Follow on G+ Follow on Twitter Subscribe the Facebook page Subscribe the RSS feed Receive notifications of new posts by email

Stellaris – Release Date and Gameplay Footage

By on March 17th, 2016 2:01 pm

Stallaris | A real-time (pausable) Space Grand Strategy Game by Paradox Development Studio and Paradox Interactive

Paradox Interactive has finally announced the release date for Stellaris, their own foray into the space 4X genre. The date is set for May 9, 2016. The game still holds their typical moniker of a Grand Strategy game and uses their pause-able real-time system. The main difference is that their games focus on more intertwining diplomacy, structured warfare where you need a justification for war and spoils are won in peace deals, and management of internal politics.

This said the game will still hold the trappings of the 4X genre: unlike their other games everyone starts on equal footing and you expand outwards while exploiting new worlds before bumping into other races where the game will slowly adopt a more Grand Strategy format as diplomacy begins to play a larger role. Exploration will also play an important part in the game as well as it will heavily affect research.

Here is the official teaser trailer of the announcement:

List of Features (from the Official Store Page):

  • Anomaly Events – Strange discoveries offer emergent storytelling
  • Deep & Varied Exploration
  • Enormous procedural galaxies, containing thousands of planets
  • Numerous playable species, each with their own traits and engineering styles
  • Vast number of Unique Random Species
  • Advanced Diplomacy system
  • Ship Designer (even civilian ships can be customized)
  • Stunning space visuals
  • Up to 32 players Multiplayer


This is barely scratching the surface, the game will also feature a lot of unique little mechanics like primitive and pre-FTL races you can uplift, observe, or probe if you are into that sort of thing. These races sometimes can become new factions in time (either by your help or on their own). Ancient races on the decline that guard their remnant empires will also populate the map. Existing faction might even splinter into other factions as their society diverges, and not just politically but biologically due to genetic research.

The list goes on, for those interested to know more about the game you can check out our first announcement article which is loaded with information and you can also go see their official blog diaries which explain each mechanics in detail. New developer diaries go up on Monday.

There is also a hefty gameplay footage video (courtesy of PC Gamer):

Stellaris seems to be a very promising title combining the insane level of details the average Paradox game tends to boast, but following a more 4X formula and being set in space thus departing from their historical roots. This said the game is very different from what many space 4X players are used to. This diary on warfare and this diary on empire management are two good examples of how their games tend to differ greatly.

Stellaris is definitely a title to keep an eye on if you’re into Grand Strategy games and wondered what it would be like if set in space or if you’re a space 4X fan that wants to see the genre handled differently. Stellaris will be released on May 9, 2016. It will be available on Steam and the Paradox Store (other retailers yet to be announced).



     Subscribe RSS

Tags: , , , , , , , ,


  1. Mythox says:

    thx for this awesome coverage !!

  2. dayrinni says:

    The dawn of a new age of 4X games is about to begin! I am excited!!

  3. SugeBearX says:

    I’m looking forward to this game…the $150 worth of DLC? Not so much…

  4. Jeff P says:

    Stellaris looks incredible! It reminds me of Distant Worlds with better graphics and UI. I’ve never played a Paradox game, but from the Weekender preview, it sure seems that this is not a game you could play casually over time, the intricacies and myriad details forcing the player to be pretty dedicated (much like Distant Worlds.)

    • dayrinni says:

      Let me tell you what’s going to happen when you play a PI game.

      First off, it’s going to be a very good game outside of the box. You’ll have a lot of fun learning their unique and custom (and complex!) systems.
      Second off, the game will probably crash and have a lot of bugs (some minor, some not so minor). But they’ll fix most of the stability bugs and minor annoyances. For the major game play issues, they’ll fix them in expansions.
      Third off, when you get bored of the game, you can then mod it. There is a HUGE HUGE mod community for PI games and they routinely push out amazing mods for every PI game. These mods range from just simple game play tweaks (ie: fixing issues or switching a few things) to total conversions. It is truly amazing.

      You can realistically get years out of a single PI game – play the base game, then play a whole slew of mods.

      One downside with modding is that the game engine does not usually stand up well to huge expansion TC mods. It becomes very unstable and strange bugs surface. So watch out for that.

      If this game is as good as I think it is, it will be king for the next few years and any future 4X game will have a hill to climb to beat it (only a proper MoO or DW2 could contest). PI does not play around with making bad games and they always do a top notch job. I say this as a both a player and indie developer of a 4X game.

      But either way, it’s a great time to be a gamer for 4X games. Like I said – this will usher in a new age of 4X gaming and that is AWESOME for the past few years has been pretty bad.

      • Jeff P says:

        Thanks for the reply! It is good to hear that at least some game companies still care about the quality of their product. I usually wait until a game has come down in price before buying, so hopefully the bugs will stomped before I take the plunge.

        • dayrinni says:

          You’re welcome.

          PI games are some of the few games I’ll buy right out the gate. Even if the game is not perfect, I know it’ll either get there from PI’s work and the contributions of the community. It’s sorta like the Elder Scroll series. Both developer and modding community do their equal parts.

          Just my 2 cents.

      • jerowen says:

        PI games are indeed for me also games I dare to buy right when they get released. They may be buggy at first, but they always fix them sooner rather than later. The only downside is that PI usually releases large amount of DLC, of which most of them I find a bit overpriced for what they add. Don’t get me wrong, the DLC’s are nice, just not 20 euro/DLC nice. :)

        Also Dayrinni, can you tell me which EU4 mod in particular you like? I’ve never tried modding it (except for the longer timeline one), and would like to try a total conversion mod.


        • dayrinni says:


          I have yet to play EU4 even though I have it. I focus mostly on HoI3 (LOVE THIS GAME) and CK2 (this game is also hilariously funny). I am completely stoked for HoI4 though, do have some concerns with them streamlining the game too much – but we’ll see how it goes.

          I never buy the DLC because the modding community usually does a good enough job (at least for HoI and CK2) to prevent me from being forced to do that – YMMV :)

  5. TimmY says:

    This is it!

    I’m really excited about this and I hope the things that will make this game unique will be accepted by most of the players, even tho I’m concerned about some of the limitations, like controlling the planets and I think the number of ships you can have.
    We shall see. :)

    • dayrinni says:

      As mentioned in my previous post, you can look to the modding community to fix those limitations. Or you can edit the ini files. These types of restrictions are usually defined in there.

      Yes, I am a huge fan of PI games :)

      • TimmY says:

        I know there will be mods. But I don’t want this game to need mods in order to be played.
        I have pretty much all PI games in my steam library but somewhat this is new territory for them.
        I will have to play the game myself to see if it’s more than a total conversion mod for CK2 or EU IV :)

  6. irished says:

    Just watched a twitch tv spot on this and wow I gotta say I’m really hyped for this one now.

  7. tiny_big says:

    I’ve never played a Paradox game and the structured warfare makes me nervous. I’m a fan of the take what I want warmongering style of classic 4x games – not sure whether I’ll like having to set “war goals” ahead of time… Everything else sounds exactly like what a complex empire builder should be though.

  8. hakkarin says:

    This game is releasing earlier than I had expected. This leaves me concerned. I can’t help but feel that this game is nothing but a side-project for Paradox while the real effort is invested into HoI4.

    Plus the game releases only a few days before the new Doom, so I don’t think it will keep my attention for that long.

  9. Mark says:

    I know we often throw around the term “potential” with a tounge-in-cheek chuckle, after being burned so many times, but I really think Stellaris truly has the potential to be the next truly great 4x space game.

    Its just a pity about the complete lack of tactical combat, but meh I didn’t really expect it in a Paradox title anyway. Would have been so nice though.

  10. Gary Vandegrift says:

    I haven’t purchased a new game on release since XCOM: Enemy Unknown. I think Stellaris will be the next one :)

  11. Jeff P says:

    I watched the “MORE” concept video and was blown-away. Foolishly, I contributed to their Kickstarter project; I now believe that game will never see the light of day after serial delays over two years.

    “Predestination” is another game that I contributed to their Kickstarter project; while the game continues to progress, the late-Alpha graphics are crude and the UI is a mess. Sadly, I doubt that the title will ever amount to a playable game.

    I’ve read the Stellaris previews and watched all the “let’s play” videos suggesting that Paradox is developing an exciting and genre-bending new game, and I’m very impressed.

    Have I “drunk the Kool Aid” again or is this the real deal?

    • Gary Vandegrift says:

      Both M.O.R.E. and Predestination are actively being worked on. M.O.R.E.’s latest dev update was on March 11, and Predestination’s latest dev update was on Feb 16. I have been playing the early access for Predestination, and if you look at their Steam page, you’ll find the reviews are 81% positive. Both games will be incredibly late based on their Kickstarter estimated delivery, but I believe we’ll see both titles released, and they’ll be good games. Even some of the most recent AAA titles like Wasteland 2 were at least a year late in delivery.

      In any case, Paradox is not an indie dev, and whether or not you end up liking Stellaris, I’m sure it will be a polished title.

      • Jeff P says:

        Re: Stellaris, I certainly hope you are right. The videos and previews are very impressive. I’m not nearly so sanguine about either MORE (still pre-alpha) or Predestination (and I have tried to play Predestination.)

        • Gary Vandegrift says:

          I guess I’m just more of an optimist than you :)

          Oh, and I really hope Stellaris is all it looks like, because I seriously think I’ll be buying this game on release, something I haven’t done with a game since XCOM: Enemy Unknown in 2012.

    • Wodzu says:

      I strongly believe that M.O.R.E. will be finished. They are constantly working on it, day by day. Lets have some faith ;)

      • SQW says:

        I prefer to put my faith in something more likely to occur. Like the 2nd coming of Christ. Tomorrow. As I’m being hand fed breakfast by Jessica Alba.

        • Vendor-Lazarus says:

          That would be something.
          And maybe we could play Star Citizen at the same time.
          Me in a chase cam and you in a cockpit cruising around in an offline LAN-Party battling Leonardo da Vinci and Nikola Tesla!

        • Gary Vandegrift says:

          Well, Christ coming back is a sure thing. We just don’t know when.

          You being hand fed breakfast by Jessica Alba? Not gonna happen :) If you say it happened, I want to see pictures! LOL!

      • neil says:

        Most likely they will run out of money and have to release a broken mess. I really don’t understand the project management. Massively behind schedule and they waste time on things like ground combat.

  12. Wodzu says:

    I must say that I really like the trailer, music is somewhat similar to the Deus Ex soundtrack.

    And maybe this time (compared to other 4X games), A.I. will be good? :)

    • neil says:

      A.I. in 4X games won’t have any chance of being good until developers start treating it seriously. That means, for big studios, employing a team experts to focus on just the A.I.

      Maybe the buzz around deep learning might encourage some studios to take A.I. more seriously?

    • SQW says:

      The thing that trips current generation of AI are options. Just like how engineers of AlphaGO puts it, the more options available to the AI the more likely it’ll mess up.

      And the current design spec of 4X game is all about cramming more mechanics for players to interact with which will just completely befuddle any game AI.

      Grand strategy games however, hides its limitations better than 4X because interactions are more broad and cause/effect more opaque. Who knows, maybe the key to good game AI is not actually a good AI but a really well disguised one.

      • neil says:

        I’m in total agreement with you on this. Unfortunately most 4X fans whinge like crazy when features are omitted or stream lined. ‘Dumbed down’ – they will put it. They whine even more if the AI ‘cheats’. I don’t personally see what is so smart about abusing a hopeless AI.

        I will say that, if you were to put the AlphaGo team in charge of making the AI for a 4X game, they would still do a much better job than anything seen in any current 4X game, regardless of the games complexity. In my opinion, of course..

      • Wodzu says:

        In my opinion, the things that trips A.I. are: lack of developers skills or not putting enough resources in it. They are mostly focused on graphics because graphics sells in opposite to the A.I. You can’t sell game showing A.I. on a picture..

        MOO2 had decent A.I. (atleast for me) and had plenty of options. MOO2 just had been blessed with skilled developers who knew how to pull it off. And you know how A.I. in MOO or Civilization worked that day? (and I am pretty sure it still works that way for latest Civ) It was hundreds of “IF” statements.

        • dayrinni says:


          Coding up good AI is definitely challenging and takes a lot of work. Then, for every feature or change added, you have to go back to the AI and make sure it works OK.
          It is also hard to develop AI during the actual design and implementation – especially when you are taking feedback from players or trying new features that may or may not work. Games are truly living things and they change a lot. I think maybe developers have good intentions going in but sometimes get to a crunch time where they have to decide what to focus on (ie: graphics or AI).

          In my experience, I’ve typically used a series of flags and if statements to dictate varying behavior. In more complex situations, I’ve used modules which can be chained together.

          My MUD, uses if statements and flags to drive the mob/NPC behavior and it mostly works.

          For the 4X game, it’s a combination of both. For the tactical combat it uses modules. This system is very complex and isn’t perfect (I have some tuning to do). For the galaxy view, it really is a series of if statements like Wodzu mentions. The probabilities of these if statements is driven by the overall tendencies of the empire. IE: if it is more into expansion, maybe it gets +25% to trying to build a colony ship where as the ones who focus more on military get no bonus.
          All of this has taken substantial amount of time and effort on my part for decent results. It isn’t perfect but it isn’t awful. I intend to devote more to this after alpha ends.

          Hopefully, some of this has been informative and interesting to read.


        • neil says:

          MOO2 had weak AI. It was trivial for a remotely competent player to beat, even on highest difficulty. The best thing you could say about it was that there were far worse examples out there. I could well believe it was just hundreds of IF statements, but that is certainly no endorsement of that approach!

          It is generally accepted within the AI community that a machine learning approach is not only desirable, but essential for tackling any complex problem. There have been some very exiting advances in AI in recent years, of which AlphaGo is the most public example.

          If the dream of competant 4X AI is to happen, it must come from utilising this leading edge of AI research, not some old dinosaurs hard coding big webs of nested conditionals.

        • Wodzu says:

          For you it was weak and trivial to beat, for me it was competetive enough to keep me in the game. You probably get bored with the game very fast, it is a shame. I’ve had a fun with it for years :)

          I don’t know what is generally accepted in the A.I. community, for me it absolutely does not matter what is under the hood of algorithm as long as it imitates human enough. That is what is important. If 10000 IFs statements do a better job than machine learning apporach, than I am all up for it.

        • neil says:

          I enjoyed it in single player until I was experienced with the game. After that multiplayer was great fun despite some major flaws/annoyances. Anyway, I was a fairly weak player in multiplayer, yet the AI was no challenge to me. If a weak human player can thrash the AI on hardest setting with ease – then the AI must be weak by any reasonable definition.

          The main point I am making is that the recent successes in AI have come through machine learning techniques. Hard coded rules just don’t work outside of very simple problems. No one would attempt tasks such as image classification based on hard coded rules. Playing 4X games is most definitely not a simple problem. What is ‘under the hood’ is important, because masses of hard coded nested conditionals is a terrible solution. Every 4X game to date has had awful AI.

        • Wodzu says:

          “No one would attempt tasks such as image classification based on hard coded rules.”

          But there are rules. For example if you want to detect face on an image, you need have set of rules for eyes, nose, mouth and so on. This is pretty simple stuff, every camera or smartphone is able to detect face. How do you think it would work without rules? You have set of rules and tolerance for them expressed by neural networks or genetic algorithms.

          “What is ‘under the hood’ is important, because masses of hard coded nested conditionals is a terrible solution.”

          Even for your machine learning A.I. you would still need rules. It seems you are under a great impression that some powerful computer at google beat human in GO. GO which itself has 10 rules. I am not impressed by it, even I wrote 15 yeas ago a software to play GO. It was very weak (it ran on Pentium 133) but I am not google and it took me a week in my spare time.

          IMHO it is nothing wrong with IFs, have you beat Civilization 5 on hardest level? I couldn’t – but I did not play in any specific style recommened on the Internet (like get to this tech and then get to that tech and so on).

          I agree with you that A.I. is awful but it is not because of use of particural technology. It is awful because there is a lot of rules and A.I. does not get enough attention by devs.

        • neil says:

          That is not how modern effective image classification algorithms work. Hard coded rules were abandoned long ago. Machine learning techniques that are successful, such as deep learning, do not simply fine tune parameters on some predefined features.

          Whether you personally think AlphaGo is a big achievement or not is utterly irrelevant. It *is* a big achievement, a very big achievement, to beat the world champion at go. Many thought this would be impossible for decades.

          Comparing your personal go project to AlphaGo is, frankly, just bizarre. Anyone that can program can write a weak game playing algorithm. I wrote a computer chess program as a teenager. I don’t try to compare that to deep blue.

        • Wodzu says:

          I’ve never compared my program to AlphaGo, you clearly misunderstood my comment. I wanted to say that it was matter of time when computer will beat human at GO due to constant increment of CPU power and simplicity of GO in terms of rules.

          I am not impressed by it and I put AlphaGo only little bit higher than Deep Blue. There is too much hype around it. Five years old Watson is much bigger achievement, in my opinion.

          I hope you do realize that I can have my own opinion in that matter, right? ;) You will not change it by using adjectives like “awful”, “weak”, “trivial” in terms of 4X A.I. Neither generalities about “modern machine learning” will help.

          I will not be discussing this topic over and over again, since you are ignoring my questions and trying to be nit-picking just to force your opinion.

          So, have fun with 4X games regardless of A.I. technology which is behind it ;)

        • neil says:

          In fairness you wrote “I am not impressed by it, even I wrote 15 yeas ago a software to play GO.”

          Anyway, whilst Go could have eventually fallen to graph searching based techniques as computer power increases, that would likely have been decades away from now. I don’t think it is fair to belittle the huge achievement of AlphaGo unless you can point to other Go playing programs that come close to it.

          If you view Watson as more impressive, that’s fine. Watson was also based on machine learning! My point is to argue against systems based on lots of hand crafted rules. Watson is not such a system.

          You object to the generality of the statement “modern machine learning”. I mean the techniques that have shown success in the last decade or so. Things such as deep learning, SVMs, boosting, random forests. I have read hundreds of computer vision papers. The only ones advocating using hard coded features and rules for image classification are over 30 years old. Machine learning has proven superior to hard coding for complex problems.

          The only question of yours I didn’t respond to I am aware of is whether I beat Civ 5 on hardest level. I did only once I think, but I grant you it is tough. Hardest level was tough, however, only because the huge bonuses the AI got and victory conditions you couldn’t stop the AI getting due to those bonuses. With a world domination only Civ 5 game I’d be confident of always beating the AI on hardest.

        • Wodzu says:

          They’ve used 1920 CPUs and 280 GPUs for the hardware of AlphaGo! That is just silly, really. I have’t known that, now when I’ve checked it out I am smiling:) There is no way I could provide you other program playing with the same strength cause I don’t have tons of hardware and power plant at my disposal:)

          But you know what, we can run AlphaGo on my laptop and compare it to some other Go software and we will see which would be better ;)

          I am sorry, I can not call this a big achievement, it is just crunching an enormous number of positions due to its supercomputer.

          Anyway, I don’t have a grudge against AlphaGo and I am not saying that hardcoded rules are better than machine learning approach. Of course I would like to play against such A.I. but this will not see the light of day for who knows how many years.

          What I am saying is that is possible to create a decent A.I. for 4X game with hardcoded rules. I am pretty sure you’ve lost quite a few games against such A.I. before you master to find its weaknesses.

  13. Wodzu says:

    “Coding up good AI is definitely challenging and takes a lot of work. Then, for every feature or change added, you have to go back to the AI and make sure it works OK.”

    Looks like, you don’t have unit tests for that ;-)

  14. SQW says:

    What the game AI industry need is something like how the graphic engine biz works right now.

    We desperately need a company that has a base line AI that can also learn from experience and just license it out to gaming studios like Unreal, Cryengine etc.

    The most amazing thing with AlphaGO’s AI is its ability to learn from experience so instead of coding all the possible moves, it just plays itself a hundred million times and remember the best case scenarios. I wish Google has a gaming division. =/

    • Mark says:

      Actually I heard that the DeepMind team were seriously considering “Starcraft” as their next experiment in teaching a machine to beat the best human alive. They wanted a complex game that involves some random elements and *doesn’t* have all the information immediately available (like Chess and GO do) and Starcraft fits the bill. I’ll bet lessons learned from that game could quite easily be applied to a 4x TB game too.

      So you might eventually get your wish, although how long it takes for this research to make its way from Google’s computer science labs to a commercial 4x game is anybody’s guess.

    • SQW says:

      I don’t think 2000 APM will help AI in a TB 4X game. =P

      • Mark says:

        No but the programming architecture necessary to visualize abstract land warfare positions, plan, execute complex strategies and adapt to counter strategies so that all those APM’s actually do something remotely useful might translate quite well to a TB game.

        I dont think the fact that Starcraft is RTS makes it all that different. To an AI, human RTS is just TBS with really fast turns.

        • Wodzu says:

          I’ve read that there is an A.I. “scene” for Starcraft where people implement their own A.I and match them against each other. I wonder how much it had eveloved compared to the original A.I. in Starcraft.

        • Mark says:

          Yes that’s a good question, it would have to be better than the vanilla AI, but how much better?

          Apparently this is why the DeepMind Google team are interested in Starcraft in the first place, because none of the current AI’s can yet defeat the best human players and they want to change that just like they did with GO.

          It seems that their goal is to make sure that machines beat us at absolutely everything. :)

        • SQW says:

          I think SC is a bad choice for AI. The computer has such an over whelming advantage in micro, a merely decent macro AI can easily beat a top human player. In an engagement of equal number of units, the AI is capable of always beating the human player due to the sheer disparity in APM.

          TB game is the only way to show AI can truly adapt and strategise against a human opponent. Of course, I’m more than curious to watch TWO AIs battling it out. =)

        • Mark says:

          “In an engagement of equal number of units, the AI is capable of always beating the human player due to the sheer disparity in APM.”

          Apparently that’s not the case with Starcraft. They say that the depth of strategy is a lot more involved than merely who can click faster as explained in this article.

          That’s the whole reason why the Deepmind team want to tackle Starcraft, because the best human players currently *cant* be beaten by any existing AI.

        • SQW says:

          Well those SC AI mentioned are more of school project level in complexity compared to what google is capable of – inability to learn beyond hard coding which become easy to exploit after a few matches.

          At a small scale. if you have an AI devoted purely to manage blink stalkers (I know, it’s SC2), do you think even the best Korean player can beat it with equal number of stalkers?

          I’m not saying a RTS AI is easy to program, I’m just saying having an RTS AI beat a human is kinda like cheating. The ultimate goal is to create an AI that can mimic human lateral thinking and intuition, not one that can also leverage its APM prowess to give itself a leg-up.

          I’d like to see several AIs fighting each other in CIV or MOO and have researchers monitor their actions and make tweaks and adjustments to eliminate bad plays and see where it leads in a few years. =)

        • Mark says:

          “The ultimate goal is to create an AI that can mimic human lateral thinking and intuition, not one that can also leverage its APM prowess to give itself a leg-up.”

          Oh I agree, in fact I wouldn’t consider that the Deepmind team had succeeded at SC unless the AI could defeat a human expert while artificially throttled down to an APM level representative of a human player.

          If it could *still* beat a human player while restricted to human-level APM’s then that would be a truly great achievement in AI engineering.

Related Articles:

Post category: News & Announcements, Videos