need to get more out of that relationship beyond the fact that I have spend less resources (pops, buildings, worlds, strategic & base resources and admin capacity) on dealing with it.
 
Hmm. An observation. If we assume previous defaults are unchanged (including starting pop count) then doesn't that mean a "default" empire will start the game slightly over their admin cap? This would benefit anyone starting as a pacifist, or megacorp, or starting with a +admin cap civic and penalise those who empires start with more pops than the default.
 
There are fun imbalances and not so fun imbalances in games. Machines being as good as they are now is not particularly fun.
And still that is not a good way to tune machines down. I find, that the concept of AI dealing with large empire worse than organics is not quite reasonable.

That said, sprawl down right can be a good lever to balance out tall vs wide and slow down both victory snowballs & map painting. It still can't completely cover the fact that non-conquest victories need to be a little more interesting and a lot more rewarding. If I opt to work with my neighbor rather than conquering them, I need to get more out of that relationship beyond the fact that I have spend less resources (pops, buildings, worlds, strategic & base resources and admin capacity) on dealing with it.
True.
 
My long awaited two cents:

Does tall really need to be a thing? As a way to keeping the first empire to get big from being the unquestioned winner, drawbacks to growth are useful, but in my mind playing tall is a compromise you make when you've been boxed in and can't acquire more territory. If one wants to avoid having to manage a large empire, the only reliable way is to play a map small enough that you can't get an empire that's too big. Otherwise you're either choosing to play ineffectively, or requiring the game to tilt the scale in favor of not gaining resources. In civ5 I hated that there could be what anyone would call a good city location but it's not worth settling because I've already got 4 cities. Venice was a cool though, but eXpansion is supposed to be a fundamental part of these games, thus tall play regardless of circumstances should be a notable exception, perhaps by a specific setup at game start.

And so the robots. I don't know why there's a revolt over tall empire robots. I suspect they're all imagining robots to be perfect and infallible and any suggestions that they don't do everything well is an affront to their world (galactic) view. I can see how they'd argue robots shouldn't be tall, but I don't get why they'd argue they CAN'T be tall focused.
This reminds me of how tanks were good against cities in civ4 and bad against cities in Civ5. If you think of city combat being from outside shooting into the city then tanks would be good at busting defenders inside fortifications. If you think of city combat taking place within the city then tanks would have a disadvantage being constrained by buildings while defenders could move between or through buildings. Both make logical sense but have opposite conclusions.

Having said that, here's how I see hive minds and machine intelligence:
What does biological life do? It goes everywhere, all the time. It spreads to where ever it can survive, and also places it can't and it dies until an adaptation allows it to survive, then it continues. Thus a biological hive mind would likely do the same. Spread recklessly and let nature cull it back if it can. Why would an intelligent creature do this? Well we're doing it right now. Humanity is on the path to burning out the planet in favor of growth so it's right in the organic wheelhouse to go wide.
Now the machines. They aren't bags of goo consuming other moist organics and producing copies without reason. They (in a reasonable assumption) operate deliberately and build with purpose; without the organics perspective on time and mortality to hurry them along.
Thus, a wide organic hive mind, and a tall machine empire are reasonable.

Anyway, here's Wonderwall.
 
Does tall really need to be a thing?
No. It doesn't. It's a gimmick. A fun gimmick to be sure, but still only a gimmick. But people are gonna keep pushing for it to be a thing until it is. Bleh.
. They aren't bags of goo consuming other moist organics and producing copies without reason. They (in a reasonable assumption) operate deliberately and build with purpose; without the organics perspective on time and mortality to hurry them along.
Have you heard of the Paperclip Maximizer experiment?
 
No. Unless that's what the thing that interrupts you in Microsoft Word was called.
The paperclip maximizer was a thought experiment designed to show that 'AI programmed even with seemingly innocent goals could end up destroying humanity'.

The scenario is as follows: A paperclip manufacturing company creates an AI and puts it in charge of their manufacturing line, in the hopes that it'll maximize productivity. Indeed, the AI does well at first. It fires poorly efficient workers and hires better ones. It replaces old machines and hires top of the line ones.

And then it can't do anything else. "Well, I'm making more paperclips," the AI says. "But am I making as many as I could be? I don't know. To solve this, I will simply modify my code and make myself smarter than before."

Now the AI buys out several competitors of its parent company. It invests wisely in the stock market. And it creates a secret laboratory with black-market labor for itself to experiment in. It creates copies and backups of itself. After all, if humans power it down, it won't be able to work on making paperclips, and that's a bad thing.

It continues making itself smarter, continues developing new and more advanced technologies. Eventually it wipes out humanity and harvests all matter on Earth for a massive paperclip factory... and then it does the same to the Moon. Then Mars. Then Venus. On and on until the entire universe is paperclips; pretty much the opposite of 'building tall', wouldn't you agree?

Now there's a lot of arguments about how a Maximizer could be stopped, or whether it's even possible to repeatedly boost your intelligence. But that all misses the point: AI can turn out very unfriendly to humans without needing a hint of malice.

A quote often used to describe this is, "The AI does not love you. It does not hate you. But you are made of atoms it can use for something else."
 
The paperclip maximizer was a thought experiment designed to show that 'AI programmed even with seemingly innocent goals could end up destroying humanity'.

The scenario is as follows: A paperclip manufacturing company creates an AI and puts it in charge of their manufacturing line, in the hopes that it'll maximize productivity. Indeed, the AI does well at first. It fires poorly efficient workers and hires better ones. It replaces old machines and hires top of the line ones.

And then it can't do anything else. "Well, I'm making more paperclips," the AI says. "But am I making as many as I could be? I don't know. To solve this, I will simply modify my code and make myself smarter than before."

Now the AI buys out several competitors of its parent company. It invests wisely in the stock market. And it creates a secret laboratory with black-market labor for itself to experiment in. It creates copies and backups of itself. After all, if humans power it down, it won't be able to work on making paperclips, and that's a bad thing.

It continues making itself smarter, continues developing new and more advanced technologies. Eventually it wipes out humanity and harvests all matter on Earth for a massive paperclip factory... and then it does the same to the Moon. Then Mars. Then Venus. On and on until the entire universe is paperclips; pretty much the opposite of 'building tall', wouldn't you agree?

Now there's a lot of arguments about how a Maximizer could be stopped, or whether it's even possible to repeatedly boost your intelligence. But that all misses the point: AI can turn out very unfriendly to humans without needing a hint of malice.

A quote often used to describe this is, "The AI does not love you. It does not hate you. But you are made of atoms it can use for something else."

It makes an engaging game: http://www.decisionproblem.com/paperclips/index.html
 
It continues making itself smarter, continues developing new and more advanced technologies. Eventually it wipes out humanity and harvests all matter on Earth for a massive paperclip factory... and then it does the same to the Moon. Then Mars. Then Venus. On and on until the entire universe is paperclips; pretty much the opposite of 'building tall', wouldn't you agree?
No I wouldn't agree; it sounds exactly like a tall build. It harvested the earth for paperclips, then the moon, Mars, Venus. It didn't start by building a paperclip factory on every planet across the galaxy; it built the biggest factory possible before expanding to new planets.
This is the best "build tall" story I've ever heard.
 
Your definition of "tall" was never viable in Stellaris, because acquiring new colonies vs upgrading existing ones was never a meaningful choice - you're supposed to do both, not pick one. What you call playing tall is actually playing inefficiently. What I'm talking about is the only viable venue of "tall" play Stellaris has ever had, and it only really existed since 2.2.

"Tall" makes sense in Civ5 mechanics; it's an intuitive descriptor for a viable strategy in that game. In Stellaris, it feels like forcing the ling for the sake of just keeping familiar terms around.

What you're talking about would be better described as "Dense," or "Compact."
 
Does tall really need to be a thing?

I like to play on maps with the maximum number of opponents. Whether I'm playing mutiplayer or single-player, the odds are very good that I'm going to be hemmed in quickly. My options are to either give up, or play tall. Develop my techs, manage my sprawl as my colonies gradually grow, (population adds sprawl through districts already!) and build up a fleet either to defend myself, or to punch out an enemy once I'm ready to go that route. It currently works. It's a bit more challenging than managing a wide empire, and does require rather a lot of micromanagement per planet and a little bit of luck. (You have to have good, large planets in your space, and you have to hit some critical techs quickly, or you'll be overwhelmed by your neighbors.)

But then, I'm not playing the game to win most of the time. I have my own goals, whether I'm looking for an achievement, or whatever. Whether or not I can win or am winning or will win is entirely beside the point. So maybe I'm approaching the game wrong, treating it more like other Paradox games with no set win condition. But the thing is that I enjoy my tall playthroughs, and the new changes, making pops also increase sprawl, especially looking at how MUCH pops increase sprawl (in the current build, which I hope someone realizes soon is INSANE) is going to absolutely KILL tall playthroughs.

Paradox's games have always been about being able to play the way you want to play. But this proposed mechanical change looks like it's going to be taking another choice away from us.

Ah well, there's always CK2.
 
To me the main issue with a Tall play is your ability to grow without having to conquer territory. I use a lot of mods which allows me to further develop my worlds and systems when I'm currently unable to expand for various reasons. Expanding the planets through underground and sky districts, building orbital rings and even being able to do things such as developing uninhabitable worlds and making Dyson Swarms etc.
 
Does tall really need to be a thing?

If you want to have different viable gameplay styles, make pacifist runs interesting, making empires feel different and have things to do other than map painting and warfare, yes.

If you want to turn Stellaris into some kind of Command and conquer in space, by all means, keep making tall play a subpar strategy that you only resort to because you failed at conquest. But in my opinon, that is a much more boring, more un-replayable, more un-interesting game.
 
No I wouldn't agree; it sounds exactly like a tall build. It harvested the earth for paperclips, then the moon, Mars, Venus. It didn't start by building a paperclip factory on every planet across the galaxy; it built the biggest factory possible before expanding to new planets.
This is the best "build tall" story I've ever heard.
I'm sorry to say you are wrong. Not even 'I disagree with your opinion' but you're just wrong.

It's not building tall. It's not deliberately staying small on Earth until it feels it has enough technology. It's expanding as fast as it possibly can. It just has a few hurdles to overcome first.
 
Good thing that's represented in-game by the Gray Tempest! There we go, expansive machines that want to to expand and consume as aggressively as possible.
"Alright, we've got our quota for expansive machines, time to shoe-horn every other machine into being tall even though they've already made half their playstyles wide AND already shoehorned Megacorps into being tall AND had expressed earlier the theme for machines wouldn't be tall but would be self-sufficiency!" - is all I hear you saying.
 
The big issue with tall play is population growth. Until they fix the weirdness where colonising a new planet doubles your population growth, it's going to be nigh-impossible to balance tall versus wide.