No Man's Sky: the promise of procedural generation vs. the reality

Originally published at:


It amazed me that there was so much hype about what was in people’s heads about what this game would be, vs what was actually going to be in the game.

It was clear from the beginning that the basic architecture of the universe – the generated planets and landscapes – would completely prohibit the enjoyment that is found in Minecraft: that of making and changing.

You can’t build huge monoliths to yourself for others to find, you can’t make things together, you can’t leave your mark on the universe in any way except for meaningless additions to a catalog of species.

You can’t even meet each other, and has been made clear now despite hemming and hawing from the developer.

This should have been clear from the mechanics of the universe, but I think players are just now starting to realize how empty that might make the game.

I grant that there is probably still a lot of fun just gathering materials, and gawking and beautiful landscapes, but I don’t see how that will maintain people’s attention the way games that allow you to make and change the universe do.

The creators are definitely at fault for the over-hype: they prevaricated over whether people could meet each other, and they described a world where you could do anything. But you can’t do anything in a world that you cannot change.

I think it’s a beautiful start, and maybe in five years we’ll have galaxy-wide games like this where you actually can have empires, monuments, farmers, and traders, all created by people, and not by a few AIs put there by the developers just to give you specific things you can do to make you feel like you can do anything.

(NB: I haven’t actually played a real computer game in years, but I’m opinionated nevertheless.)

Edit: The game that was in people’s minds aside, it does look like it can be very cool just to go around and explore:


Looking forward to the No Man’s Sky High amount of butthurt that Game XYZ failed to live up to [unachievable standard].


I’ve managed to avoid the hype trap for No Man’s Sky, having learned my lesson long ago. It even looks like the kind of game I’d really enjoy for a while. Sadly the PC port has proven to be shoddy with no proper translation of controls, crashes, and poor performance. I’ll pick it up when it bargain bins on Steam, assuming they get the PC issues fixed. If they actually had delivered on real multiplayer and put a worthy effort into the PC release, I might even think it worth the $60, but alas, the hype train has derailed.

1 Like

That’s not inherent to procedural generation though, they could (in the it’s not absolutely impossible sense) track the major changes to the world and reflect that in later players views of the world. That’s just a huge amount of data and work for a really small team to add to an already ambitious project

A lot of the confusion comes from how long this was in such a large spotlight so there were a number of features that Hello Games wanted to do at some point but didn’t make it. I’m hopeful for future patches because it sounds like they still have plans to update and upgrade the game, maybe adding some of the features that people were hoping for.


huge penis shaped monoliths no doubt.


Sometimes a monolith is just a monolith.



I’m with you.

This game looks like an incredible technical achievement, especially when you consider how small a team Hello Games is.

But as a value proposition, $60 for a game they’re still patching planned functionality into is unconvincing. I’ll be all over the Greatest Hits™ release, or whatever they call it when and if they put the game out in its final form.


One of the things about getting older is that I see things hyped as “new” that are really decades old. Procedural generated galaxy with thousands of worlds? We had that in Elite in 1984. Although it had its charms, procedural generation then and now can’t beat an actual designed game world.


I should not have bought it. I knew it was going to be disappointing, but … I wanted to see it.

And let me get this straight, the problem of the game is not the procedural generation, although the video game me a great laugh and yes, it is like that, you get to see absolutely bizarre creatures that make absolutely no sense at all.

The problem with the game is that there is not enough game on it.

Every single planet you go you can do the same 6 or 7 things, none of them too complicated, but none of them too rewarding either. Everything is a very mini minigame. Nothing points you toward anything of meaning, except the supposed big quest that sounds just like “get moving”. So you go to a planet, do those few things (talk to an alien, choose 2 or 3 options to get a reward that you probably already got; go to a base and try to get a new blueprint; find a monolith and learn a word or three…) And with that you probably manage to get a better ship component, or a marginally better ship, and … go to the next system to do it all over again. And again. And again. And again. And again.

On the tension between having something that lets you define your own meaning and your own adventure and something so open that you cant do anything meaningful, it is falling on the second part of the spectrum. You can explore infinite worlds that are more or less all the same, to do infinite times the same minipuzzles, to mine stuff to build stuff to mine stuff faster to keep moving to another planet to do the same thing to…


I’m with you there. There is nothing in No Man’s Sky that hasn’t been done before. The only thing new about it is the particular combination of all those features, which is, admittedly, an exciting thing when you’ve wanted to see those features brought together. Even recent games have shared most of No Man’s Sky features without nearly the hype. Starbound has a practically infinite universe with procedurally generated planets and creatures that you can visit, explore, and exploit for resources…as well as destroy and build on. It’s also an indie game. The feature set includes multiplayer and many more things to do on those planets. Of course, it’s a 2D sidescroller, and while quite pretty, doesn’t have the visual awe appeal of No Man’s Sky. That visual awe in combination with procedural generation seems to be the kicker that got so much hype for No Man’s Sky. A lot of us out here are hoping there will be a AAA attempt at the whole space exploration/survival/builder concept. We can see all those wonderful features individually in a myriad of games… we just want it all in one!

Now what I’d have really loved would have been a combination of Kerbal Space Program with No Man’s Sky’s procedural generation. KSP is an amazing experience, and reaching a destination is incredibly rewarding, but after you reach it, there just isn’t much to see or do. KSP’s realistic take on orbital mechanics and space travel is mind blowing and gave me such an incredible respect for everything NASA has done. Taking off and landing on a planet in every other game now just seems silly. Why are you flying straight up?! That’s not how it works!


We just never learned from the Spore debacle.


It seems like nontroversy to me.

I have been interested in the game for quite a few years and casually following its development, yet was not aware of the alleged “hype”. What I read was basically “wander around and explore in a procedurally-generated universe”, which seems to be exactly what they delivered. I actually like that it doesn’t offer/require trade, combat, building, etc because those already exist in so many other games.

Anybody who can’t handle emptiness should probably avoid space exploration.


If you want procedural generation, look no further than Dwarf Fortress!


That’s sort of a false premise though. You’re assuming a given game, or any game would be better, more enjoyable and preferred by most people if it included mine craft style building. But that isn’t the case. Not everything needs that. I personally don’t particularly like that. And as much as plenty of people liked those building elements in (for example) Fallout 4. The some what weird focus on those elements in F4 has been a major point of ciriticism against it, and just as many people seem to dislike those aspects.

The idea of a game focused pretty specifically on exploration as opposed to combat, building, other game mechanics is an entirely valid one. And that can be just as entertaining as a game based on building things. Even the survival mechanics make a kind of sense given their rabid popularity in indy PC games as of late. Though they’re the one thing making me question if I’m really interested here, another mechanic I’m not fond of.

A game (or any piece of media) isn’t a failure because it isn’t a different type of game. Its a failure if it sucks at executing the kind of game it is. Its a bit like arguing that Mario cart is insufficient because the basic architecture prevents the pure joy of amassing armies of units to battle your enemies.

I think the concern about longevity is a valid one. It seems sort of grindy, and reviews have pegged the story and setting as sort of vacant and thin. So I do wonder how long the “fly around poking shit with a stick!” thing can sustain itself. As much as it ticks a number of boxes for me.

I too largely ignored this thing till it launched. I knew something about spaceships and life changing magic. That’s about it. I’ve been burned a few too many time the past few years to pick up anything (outside of The Witcher) at or near launch. Too many borked games. Or developers I like and trust dragging their butts on the carpet. I will probably buy it at some point. Haven’t decided if it will be at full price or wait for a sale. I’d like to throw more money at indy developers, frankly.

That’s increasingly the impression I’m getting. Together with some rather nonsensical aspects to the setting. Like you’re exploring the galaxy, discovering things being the entire purpose of the game. But each planet is defended by the same security drones (cramming combat in where it doesn’t really fit), and contains the same buildings and the same friendly aliens. Sounds like it would get old fast. It doesn’t sound like there’s much to really explore or discover.

1 Like

I was considering finally upgrading to a PS4 to get this game. I am glad I waited. Though even the simple exploration part intrigues me.

I am fine with weird looking animals. It harkens back to the day of weird ass aliens on the cover of sci fi novels and pulps.


As a gamer designer, I want to believe in procedural generation, I think it has great promise - for certain tasks, at least. (There’s some really neat tool sets that use procedural generation systems, for example.) But there’s this fantasy that it can replace human design entirely and the reality doesn’t live up to that. I got brought on to work on a game where the original team had gone to great expense to build their own engine from the ground up to allow for procedural generation of various elements, including level layouts. The problem was, they had so many bad level layouts, they essentially had to create a whitelist of acceptable seeds to use for level generation. On top of which the level generation system created flaws in the environments and the levels just weren’t that interesting (it didn’t help that the components they were made out of weren’t themselves that interesting, either). They would have been better off just hand-crafting the levels, and it would have been less work, ultimately.
The No Man’s Sky demo hand-crafted things on some level, at the very least picking the best seeds for world generation to show off the best the game had to offer.

And here we have the hazard of a very small team creating a game with hugely ambitious procedural generation systems - that becomes the focus, to the detriment of anything else. It didn’t help that there were these fantasies about what it was going to be, but the resources of the team meant that it was always going to be limited to exploration and not much else.

A problem was that a lot of people seemed to think this was that AAA game - even though it’s a small indie team as well. In fact, the team for Starbound is almost exactly the same size as No Man’s Sky’s. Which shows how ambitious and impressive No Man’s Sky really is.

It is, unless you build up a fantasy about what you want the game to be in your head that’s not based on what’s actually been presented. It’s the same problem that games on Kickstarter have.


I think that Roguelikes are a genre that has embraced procedural generation without the downfalls at least. Games like Dungeon Crawl Stone Soup or ToME use it to a degree, but set it in a world framed by human-made lore and gameplay.

For what it’s worth, I find NMS to be pretty interesting and kind of fun, but it definitely has a hype problem and the devs weren’t clear enough on it.

Edit: Sorry, Shuck, didn’t mean to reply to you. I mean, nice comment and all but I clicked the wrong button :stuck_out_tongue:


I am not so sure about that. “Other people are responsible for my runaway fantasies!” sounds like nontroversy, as well.

I strongly disagree, on the basis of living in a procedurally-generated universe. Humans themselves are procedurally generated, so I think the trick is mostly that of using sufficiently robust procedures. Human design has only the illusion of self-direction, but seems neat because it conforms to its own expectations.

1 Like