Stellaris Dev Diary #325 - 3.10.3 "Pyxis" Released [d2aa] + Further Beta Plans

[SirBlackAxe]

Something to keep in mind is that tech speed is greatly influenced by which DLCs a player has.

Splitting the tech/tradition slider so players can slow research without kneecapping unity is absolutely key here, as I doubt it's possible to have tech speed balanced just as well for someone no DLC vs. someone with e.g. Galactic Paragons and Ancient Relics.

 

[mrt1212]

Your statement about unity over science being known by experienced players to be objectively wrong is... objectively wrong (sorry, couldn't resist it), since I am an experienced player and this summer posted a ridiculous unity over science counterexample (unity output > highest of phy, soc, eng output throughout the game), with direct control of around 250 system and 170 planets, that yet managed to be, if not second to none, then second to very, very, few high tech builds, as it achieving shield repeatable 100 in the mid-2330s and had around 70 fleets of 3 million fleetpower each in the early 2340s when I got bored of building ever more ascended research ringworld... in a huge galaxy in unmodded Stellaris with default settings (apart from difficulty). Granted, that game had a really good start and I only expected that level of performance 2-3 decades later based on the earlier games in that test series optimizing the build, but still, it is a counterexample. I thought you had participated in some of those discussions, but perhaps I remember wrong.

Granted, that was a special case and I certainly don't recommend people play with unity over science currently unless they have a really good plan based around ascensionist wide gameplay (but not as ridiculously wide as I did for demonstration purposes, since some of the game mechanics I took advantage of to achieve that were eliminated as part of the 3.10 nerf), but it is an option available to experienced players.

So as far as I am concerned, the interesting thing about this will be to see how forum consensus deals with the unity/science interplay with the changes. Given the extreme reluctance the player base has had embracing unity since 3.3 despite its strong mathematical properties, I am not as sanguine as you about players embracing a more balanced approach, that is superior even now, or occasionally favouring unity over science, that can be superior now, but generally isn't, rather than doubling down on science.

But I hope you are right. And it will be especially interesting if we end up in a situation where alloys over both unity and science makes sense for some warlike builds. THAT would be awesome.

 

[Vreivai]

My hypothesis is this is going to help in the mid/late game (because the player gains science faster than tech costs go up on live) but hurt the early game.

I do feel in today's game that the first few techs are already glacially slow to acquire. 2-4 years for a tech seems about right, more than four years feels slow outside of super-special techs. With both increasing the T1 tech cost to 2500 and slightly nerfing researcher output, it'll tack on a year or two for those first techs (baseline, pre-expansion).

If this does indeed turn out to be the case, I think they could just bump up the base research each empire gets (iirc all empires get 10 each) to 20 or so. It means we'd start with slightly more science than live (to counter the +500 tech cost) but wouldn't really make any difference in the mid/late game.

I think it's more going to be the opposite. Sure, it'll be slower in the early game, but T1 techs are only 50% [derp, 500 is 25% of 2000, not 50%, T1 techs are only 25% more expensive] more expensive. T4 and T5 techs are multiple times as expensive, and most of the reduction to research output (loss of +60% researcher output from techs, reduction in research speed increases and increased researcher upkeep from increased research speed) only comes into play later in the game. From some quick math I did, it looks like higher-tier techs are going to see an effective increase in cost of 10x+ when you factor in the lower research output.

 

[-Marauder-]

I think it's more going to be the opposite. Sure, it'll be slower in the early game, but T1 techs are only 50% more expensive. T4 and T5 techs are multiple times as expensive, and most of the reduction to research output (loss of +60% researcher output from techs, reduction in research speed increases and increased researcher upkeep from increased research speed) only comes into play later in the game. From some quick math I did, it looks like higher-tier techs are going to see an effective increase in cost of 10x+ when you factor in the lower research output.

It's going to be slow and miserable across the board. Someone early did the math comparatively. Even without the impact on economy and all, we're looking at finishing tech trees in 2700-2800. Hardly any game lasts that long. Or is particularly fun at that point.

 

[Millbot]

I'll also throw out there, that I hope they take a look at how special projects feed into things. A good deal of those do slow down our research and it's often for engineering, but there are a fair number that also impact society and a few that impact physics.

It's probably worth asking if those research gated projects, that also slow down research in a specific field, are adding to the gameplay experience or if they are actually harming. I kind of learn towards harm because they really screw with pacing. As in it does feel like the overall game expects you to hit certain techs by a certain point and that is often not factoring in a number of special projects that have been added in over time. It's also a factor in why we hit repeatable techs in the physics tree so much earlier than other trees. You don't get any species modification projects that eat into physics research and outside of some early contact stuff, you don't see too many projects that require you put physics on hold to do them.

I'd argue that the devs could achieve both the time gating goal and the goal of having that gate scale with research output, but going with a setup where the special project can be done without putting the current research on hold and the time it takes will factor in what your research output is. So you could have the species modification project go without slowing bring engineering or society research to a complete halt and still have it be quicker for the guy with 80 sprawl and 3 research labs, while the dude with 200 sprawl and one research lab has it take much longer still (pretty sure no one does that, but I trying to throw out a extreme example).

Now I'm going to somewhat disagree with people on the megastructure front. I'd argue the ones we need to unlock galactic wonders should actually be things we can get midgame. I'd even argue that if building a certain number of galactic wonders aren't going to be a win condition. Then those should be something that we can start getting online towards the final part of the midgame. Otherwise that's going to put all the megastructures in the position that a freshly build ringworld sits in, where there isn't much point because the game is essentially over by the time you finish having it fully setup. Heck, that would make building ringworldsfrom scratch not even worth doing in some ways because it doesn't just take a long time building the thing, you then have the time it takes to colonize and then build up and populate. I will say, I'm perfectly fine if that means having to adjust costs and output of megasructures again, even if that makes them seem a bit less super duper awesome.

I'd also say the ship has long since sailed on the any concept of limiting players on only building some of the megastructures. We've always had the option of building at least one of each from scratch, we have a relic that doubles that number and we've beenable to get them all up and running for some time now. Again, only way you could really justify making megastructures end game, is if you either made them super powerful or made having a certain number complete a win condition. That said, we do kind of need the option to have more of those, than the current setup of get the high score, which is usually just conquer everything or get so big that you outpace everyone in everything else.

Probably a big reason why they are doing the heavy hit to research speed though. They want to see if anyone can find creative ways to either hit repeatables or get certain things up and running well before they should be able to. For every 10 people bemoaning how they think the devs don't know what they are doing with the research changes, there is probably one player happily rubbing their hands together thinking about which traits, techs, events and relics they can try gaming to get well into repeatable techs or have all the megastructures well before endgame and some of them are probably going to find some obscure crap, where they'll actually have to use the game commands to make it reliably happen for actual testing purposes.

 

[mrt1212]

I'll also throw out there, that I hope they take a look at how special projects feed into things. A good deal of those do slow down our research and it's often for engineering, but there are a fair number that also impact society and a few that impact physics.

It's probably worth asking if those research gated projects, that also slow down research in a specific field, are adding to the gameplay experience or if they are actually harming. I kind of learn towards harm because they really screw with pacing. As in it does feel like the overall game expects you to hit certain techs by a certain point and that is often not factoring in a number of special projects that have been added in over time. It's also a factor in why we hit repeatable techs in the physics tree so much earlier than other trees. You don't get any species modification projects that eat into physics research and outside of some early contact stuff, you don't see too many projects that require you put physics on hold to do them.

I'd argue that the devs could achieve both the time gating goal and the goal of having that gate scale with research output, but going with a setup where the special project can be done without putting the current research on hold and the time it takes will factor in what your research output is. So you could have the species modification project go without slowing bring engineering or society research to a complete halt and still have it be quicker for the guy with 80 sprawl and 3 research labs, while the dude with 200 sprawl and one research lab has it take much longer still (pretty sure no one does that, but I trying to throw out a extreme example).

Now I'm going to somewhat disagree with people on the megastructure front. I'd argue the ones we need to unlock galactic wonders should actually be things we can get midgame. I'd even argue that if building a certain number of galactic wonders aren't going to be a win condition. Then those should be something that we can start getting online towards the final part of the midgame. Otherwise that's going to put all the megastructures in the position that a freshly build ringworld sits in, where there isn't much point because the game is essentially over by the time you finish having it fully setup. Heck, that would make building ringworldsfrom scratch not even worth doing in some ways because it doesn't just take a long time building the thing, you then have the time it takes to colonize and then build up and populate. I will say, I'm perfectly fine if that means having to adjust costs and output of megasructures again, even if that makes them seem a bit less super duper awesome.

I'd also say the ship has long since sailed on the any concept of limiting players on only building some of the megastructures. We've always had the option of building at least one of each from scratch, we have a relic that doubles that number and we've beenable to get them all up and running for some time now. Again, only way you could really justify making megastructures end game, is if you either made them super powerful or made having a certain number complete a win condition. That said, we do kind of need the option to have more of those, than the current setup of get the high score, which is usually just conquer everything or get so big that you outpace everyone in everything else.

Probably a big reason why they are doing the heavy hit to research speed though. They want to see if anyone can find creative ways to either hit repeatables or get certain things up and running well before they should be able to. For every 10 people bemoaning how they think the devs don't know what htey are doing with the research changes, there is probably one player happily rubbing their hands together thinking about which traits, techs, events and relics they can try gaming to get well into repeatable techs or have all the megastructures well before endgame and some of them are probably going to find some obscure crap, wehre they'll actually have to use the game commands to make it reliably happen for actual teshing purposes.

I think the thing that I've mostly experienced since 3.9 being kind of a slouch-pace-keeper with research is that the AI builds them so I can just conquer them, no sweat off my back, and the Nexus and Command Center are the real prizes of the AI getting there first. I don't know if this is exactly a problem and I certainly enjoy skimming the cream off the top like this while everyone else rushes to be firsties to even produce the cream, but I do wonder if the delay of their onset would evaporate that kind of dynamic.

 

[Vreivai]

It's going to be slow and miserable across the board. Someone early did the math comparatively. Even without the impact on economy and all, we're looking at finishing tech trees in 2700-2800. Hardly any game lasts that long. Or is particularly fun at that point.

Could you provide a link to that? I'd be interested to see it. I was thinking more like 2600 based on my progress in a game I'd played to 2431, and even then was worried I was being a little too pessimistic despite trying to be conservative in my estimate. Also was more a ballpark estimate.

 

[Thiend]

Your statement about unity over science being known by experienced players to be objectively wrong is... objectively wrong (sorry, couldn't resist it), since I am an experienced player and this summer posted a ridiculous unity over science counterexample (unity output > highest of phy, soc, eng output throughout the game), with direct control of around 250 system and 170 planets, that yet managed to be, if not second to none, then second to very, very, few high tech builds, as it achieving shield repeatable 100 in the mid-2330s and had around 70 fleets of 3 million fleetpower each in the early 2340s when I got bored of building ever more ascended research ringworld... in a huge galaxy in unmodded Stellaris with default settings (apart from difficulty). Granted, that game had a really good start and I only expected that level of performance 2-3 decades later based on the earlier games in that test series optimizing the build, but still, it is a counterexample. I thought you had participated in some of those discussions, but perhaps I remember wrong.

Granted, that was a special case and I certainly don't recommend people play with unity over science currently unless they have a really good plan based around ascensionist wide gameplay (but not as ridiculously wide as I did for demonstration purposes, since some of the game mechanics I took advantage of to achieve that were eliminated as part of the 3.10 nerf), but it is an option available to experienced players.

So as far as I am concerned, the interesting thing about this will be to see how forum consensus deals with the unity/science interplay with the changes. Given the extreme reluctance the player base has had embracing unity since 3.3 despite its strong mathematical properties, I am not as sanguine as you about players embracing a more balanced approach, that is superior even now, or occasionally favouring unity over science, that can be superior now, but generally isn't, rather than doubling down on science.

But I hope you are right. And it will be especially interesting if we end up in a situation where alloys over both unity and science makes sense for some warlike builds. THAT would be awesome.

I was part of it, and I actually agree, but using unity over science properly is genuinely difficult - it's kind of an edge case, although you can get dramatically more out of unity than science (to a point, you need science for Ascension Theory) if you know what you're doing.

 

[Nevars]

Rebalancing tech and nerfing ship cost reduction is a long time coming and I welcome them.

But I will be against any and all attempt to put a hard cap on fleets number, this is war on galactic scale not school yard fight.

If anything, we need more ship, not less of them.

 

[Vreivai]

If anything, we need more ship, not less of them.

Ships are a major contributor to late game slowdown. More ships would be heavily detrimental to the game. And especially if Paradox wants to go the route of lengthening the game, as these changes appear, then something has to give. And with how the game is currently set up, I think cutting ship numbers would be much easier than consolidating pops.

 

[Millbot]

I think the thing that I've mostly experienced since 3.9 being kind of a slouch-pace-keeper with research is that the AI builds them so I can just conquer them, no sweat off my back, and the Nexus and Command Center are the real prizes of the AI getting there first. I don't know if this is exactly a problem and I certainly enjoy skimming the cream off the top like this while everyone else rushes to be firsties to even produce the cream, but I do wonder if the delay of their onset would evaporate that kind of dynamic.

Honestly, I'd like the devs to revisit how megastructure restoration works and how conquests works with them. My key complaints.

-If you're competing with other empires that can maintain similar tech levels, you're actually more encouraged to let them restore a ruined megastructure and then conquer the system in it. If there is no colony in that system and it's on your border, it'll cost less influence. Also by letting them restore it, you don't have to spend the allows. Upping the odds of getting the megaengineering tech card isn't enough.​

​

-Worse, with how the game is currently coded. Once an empire starts restoration of a megastructure or starts on construction past the site construction. The process will continue, even if the empire that started the process, no longer own the system IIRC. Been a year since I last saw that with one of the few war in heavens I fought, so they might have fixed that nonsense.​

​

-It doesn't really make any sense that my empire could be way behind an empire with a science nexus and then I get full benefits of that science nexus despite never ever researching how a science nexus works. Heck, I can even continue the build process if it was a fresh science network that was completed.​

What I would like to things better.

Restoration should be rewarding enough that people don't want to just wait for some other empire to do it.

-The first time an empire restores a ruined megastructure, they should get an empire buff "Megastructure Restoration," that reduces megastructure build time by 5% and costs by 2.5%. I'm suggesting modest because that then leaves the door open for the idea that additional restorations could bump those bonus by very small amounts. So the second restored megastructure might bump the bonus up from 5% to 5.2% megastructure build speed and reduce the cost by 2.6%.​

​

-If the megastructure is not a galactic wonder. The empire will get the tech card to research it. If they already have the tech card, then it'll add 10% to the research progress. If they've fully researched the tech, they'll get a fraction of the relevant tech points as a consolation prize. Galactic wonders would work similar, but it would be setup so that you don't get the tech research cards and you only get the tech progress if you have the perk. Real trick is making it so that can be gotten retroactively. AKA if I restored a dyson sphere to unlock the ability to select the perk, when I've my dyson sphere tech research card, the came will credit it as partially researched.​

​

-Regardless of what structure is restored, the game will flag that empire as being able to get full benefit from megastructures of that type. So if I restored a dyson sphere, any other dyson spheres that come into my possession, I can get the full energy output.​

​

-Get a lump sum of unity because it should be "hey, guys we restored this awesome ancient megastructure that some dead civilization built because we're such badasses," while also getting a lump sum of minor artifacts (yes, the devs do need to adjust some things with that because this would be the point where people might be at artifact cap, but I'd argue letting us better track CDs that use them, might alleviate this somewhat).​

​

-Of course, it would still be beneficial to have the ruined megastructure in your territory, to up the odds of getting the relevant tech cards.​

The above would probably make it much more exciting to not only find a ruined megastructure, but also restore.

As for conquest and megastrucgtures.

​

-First and foremost, restoration or build progress on stages passed laying out the build site, should pause if the initiating empire no longer controls the system. Once that empires regains control, said progress will continue until they are either no longer in conrol or the process is complete. In the vent that the war ends and they lose ownership, all restoration or construction progress should be reset. The new owner will have to start from scratch if they want to restore the megastructure or start the beginning for the next phase of consruction on a new megastructure. The original owner could and probably should get at least a partial refund on the alloys, maybe a complete one.​

​

-If an empire conquers a megastructure that someone else has either restored or built from scratch. They should be forced to do a special project get the structure online for their use because I think it oculd be argued that the prevous owner will do some sabotage to the structure upon realizing they will loose control of it. This would slow down the conquest snowball, while still allowing people to benefit from conquering a system with a megastructure. You just won't get immeadiate use out of it and you might not get full use depending on what you've researched, restored and/or invested APs in.​

​

-The next thing is upon completing the special project to access their new megastructure. The new owners should get hit with a penalty if they have not either restored a similar megastructure or fully researched it. So for example, if I jack someone's matter decompressor and never restored one, nor bothered to invest in the galactic wonders perk. Then I'm not going to get the full output of it once, it's back online because my empire doesn't fully understand how the thing works. If I'm out of perk slots, well to bad for me, if I wanted a fully functional matter decompressor, I shoudl have either gotten the perk or found a ruined one and restored it myself. This adds both value to the the galactic wonders perk and megastructure restoration, while also slowing down the conquest snowball. Though if I jack someone's strategic coordiantion center, but havne't gotten the tech card for it yet, I'll have an increased chance of drawing that tech card noq.​

​

-Finally, if an empire lacks the reserach for a megastructure, as in it's either incomplete or they haven't pulled the tech card for it and gain control of someone's incomplete megastructure for that tech card. Then they cannot click on that megastructure and tell the game to start building the next phase until they have completed the research. In the event that we're talking about galactic wonders, if they aren't functional ones, then the game should treat their upkeep similar to that of a runined megastructure. So If I steal someone's system with an incomplete ringworld, while never picking up galactic wonders, the game won't punish me with upkeep for a ringworld that I can never complete. I mean, I"m already being somewhat punished because I can't complete it and maybe I should have let the previous owner finish; especially, if I happened to wipe them out.​

Those changes would make the whole conquest thing make a little more sense, still retain value for people that conquer someone else's system with a megastructure, but also adds some reasonable costs to that conquest.

 

[mrt1212]

-It doesn't really make any sense that my empire could be way behind an empire with a science nexus and then I get full benefits of that science nexus despite never ever researching how a science nexus works. Heck, I can even continue the build process if it was a fresh science network that was completed.​

Yeah, thats what made me think it was cute if not clever play among a small handful of other reasons to let AI empires do some of my work for me.

 

[Ajacoba]

It's going to be slow and miserable across the board. Someone early did the math comparatively. Even without the impact on economy and all, we're looking at finishing tech trees in 2700-2800. Hardly any game lasts that long. Or is particularly fun at that point.

Doesn't the graph completely ignore the rubberband and the other sources of progression, like stealing tech from AI or reverse-engineering debris?

 

[-Marauder-]

Doesn't the graph completely ignore the rubberband and the other sources of progression, like stealing tech from AI or reverse-engineering debris?

You can only do that if someone has researched the technology already. We're talking about the trailblazers here. It doesn't speed up tech progress in any capacity, it just makes it so that weaker Empires will keep continuously pace in some misguided attempt at equity of outcome.

 

[Panzerslothen]

But to the more crucial point - does it feel like it's 'there' yet, really?

And per the popular meme:

View attachment 1060767

See I don't even think this meme is entirely accurate (enjoyable and amusing though it is). It's less a matter of relearning the
entire game - we've not really had to do that since 2.0 or, at a push, 2.2 or 3.0. Everything since then has been a greater or lesser degree of iteration. Core mechanics have remained relatively consistent since 3.0 (or 2.0, depending on how you want to imagine it).
Instead, it's a matter of (re)learning those new and changed features, often in subtle ways.
I get that this is possibly a semantic argument (although I don't really think that it is).

I get what you mean about 'is it there yet?'. Although I think the Custodian initiative is one of the best things since sliced bread, it has had an effect of making some people think the game is in a permanent beta or pre-release state.

That said I am now at 2098 hours (half of which was modded, non-ironman, non-achievement runs under pre-3.0 editions), and every run still teaches me something. Stellaris has a depth to it, but with that depth of course comes complexity.

 

[Panzerslothen]

Granted, that was a special case and I certainly don't recommend people play with unity over science currently unless they have a really good plan based around ascensionist wide gameplay (but not as ridiculously wide as I did for demonstration purposes, since some of the game mechanics I took advantage of to achieve that were eliminated as part of the 3.10 nerf), but it is an option available to experienced players.

Ascensionist.... WIDE?!

You had my attention; now you have my interest. Teach me your ways, oh great sage!

 

[ashbery76]

The game should be balanced around the core Stellaris gamer ala the single player and not Tube steamers or MP that want the game finished in a afternoon.

Epic long games that tech lasts until end game.

 

[prismaticmarcus]

I haven't seen such childish squabbling in a long time.

Let's remember we're here to talk about a game, not each other.

 

[Enailios]

I haven't seen such childish squabbling in a long time.

Let's remember we're here to talk about a game, not each other.

Was gonna ask since i’m new here on the forum (soon 3k in the game!) if this was common? While i applaud the passion and lively debate (on a mildly inscrutable but crazy fun game), do wonder if some here be 12 year olds (scratch out a word) who are a tad too bored. In any case, i’ve noticed the moderators’ hard work and I thank y’all for it

 

[Starfury]

Ascensionist.... WIDE?!

You had my attention; now you have my interest. Teach me your ways, oh great sage!

That particular strategy won't work quite as well in 3.10 as it relied on a (now removed) governor trait generating 9x monthly unity per blocker cleared.

I've played a 3.9 game where I used that trait as well as the explorer destiny trait to generate unity out of the wazoo. I ended up with ~30 max ascended worlds with <100 empire size in the 2280s. (Though nowhere near or on track towards the economy Peter had in the 2300s unless I went on a nihilistic acquisition spree of galaxtic scale)

Needless to say I'm not surprised both traits didn't make the cut for 3.10