I picked it up this afternoon. :)

I'm only a few hours in as I'm going slow... but damn if I'm not totally into this game and I'm glad I didn't hold off on it. I've played many 4x games, but Stellaris really gives me that "unknown universe" feeling that no other 4x game has ever done for me. I'm always checking up on my survey ships to make sure aliens aren't stalking them.

The UI is so damn impressive too. I've never played a Paradox game and I figured this game would be too much for me to really want to get into, but they have done an excellent job, at least in the early game.

I'm not usually able to get into Paradox games, but this one has me completely hooked. I'll echo everything said here.

Sweet! I think the early game beats the pants off anything I've played, ever. The trick is to raise everything else closer to that standard.

For those scared off by the typical learning curve of a Paradox game, this doesn't have it. The tutorial was pretty well done. If you are familiar with Space 4x concepts, this game will not be hard for you to grasp. The UI is excellent, although I would love to make the text in the pop ups a little bigger without changing screen resolution.

This is really an excellent game for me so far. I am sure I will find some things I don't like, but the innovative ideas around the tech tree are awesome.

I love when something is done building on a planet, you get that green pulsating circle.. just little things like that make me giddy. First RTS game I've actually liked, typically I like turned based. :)

It has been a complete experience so far, starting with customizing your race.. something I NEVER do, but did in Stellaris.

I really like the game so far. It's my first day one purchase in eight years, and I'm glad I didn't wait. Exploration is usually my favorite part of 4x games, and this has been fantastic. I read on the Paradox forums that there may be some sort of archeology victory condition. I'm looking forward to seeing how that plays out. I'll also add a big +1 to the UI praise.

My one complaint is sectors, and it appears that I'm not alone. The tutorial prompted me to do it long before I needed to. I think I had only three planets at the time. At least in the one game I've played, there are only so many planets to settle early on, so it wouldn't have been that difficult to mostly develop the planet before forming the sector. I also don't like being penalized for deciding against including a system while initially creating it.

What just bit me was finding out that they aren't able to use strategic resources from outside of the sector. I ended up having to reload after that. Now I have to convert the two Betharian Power Plants into standard ones, then upgrade, then reform the sector all before I can settle another planet. I don't necessarily disagree with the game design, but it would have been helpful to know beforehand.

As far as being inefficient, it seems like a no-brainer to prioritize food tiles until it's fully (or mostly) grown, but they definitely don't do that. It would be one thing if there was some dire need for something else, but there just isn't. Maybe the sectors without governors could be controlled by some empire-wide template. I actually like the sector concept, but it needs some tweaking and maybe a little more granular control.

All things considered, a big thumbs up so far.

FYI while there are a lot of issues right now with sectors -- one of the worst seems to be that they don't have the influence cost required to upgrade the planetary capital buildings -- this is actually not a no-brainer. There isn't a huge difference between, say, having a surplus of 5 food versus 2. It's not a linear improvement, which I think is fantastic. You can only jumpstart growth to a minor degree. The strategic resource thing is another concern. The project lead has basically commented that they are paying close attention to sectors, so they know it's a problem. I'm interested to see what they do with it, it appears it hasn't been fully decided yet, which makes sense at this point but it's something I think they need to take significant steps on ASAP. Then everyone will move on to something else to complain about :).

Interesting. That's what I get for making assumptions. I guess I've played too much Civ :)

I made the same assumption at first. Basically you get 1.0 toward your next growth target with 1 food surplus, and if you have like 7 food you get like 1.5 or 1.6. So it helps ... but you can't just grow pops like 4x faster or whatever just because you have more food.

That is good to know. I know most LP's I've watched seem to focus on minerals.

Discovering more stuff that I was wrong about. So I should announce here. This shouldn't be a problem. I'm mostly through digging through the files to see what mod possibilities there are though basic editing. Getting to where they deal with the resources I find this under influence:

"sector_ignore_cost = yes"

In other words, sectors shouldn't have to pay the influence cost for stuff. Which causes exploits(giving a sector a frontier outpost, for example), but also means they should be building those key planetary capitals without worry. I need to stop believing what I read on the forums, there's so much misinformation right now about sectors :).

That's probably the way to go at first. There's no shortage of things you need minerals for -- especially early.

After a decade or two though I found energy much harder to come by than minerals. For power you need power plants. You can't build power plants upgrade without an admin building. You can't build an admin building without 5 pops. Even with what Brian pointed out, I still think I'm going to go with food for the first four tiles. After that, I may very well change things up.

This isn't meant as any sort of advice though. I'm pretty much just winging it my first run through. But that's been my observation to this point (just finished colonizing my eighth planet).

I agree with focusing on minerals. Basically you need minerals except when you run out of energy to do the upkeep on the buildings getting you the minerals. Then you need more energy, so you can pay for the buildings that are getting you the minerals :). It's sort of a tug of war between those two.

Also, if I may vent a little. I found this near the very bottom of the biggest 'catch-all' file that we can easily change(defines.lua if you care). If you've played the game much you know how important each planet's 'capital' building is. They increase production on the four adjacent tiles for any food, minerals, or energy produced there. Getting the right buildings in those places is super-important.


I ... guh ... WHYWHYWHYWHYWHYWHYWHY??? If that does what I think it does, simply changing that 1 to like a 3, or 4, or whatever the heck the biggest number is, should do a crapton to help the AI not being a complete idiot(and it sure explains why they are so hesitant to build up sector capitals).

I've found, and it's just plain logical, that upgrading those capitals as soon as humanly possible(i.e.,. when you hit the pop requirement for them), is huge. I'm both happy to have found this, and at the same time I want to scream. Somebody put that in there. On purpose.

WHY? :)

Understanding how important planet / colony buildings are was a huge moment for me. I had been driving my economy almost entirely from mining planets that I did not colonize; once I realized how valuable a single power plant (let alone a Betharian Power Plant) was, it changed my world. I, like most, opted to push for a huge mineral surplus early. Which was awesome, until a few hours and a few outposts in when I started facing huge energy concerns. I was worried, because I had gone a long while since finding a system that had ANY energy available in it, until I started looking into planetary upgrades. Between surface buildings and Spaceport modules, I was able to completely flip around my energy problem in no time at all.

Almost makes me want to start over so I can build things up properly from the beginning!

This has kinda been the view-from-the-peanut-gallery question for me since the discussion about sectors started to emerge.

My gut reaction -- and still my guess at this point honestly -- is that the limits on human control might be an attempt to cover up some AI shortcomings (that may or may not have even appeared at early stages of the game). Maybe humans kicked the crap outta the AI routinely in testing if they had full or highly expanded control, so this could be a means of artificially preventing that kind of imbalance.

I ran into the sanctuary. Interesting

There's some truth to that Jon, but I think it's more in the inherent design category than Stellaris-specific AI weakness. There's got to be a better way to say that :P. What I mean is, they have talked about that fact that no AI will compete with a good human player. That's more of an issue here than in CK or EU because you've basically got symmetrical starts going on, not a wide variety in nations, holdings, and so on which is the opposite.

I do think that issue is part of sectors, but it was announced pretty early in the process and it does have other things associated with it; factions that can potentially rebel based on sectors, the whole limiting micromanagement thing, and so forth.

My educated guesstimate is that feature creep and development schedule stuff happened. The AI behaves very reasonably in some parts of the game -- and in others it falls flat on it's face, then gets up for the sole purpose of falling again. I would surmise that they probably tried to add stuff too close to release instead of cutting it off early enough to test what was there. I don't know though, we're all just guessing on that score. Just going by my limited experience in software and what I've seen and heard from the team.

That kind of begs a question, though it's probably one that lends itself to a different thread rather than an outright sidetrack here. But I'll risk it, trusting us not to all go bonkers with it:

How many "good human players" are there? Or rather, what percentage of players would meet that threshold?

I'm not comparing them to each other mind you, simply "good enough to soundly beat the AI routinely & repeatedly" will suffice. Is that 90% of the player base? 20%? 50%? I honestly don't have the foggiest idea, not really.

And if it's a big number (which I'm guessing it's at least in the 33-67 percent range with these games), then shouldn't there be some degree of groupthink possible that would lead to an improved AI? I mean, if that many people are "good" by the beats-up-AI standard then surely there are some common traits/styles/choices among them.

I'm not looking for cheats, handcuffs for the human player, Maddenitis for the AI. I mean genuinely better/smarter/more competitive AI.

What is the AI missing, then, that makes it unable to compete adequately with X number of players?

Heuristics is the short answer. Putting heuristics (the use of a sub-optimal but passable strategy when a perfect strategy is impossible or incredibly time-consuming to determine) into computer logic is nearly impossible - but we're getting there. Essentially, the complexity of the game makes it ridiculously difficult to make that work from an AI standpoint. It's really hard to break down the disparate elements of the game into a simple 'do this' answer to a situation where there are hundreds of variables.

And to be honest, the first part of your question is perhaps more important than the second. For strategy games like this that are incredibly complex and have a difficult learning curve, you're probably at least greater than 50% "good" rating for gamers. Maybe even well clear of that number. But then you get into the issues with the company itself and its solvency. They are trying to make their game as accessible as they can to generate more sales - they don't necessarily WANT a ridiculously good AI, because they want the new business from customers who AREN'T as savvy as the average player, and kicking them around the arena in their first go-around isn't exactly good for business.

Maybe in a game like Civ, with 6 or 7 different difficulty levels, you can really prime the pump on the highest difficulty level. But Stellaris has "Normal" and "Hard" modes, and that's it. Which leaves little margin for error in terms of how good you can make the AI.

Yeah I kinda had a feeling that was the case but honestly wasn't sure how high to set the bar since, well, we may not be an "average" focus group around here to draw conclusions from;) And I'm sure this place does color my perceptions, especially on games of this sort.

By the same token though I also see some fairly well-trafficked YouTubers playing games (not this one yet obviously but HOI comes to mind) and doing incredibly dumb stuff than even I end up #smdh about while only half watching (I usually see these when my son is watching them). That leaves me wondering if, you know, maybe I don't suck quite as bad as I think I do most of the time.


And this leads back to my bit about scaling the AI up/down without doing stuff that feels very gamey. Cause that's probably the easiest way to scale the AI ... but it sucks when you know that's happening & you're on the wrong end of it.

There is a mod out to raise initial planet cap from 5 to 10. Debating on downloading it and seeing if it works with my current game where I'm at 5 planets and was about to create my first sector...

Dola -

It does work on current games. I'm going to use it at least until Sector AI improves.

Is that on steam workshop EF?

for those of you who have been playing non-stop, and are in no shape to go to work. Paradox understands and has prepared this: :)

https://forum.paradoxplaza.com/forum...nday.930484%2F

Yes, it is. I searched for more recent and it was on the 2nd page as of this morning.

Yeah, this is something I meant to touch on. The simple fact that we are on the internet discussing video games puts us on an entirely different tier of gamer than oh, I don't know...maybe 80% of the gaming world? Like I alluded to in the original post, this is probably far less true of this style of game (grand strategy) than most - for example, a game like Madden or The Division probably has a much lower threshold - but still something to consider.


And this is the other side of the coin - what do you define as 'good?' I first started really considering this when I was raiding at a high level in World of Warcraft. I demanded a really high level from my raiders, and the disparity between one of my guys and a random person 'off the street' was absurd. The discussions that would happen around high-end raid design were intriguing, because all of the top end players wanted an extremely high level challenge - but we're talking about something that would be difficult for literally the top 1% of the player base; things that would be almost impossible for 90% of the people playing the game. So how much development time should be given over to that?


This has long been the most fascinating part of video game design to me. I was a computer programmer for a hot minute in college, with a dream of making video games. That dream died a while back when I realized I liked playing them much more than building them, but this question still intrigues me. I have long since held to the belief that the reason AI is so sucky is that it is predictable - if you can find out what is 'motivating' the AI, you just use that to your advantage and push it into a situation where it makes a wrong move.

I never got far enough into programming to be able to really push things, but my thought was always to program in some sort of limited randomizer, in that when faced with a particular decision the AI would weight the choices, and then choose one at random based on the weighting. If it wasn't so clunky, make several different...I guess call them 'flavors' of AI that weighted different aspects more heavily than others. Make it so that the user cannot precisely predict what is going to happen. The link I put in my original post up there goes to the FiveThirtyEight article about the AI that beat a grand master Go player, and it uses a system called the Monte Carlo method, where an AI 'learns' by repeatedly attempting different moves and finding out how often it puts itself in a better position by making the move. I had never thought up anything as sophisticated as that, but at some point you have to wonder - a computer can calculate things perfectly. If we get the computer to a given place, will it be possible to defeat it? And again, we're back to the original thought - is the point that we want the AI to win all the time? Not exactly.

I just found that you don't pay monthly influence on outposts that are inside a sector. That will probably change my strategy some.


The first thing I saw while poking around in the defines file was:
TECH_COST_FREE_POPS = 10, -- First X pops do not add any tech cost
TECH_COST_MULT_NUM_POPS = 0.02, -- Each non-free Pop adds +X% tech cost

So I guess each pop over 10 on each planet increases tech costs by 2% ??

Yeah I think the outposts in sector thing is bad exploit/loophole that needs to be closed. On the tech thing, not on each planet, in your entire empire. Basically, as your pops grow you need to grow your research output also.

I've stored up quite a bit of society research. How do I use the stored research?

I am not sure what you mean by "stored up." I think as soon as you start research on a society tech, it will all get applied to that tech.

You research something. Every month you will draw out an amount equal to your current research points from what is stored, doubling your research speed until it's gone.

Example: 150 stored points and 10 research a month in whatever. You will research at a rate of 20, taking 10 out of your storage for 15 months. Then, having used up your stored points, you will go back to researching at 10.

That makes sense. Thanks.

I think my biggest complaint about this game so far is the lack of espionage.

I'm sitting next to an empire that I want to conquer. However the only thing I know about them is their pop size and how many planets they have, each which are about half of mine. Is there no way to find out things such as fleet size or technology level? Even knowing what planet types they like would be helpful.

I think it's just the speed factor involved with making videos. Good YouTubers not only have to explain what they are doing (which I bet is much harder than it looks) but they have to maintain a decent pace as to not bore viewers. So, they cut corners. There are things that even I notice when watching Arumba or quill playing EU4, and I've never even played it (quill not using any of his diplomats for oh, 50 years, or Arumba forgetting about an army for three videos).

Sometimes it's just tunnel vision. Focusing on one thing while completely missing something obvious. Probably happens to everybody.

Espionage has been talked about as a potential addition. That said, you can find out all those things you are looking for. Go the diplomacy screen and

-- mouse over relative power(where it says Equivalent or whatever) and you get a rough comparison of your fleet size, naval capacity and technology level

-- mouse over the little mini-portrait at the bottom and see their traits, planet preference, and all their modifiers for the empire's founding species

I think there should be espionage - sending spies into enemy territory, doing missions to slow down their research or kill a leader, would be fun.

How exactly would I, as a three legged squidbeast, infiltrate a top secret lab run by xenophobic sentient ferns?

Asking for a friend.

There are some of those situations for sure, but a couple of the guys my son watches regularly (no clue to their names offhand) make the same mistakes over & over & over. The videos are good, they communicate the experience well ... I'm just not sure they're really great players.

I'll just drop this here. Some of my goals are not things that everyone is going to want to have, but this first version I wouldn't expect to raise many objections. RealTime v 0.1 is now up on the Steam Workshop: Steam Workshop :: RealTime v 0.1. Next up I plan to start getting into some changes relating to the timescale, but I will also be looking into the problem of energy hoarding which, aside from the planetary management AI that I can't touch, is enemy #1 at the moment.

Mod thread with a little more info: https://forum.paradoxplaza.com/forum...altime.929961/


v 0.1 CHANGELOG + REASONING

** Core planets limit set to 30. This is just a placeholder amount until I find a more elegant solution.

** AI will now build planetary capitals before having a full planet of pops. For some reason it was set not to do this. This has a considerable impact on development of new systems.

** AI hoarding tweak -- they will no longer save minerals for things they don't need, freeing them up for other costs. Default is for them to keep 1.4k in the reserve at startup, that amount increasing over time. They'll invest these now when not needed for something specific, aiding their economic growth. Similarly, reduced amount stored up by sectors so they will use what they have.

** AI leaders; they will recruit a bit more governors and a bit less scientists. Was set to always have six scientists, but never more than 2 science ships. In other words, to always have a scientist they couldn't use. They should now move that leader over to governor and administer their planets and sectors a hair better with the extra body.

** Tried to make the AI stop hoarding influence as well. The only thing I was really successful at making them do was to use it on planetary edicts. This is better than not using it at all. I also tweaked a couple things to make them hopefully build more frontier outposts.

** Experience gained by survey scientists is now a third of what it was. They level up ridiculously fast, so this was just an easy quick change. It is probably still too high but should be a lot closer to the right ballpark.

I'll just reiterate...Pay attention to how you treat fallen empires. Yeesh.

Gulp.. I just ran into my first fallen empire... worried to see what happens next.

pain.

Eh, you should be fine if you treat them with respect. If you don't ... well, there's always next galaxy ...

They already don't like me because of xenophobia and some border tension... it's only minus 75 right now. Hopefully, that is not enough to provoke them. Now, that I know they are out there, I'm focusing my expansion efforts in other directions.

Maybe I missed it, but it'd be really nice to let the AI take over the science ships. After a short, bloody war, I had debris fields all over the place and it was annoying as hell having to manually tell the science ships to research a debris field every minute.

Given that we're talking about Paradox, you'll see espionage included in the 1st or 2nd DLC for $14.99.

You didn't miss it sabotai, it's not there :popcorn: That's the kind of thing I hope they add. You can queue up umpteen of them via Shift-Click, but telling the science ships to just 'go research all the debris' would be better.

I googled that "fallen empire" shit.

Damn.

Yeah, I'd love a System-wide research debris button.