Hey, thanks! That is exactly the kind of summary I was hoping for. I’m on the fence about Megacorps and Federations, so that’s wait-for-sale for sure. I need to figure out if I want the species packs, because I enjoy the sci-fi stories in my head with Stellaris way more than in other games and the cosmetic stuff probably helps with that. Also, I just read Gideon the Ninth so the whole space-necromancer thing is particularly appealing, heh.
You forgot to mentioned Synthetic Dawn, where you can play a machine race. They can live on any world, and you have some options like kill bots, assimilators, servitors (they take care of their living pops) or just a regular race that happens to be machines.
I lumped them in with the Species Packs, as they are mostly cosmetic but some add in new mechanisms like Necroids, Robots, Hive Minds, etc.
I actually picked up Synthetic Dawn because it was classed as a “Story Pack” in the store, and in fact I just started as a servitor empire (the default one that’s based on earth). It’s cool!
I feel like I need some space in my life and I’m trying Stellaris again. Since release I’ve checked from time to time that it’s still a boring mess.
I feel like you hit the right spot. Their system of external interactions are probably less complex than EU4. But in EU4 quantity morphed into quality; you got hundreds of actors so relatively simple predictable interactions never gave you a predictable picture. In Stellaris with some experience you can instantly see which empires will be your friends/allies/federation members and which ones you will have to fight. You can probably go against the tide a little but there’s little point to it. They can’t make space more populated because their economic model already struggles with processing so many factions. Plus they know people like exploration aspect of Stellaris so they want you to explore and colonize for a long time before you run out of space.
Yeah, I appreciate that they continue to work on it, but all of these efforts to refine the micro and then automate away the micro - it’s digging a hole and then filling it in. I remember an early interview where they mentioned that the tile “minigame” was so the player isn’t bored at the start when they have only one planet - but it doesn’t scale. The economic model they introduced in the 2.0 version is interesting, but also doesn’t scale.
I think in this sort of game, scale would be an interesting thing. I’d love if the economy was more abstract, but the internal political mechanics were more fleshed out and had teeth. That is something that could scale to the end-game. I’d love if there was a strong tension between internal consolidation and stability and rapid expansion. If you don’t want to end up a small fish, you need to push hard enough that you risk fragmentation, but if you push it too far you spend a long time putting Humpty Dumpty back together.
The original goal was to have Sector governors take over as your empire grew which would have eliminated most of the scaling issues. The problem is that they’ve never been able to get the AI in a good enough state to handle it so no one wants to automate anything. I’m sure part of that problem has been the massive overhauls the game has repeatedly seen, talk about a horrible moving target for any AI devs. Finally get the AI somewhat decent at the tile game / economy? Well, rip it out and write a new AI from the ground up that handles pops!
Sorry to beat a dead horse, but this is why I want a Stellaris 2. I’ve had fun with Stellaris and I do actually like the changes they’ve introduced. I find the pop economy a lot more engaging than the tile-based one. But at this point, just take your lessons learned from Stellaris and start fresh with Stellaris 2. I think it could be a really strong game. Stellaris is 4.5 years old, it’s not like they’d be cutting and running. Who knows, if it’s not Victoria 3 then maybe that’s what Martin Anward has been working on since moving off of the Stellaris project.
I used to think that this was OK, but as time has gone on, I’ve become more aware of how difficult it is to get AI right, and how rare are the people who can really make it work well. If I think of AI dev time as a very limited resource, I’d much rather they spend time on making the enemy players as effective as possible and not on automating away gameplay that could be abstracted away.
That goes double when the gameplay that we want automated away is actually costing a huge amount of CPU. Ultimately the game is spending a large amount of AI dev resources, designer resources and CPU resources on fixing things that don’t really need to be there in the first place.
I mean, have you ever played EU4 and wished that you could dig into the your provinces and re-assign pops to different jobs, find out why Peter in Oxfordshire isn’t happy with the government?
There is some nice flavour in the more detailed economic representation, but I think Stellaris 2 could come up with a more abstract system that also doesn’t impose a significant maintenance burden on the player or AI and doesn’t cost many CPU cycles. Then you could have much bigger maps, a bigger range of empire scales and more interesting galactic diplomatic situation.
I wonder if the whole Paradox realtime model is a bad fit for Stellaris. The realtime thing kinda makes sense for history-based games, in that you need several timescales: it doesn’t make sense that wars, or moving your troops around, takes hundreds of years, but they have to work on the same clock as your entire empire in a turn based game, and you don’t want a million turns. Realtime allows you to slow down for the war-torn years and then zoom past the development stage.
But the other side of the coin is that realtime is really very much about waiting for the fun to start, and then not knowing if you should be doing stuff. The pacing is all off, as the devs have no control over it. If you have a space or fantasy-based game, realistic time scales are no longer a factor, and you can just try to make sure that interesting decisions and fun stuff is happening as often as possible (i.e. per turn). It’s also much easier to make effective AI, as you’re not limited by real-time constraints.
That’s a fair point, but I think in this case the code would be one and the same as I don’t know why the sector management would need to be different for a AI player than a PC. That is, the sector AI being so bad is indicative of the overall economic AI being busted. Fix one and you fix the other to a large degree, I think.
For sure. But why have a layer where the efficiency of each planet depends on how well you solve a complex optimisation problem (you being player or AI)? One good answer is if you want the game to be about optimising planets, but I think a more engaging game would put the focus elsewhere. If the focus IS elsewhere, you are still spending resources fixing a problem that you created.
You could have where you basically choose between some pathways for how you want to develop the planet, and it just updates it’s development levels, population and outputs in various categories according to reasonably simple formulae. You choose what the “development strategy” for the planet is, and the “development implementation” layer is just abstract. AI dev and CPU time that is currently spent solving the development optimisation problem at the planet level could instead be spent optimising military, diplomatic and colonisation strategy.
Sounds like you want the sliders from MOO.
I tried a game with the latest patch and once I turned on sector AI it completely tanked my economy almost past the point of recovery. There are extensive patch notes on how the sector AI has been fixed.
They don’t seem to understand the only way a sector AI is going to work with a human player is to have the player able to set production goals, for example I want my Empire to be producing X excess minerals and Y excess energy and then go ahead and build the research/alloy stuff once those needs are met. It is… incredibly frustrating.
Why literally write an entire blog about how you’ve fixed sector AI when clearly no one that works there even played a game and turned it on in the midgame.
I feel like I just need to wash my hands clean at this point. There was a point in the games life cycle it was fun, and sadly it was with the terrible tile system, because at least the AI could use it without breaking.
I’m with @KevinC. Call me back for Stellaris 2.
Yeah, kind of. Planets in MOO are simpler than the sort of thing I am thinking of for Stellaris, but they are admirably self driving. Things can get a little fiddly on the margins but they mostly do things that make sense.
So I don’t think sliders themselves are what I am looking for, but I like the general idea of you set an overall policy and your planet’s numbers increase organically from there.
Edit: Planets in the original SotS are also self-driving, basically one-dimensional. But you still had a decision to make about how much to invest in the initial colony to get it going faster, and whether is was worth the expense to settle a marginally habitable planet during it’s long development period. I’d like colonies in Stellaris 2 to be multi-dimensional, to capture the flavour of different types of planets and different population mixes. Just have the numbers grow according to formulas rather than lots and lots of discrete build decisions.
Like all other things good in this game, the answer is to steal Distant Worlds design. By that I mean a simple planet economy that couples well with strategic resource mining and other more advanced systems. They’ll figure it, eventually.
They modelled it after the demense system in Crusader Kings, but it works so much better there. I think for the reasons that you state. In CK you only build buildings and their relative position doesn’t matter. It’s also the cause in CK that your vassal’s territory isn’t “yours” the same way that it is when a governor is managing a planet in Stellaris. In fact, I have met more than a few vassals in CK which I would not mind if they made a mess of their territory.
Which leads me to my biggest gripe with Stellaris, which is that even with all those patches and redesigns they were never able to make internal politics interesting or even a threat.
Dev diary #193 is up. Looks very interesting: https://forum.paradoxplaza.com/forum/threads/stellaris-dev-diary-193-signal-discovered.1444592/
I grabbed the deluxe edition off psn for 15 bucks as a change of pace for me. Hoo boy! I have no idea what I’m doing, but I’m scanning and building stations, and the game keeps telling me shit that my folks are finding shit, so…yay?
I do enjoy the chill experience thus far.
I think the biggest problem is actually that it isn’t turtles all the way down, as currently implemented, so at some point you get to a layer where deterministic effects stem from precise decisions and it becomes a puzzle instead of a strategy. And that core puzzle isn’t enough fun to be so important. If it was sliders or zoning and you knew that all the sliders and zones really did was tip the scale on each stochastic decision that cropped up, then the organic growth wouldn’t be frustrating even if it was imperfect. It would also scale to sectors and even larger groupings with no difference, the way the CK “council” does.
So instead they try to automate puzzle solving which means either: lots of budget for very advanced AI, AlphaZero-style; very simple puzzles that completely bore players; or automation that screws everything up.
Ah, more meme content…
For far too long has the galaxy remained safe from player hands. In fact, more often than not, the player has needed to be the defender of said galaxy. No more! Finally has your time come to become the crisis !
It might be interesting, but not really the most needed thing.