This Dumb Industry: The Biggest Game Ever

By Shamus
on Aug 16, 2016
Filed under:
Column

So No Man’s Sky is out, and everyone is talking about how “big” it is in terms of playable gamespace. Way back in 2010 I did a video talking about how FUEL was the “Biggest Game Ever” according to the Guinness Book of World Records. (With a qualifying asterisk that it was merely the biggest on any console.) so I guess now is a good time to revisit the topic.


Link (YouTube)

Having said that, talking about the “biggest” or “best” or “smartest” game ever is a ridiculously troublesome and unrewarding task, because you’re just opening yourself up to death by a thousand quibbles.

I made that video because I love talking about procedural worlds. I love talking about how they’re built, how we explore them, how to fill them with interesting content, and the unique rendering challenges they introduce. Which means that for me, that video has officially the Most Annoying Comment Thread on YouTube. Because none of those people want to talk about any of that.

To this day the video is still getting idiotic objections from people who are literally arguing with the title while ignoring the content, timing, and premise of the thing. Usually they do so by typing a one-word comment, which is the name of some game they think will “gotcha” my video. Often, they do this in ALL CAPS. If they’re a really special snowflake, maybe they’ll follow-up with the words “(((mic drop)))”, indicating they believe theirs to be the full and final end of the comment thread, ignoring the several hundred other people who already gave the exact same idiotically wrong answer.

MINECRAFT.

This video was made in January 2010. Minecraft as we know it wasn’t really a thing yet and it wouldn’t have “infinite” terrain for several more months.

DAGGERFALL. nuff said.

It’s called the “largest on any CONSOLE” for a reason, which is that this qualifier rules out stuff like Daggerfall. While interesting in its day, an endless flat plain of copy-paste towns isn’t really addressing the central problem of creating terrain detail. If Daggerfall is our standard of “largest”, then we can easily beat it by creating a gameworld with a single object for scale (say, a post box) and then extendending a flat plane of repeating grass texture all the way out to the maximum values allowed by floating-point numbers.

EVE! ((mic drop))

(Or some other space game.) The most annoying answer. It’s like pointing out that Stig Severinsen held his breath for 22 minutes and then some idiot rolls his eyes, saying that’s not a big deal because sperm whales can do it for 90. Or they claim Olympic sprinter Usain Bolt isn’t very fast because he’s slower than NASCAR legend Dale Earnhardt. You can’t tell if the person is trying to be annoying or if they’re genuinely so stupid they can’t tell that the comparison makes no sense.

I realize “stupidity and ignorance as performance art” is nothing new to YouTube, but this particular thread always annoyed me so much more than the typical, “LOL CONSOLEFAG” stuff, because it exists in place of a discussion I care about. I don’t mind that they insult me (which is just YouTube’s way of saying “hi!”) but I do mind that I have this stupidity instead of an analysis on what games are big and what goals or criteria we might use to decide which games “count” and which ones don’t.

I think this frustration has kept me away from the topic for too long. Since this is a post and not a YouTube video, we should be goodAside from the smart-asses who won’t be able to resist posting YouTube-style nonsense ironically. I know you were already working on your comment in your head while reading this. This ain’t my first rodeo.. So let’s talk about how big NMS is really is.

My original video began with this image, showing how large Oblivion (the game, not the realm) is:

Size is approximate.

Size is approximate.

Just to avoid confusion, the game is called Oblivion but the place depicted on the map is called Cyrodiil. At any rate, I remember that world feeling pretty big when I played it way back in 2006. But then we compare that to World of Warcraft:

Size is REALLY approximate.

Size is REALLY approximate.

Obviously this map doesn’t include the expansion realms. Then again, it’s also mostly ocean. Still, the point stands that “big” Cyrodiil would fit many times within the landmass of pre-expansion World of Warcraft. And now we compare that to FUEL:

Size is actually pretty darn accurate, since it was provided by the developer.

Size is actually pretty darn accurate, since it was provided by the developer.

While the Warcraft map was mostly water, the FUEL map is nearly all content.

At this point it would be natural to show a map of No Man’s Sky and depict how much larger it is than FUEL. We can’t do this for two reasons:

  1. The landmass of No Man’s Sky is broken into many planets and so there is no single map that can show the whole thing.
  2. The two game worlds are so far apart in terms of scale that it would be literally impossible to depict the difference the way we did with the previous maps.

So that’s why we can’t compare the two. But we’re going to do it anyway. In the process, we’re going to need millions of smartphones. I’ll explain that bit later. But first…

Interesting-ness is more important than square kilometers.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

We need to establish our ground rules or we’ll end up in a bunch of pedantic arguments arising from different assumptions. Our goal here isn’t just to make the “biggest” world in terms of cubic volume. As the smart-asses pointed out when they bought up EVE, it’s pretty easy to make a big gameworld set in outer space. Just make a skybox of fixed stars and stick the user’s camera in the center. There! You made a game AS BIG AS THE WHOLE UNIVERSE! Or better yet, make a text adventure where the player has a machine that will whisk them away to alternate dimensions. They can’t explore any of them, but there’s a little readout on the machine that says which numbered dimension they’re in. Now you’ve made a game that’s BILLIONS OF TIMES LARGER THAN THE UNIVERSE!

Our goal isn’t mindlessly marking out vast stretches of empty cubic volume, our goal is to make space that the user will find inherently interesting to explore. That’s obviously a nebulous definition and leaves way too much room for smart-asses to derail the exercise, but there’s no getting around that. “Interesting” is not a fixed threshold. Back in 1996 lots of people found the world of Daggerfall “interesting”, but it doesn’t hold up at all today.

My own measure isn’t in parsecs or square kilometers, but in time. As in, “How long can the average player explore before they feel like they’ve seen enough?” It’s a bit like looking into a kaleidoscope. It’s technically capable of an infinite variety of images, but in practice it’s entertaining for about two minutes. You could make the device more interesting with more colors or shapes, or mess with the mirrors, but ultimately there just aren’t that many minutes of entertainment in viewing random symmetrical patterns of color.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

It’s true that time is a really subjective way to measure explorable gamespace, but I think it’s also a much more meaningful measure. It’s the reason we’re making these worlds in the first place. But not only is time subjective, it’s also hopelessly muddled. The lands of Diablo II are randomly generated, but it wouldn’t make any sense to just take the average number of hours played and claim that’s the value of Diablo II’s gamespace. People don’t dump thousands of hours into that game because they can’t wait to see the next randomized version of the Inner Cloister. They’re playing because they’re engaged with the mechanics.

We can’t agree (in absolute terms) what makes a space interesting, we can’t agree on how long a particular game is worth playing, and we can’t agree on where you draw the line between playing a game for the purposes of exploration and playing a game to engage with the other mechanics.

So while I think time is a better measure of what we’re trying to accomplish when we make procedurally generated gameworlds, I can see why everyone insists on measuring worlds in terms of square footage. We can’t agree on any of those other things, but at the end of the day we can at least look at two different numbers and agree that one of them is much larger than the other.

So fine. For the sake of appeasing the bean-counters, let’s compare No Man’s Sky to FUEL.

First, we half-ass some estimates

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

No Man’s Sky doesn’t actually say how big its planets are. Indeed, the game seems to be afraid of numbers in general. You know this weapon upgrade does “more” damage, but not how much more. You know this ship upgrade makes your ship “faster” but not how much faster. Your shields get “stronger”. Your fuel tank is “bigger”. Your scan range is “increased”. Your jetpack boost is “longer”. How much? Eh. I guess just spend resources on the upgrade and see if you can feel the difference?

This obfuscated approach to numbers extends to distances. You can’t tell how far away something is, only how long it will take you to get there at your current speed. This makes it really hard to work out the size of these worlds. If you’ve ever opened up Google Earth and zoomed from the planetary level to the street-level view, you know how difficult it is to keep track of distances during those intermediate steps. Without reference objects, the view from 1,000 meters is a lot like the view from 3,000 meters. You can’t really get a sense of distance until familiar man-made objects pop into view.

So I can’t say how big the worlds are. They vary quite a bit in size, and it’s not even clear what the size range is. Does the largest planet have four times the diameter than the smallest? Eight times? Ten times? Again, without numbers it becomes very difficult to judge. When you warp into a system and see a planet fill your view, you can’t even tell if it’s a huge planet very far away, or a small planet that’s very close. You can sort of take a rough guess based on how long it will take you to get there, but this is muddled by the fact that your top speed goes up by an unknown amount as you progress through the game.

But having visited about 50 planets so far, I have gotten enough of a sense of things that I can say that the smallest planetThat I’ve seen so far. is still larger than all of FUEL. Let’s be extremely conservative and say the smallest planet is FUEL×1 in terms of area and that the largest is FUEL×3. And let’s further assume that these sizes are evenly distributed, so that the average planet size is FUEL×2.

Get on with it already! How big IS it?

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

As people have been breathlessly saying for months, No Man’s Sky has 18,446,744,073,709,551,616 worlds. That’s a lot of number. That figure might seem sort of pointlessly gargantuan and arbitrary, but it’s actually an important number from a computer science standpoint. That’s 264, which is the highest number you can store in just eight bytes of memoryActually, the highest value PLUS ONE, because computers begin counting at zero..

If you’ve ever played Minecraft you know that worlds have “seed numbers”. That’s a starting value used to feed the random number generator. Each seed gives you a unique world. Note that the values aren’t really related. Seed #5 isn’t similar to seed #6 but radically different from seed #10,000. It’s like asking someone to pick a card from a shuffled deck. The first card isn’t going to be just one off from the secondWell, it COULD be. You get how this works..

So that’s basically how it works in No Man’s Sky. The seed for a planet is an eight-byte value. Give the planet-making code the same seed number and you’ll get the same planet every time. Give it a different seed number and you’ll get something totally different. This is how you get so much output from so little input. NMS takes up 2,806,296,266 bytes on your hard driveAbout 2.6 Gigabytes.. Which means the game can generate 6,573,341,630 distinct planets for every byte of hard drive space the game takes up. It can get away with this because it’s not trying to store all the planets at once. You go somewhere and it just plugs in the appropriate number and gives you the resulting planet.

“But Shamus! How can the game store so many seed numbers? Wouldn’t just the list of planet seeds be larger than most hard drives?”

You’re a clever one. Yes, doing it brute-force like that would indeed be too big. But see, the seeds themselves can be procedurally generated. You can think of it as a nested problem: You feed the galaxy generator a seed number based on which galaxy the player is in. Then you use that galaxy to give you a seed for the star system they’re in. Then you use the star system generator to give you the seed for the planet they’re on. Then you use the planet generator to create the scenery around them. So your save only needs to store that you’re at (plantary coordinates) on (planet number) in (star system number) in (galaxy number). This might not be 100% how No Man’s Sky handles it. It’s just an example of one of many solutions to the problem.

So getting back to our comparison with FUEL…

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

If we were trying to show the map of FUEL in front of the map of No Man’s Sky to give a sense of their diverging scale, then we could create a map where every pixel represented a single planet. That would call for an image 4,294,967,296 pixels on a side. Obviously you can’t fit an image like that on a standard 1080p monitor or television. If you wanted to show a 4,294,967,2962 image and you wanted to be able to display every pixel at once, then you’d need a bank of screens that’s 2,236,962 displays across.

According to this site my LG G2 phone is about 139mm wide. So imagine we went out into the desert and put out a carpet of these phonesBut not mine, because I’m busy using this one and you can’t have it. and set each one to display a small section of our NMS map. You’d have to drive about 310 kilometers to get from one edge of this phone field to the otherThis field won’t actually show the whole image, because phone screens 16:9. So the bottom of the map will be cut off. Don’t worry, those pixels represent all the dumb boring planets nobody cares about.. But you shouldn’t, because these phones are fragile and your car will crush them.

Anyway, so now you have a carpet of phones about ten times the size of Rhode island. All those screens are being used to display a single image. Every pixel represents just a single possible world in No Man’s Sky. But if you go to the very middle of the Smartphone Plains and find the middle-most phone, you can see our comparison map of FUEL.

It’s half a pixel.

Again, this is assuming that the average world size is double the size of FUEL, which was a very conservative estimate.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

The game runs like butts on my machine, which means these screenshots don`t look as good as they ought. I`m fine with these visuals, but I know some people will accuse me of being `unfair` to the game. Sorry. This is the best I can do.

So yes, No Man’s Sky is ridiculously big in terms of area. But to me that’s not nearly as impressive as the actually impossible-to-measure interesting-ness of these worlds. Like I said on the podcast this week, this game is a technological marvel. Not because of its size, but because of the way it merges these coarse numeric inputs and turns them into varied, aesthetically pleasing worlds.

Enjoyed this post? Please share!

Footnotes:

[1] Aside from the smart-asses who won’t be able to resist posting YouTube-style nonsense ironically. I know you were already working on your comment in your head while reading this. This ain’t my first rodeo.

[2] That I’ve seen so far.

[3] Actually, the highest value PLUS ONE, because computers begin counting at zero.

[4] Well, it COULD be. You get how this works.

[5] About 2.6 Gigabytes.

[6] But not mine, because I’m busy using this one and you can’t have it.

[7] This field won’t actually show the whole image, because phone screens 16:9. So the bottom of the map will be cut off. Don’t worry, those pixels represent all the dumb boring planets nobody cares about.



A Hundred!202018We've got 158 comments. But one more probably won't hurt.

From the Archives:

  1. MichaelGC says:

    Yay! Haven’t read the article yet (actually, what’s it about? Aha. Well there’s a surprise! ;D ), but I’m just very glad that there is one.

  2. Grudgeal says:

    So, basically, if I get this right (and I’m sure I don’t):

    FUEL uses a generator to generate terrain as you move towards it, using a set of pre-set seeds to generate a certain terrain as you approach it. So the FUEL gameworld will always be the same each time you start a new game because each area has its own pre-set seed, but the actual terrain isn’t ‘there’ per se until you get close enough for the machine to start generating it using the seed number for that area and the toolbox that converts seed numbers into terrain.

    No Man’s Sky does essentially the same, only with three sets of seeds (galaxies, starsystems, planets) and goes one step further in generating the seeds themselves first, so you generate a new galaxy every new game, a new starsystem each time you move to a new starsystem, and a new planet every time you enter a new starsystem. Each of these layers have their own toolbox for converting seed numbers into their type of terrain.

    In both cases the rest of the gameworld doesnn’t actually ‘exist’ until you, the player, enter an area where the computer has to actualise it for you, and the majority of the games’ data is probably devoted to the toolbox that does the translation. It also means that, statistically, it should be *possible* for a single game to hold two identical starsystems in it where the starsystems and their planets are exactly the same, because the game used a set of the same seed numbers in the same order twice… It’s just very, very, very unlikely.

    That’s rather fascinating.

    • LazerFX says:

      I think it would be more accurate to say that we don’t know how the NMS seed is generated – there’s a possibility that it’s a nested-seed system, but they haven’t said – and without them saying, we shouldn’t really assume. There are numerous ways this could be handled (It’s got an online component that’s required to play, so there could be a server that’s storing everything and generating seeds for you, based upon an unknown-size database that’s “in the cloud” somewhere…)…

      But, with those things taken into account, yes – the terrain doesn’t exist until it’s generated.

      There’s a brilliant example of this called .kkreiger from the Demo Scene (https://en.wikipedia.org/wiki/.kkrieger) which demonstrates how to generate a first person shooter in just 96KB. Yes, less than 96 Kilobytes, or 97,280 bytes.

    • Tsi says:

      You are right. The game world doesn’t exist until it’s processed from the base seed.

      Here is the base seed for N.M.S : 42 (Making it up but I bet that’s it ; ) ).
      This number is stored in every game’s executable and used as a base to generate the same universe for every player. Otherwise If players with different base seeds ever went to the same coordinates to meet, they would see totally different planets and constellations.

      Now, If N.M.S had true multiplayer or coop. Imagine if players with different base seeds met. It would be like living in a parallel universe but being able to see people from other universes crashing into planets in yours but being just fine in theirs.
      So weird.

  3. Liam says:

    So if say 10 million players visit 10000 planets, what are the chances of 2 players having visited the same planet?

    That’s 100 billion planet visits, out of 16 quintillion, so basically zero chance :D

    Perhaps if everyone on earth visited a million planets a day for a year?

    (that’s 7,000,000,000,000,000*365 planet visits a year, or 2.5 quintillion)

    Only 1 in 7 planets has been visited.

    I’m sure birthday paradox stuff will make it more likely etc but I last did a stats course 20 years ago, so meh.

    • Sleeping Dragon says:

      Not only did it happen, it happened on the first day.

      I’m not sure how precisely the planet naming thing works but the way I’ve seen this reported a player has visited a planet that they “noticed has been named by another player already”, they in fact contacted said player and they arranged a meeting… leading to an actual confirmation that the players cannot affect and in fact don’t even see each other in the gamespace.

      I’d link the article but I’m at work where basically every site even vaguely related to vidyagames other than 20sided is blocked.

    • Matt Downie says:

      If there are 18,446,744,073,709,551,616 worlds and 10,000,000 players each visit 10,000 planets, what are the chances of at least 2 players having visited the same planet?

      Almost 100%, believe it or not.

      Consider, at the end of that period, if no overlaps have occurred yet, at this point 100,000,000,000 planets have been visited. The next planet that gets visited, there is a 1 in 184,467,440 chance that this planet will be one that has been visited before. Still unlikely on an individual basis, but if those 10,000,000 players each visit another 20 planets, the chances of at least one finding a visited planet are well over 50%. And if the chances are over 50% for the last 20 planets (each), the cumulative chance for the previous 10,000 planets (each) is going to be a lot higher than that.

      • Matt Downie says:

        The mathematical formula to calculate the chance where P = planets in universe and V = planets visited is:
        P! / ((P-V)! * P^V)

        You can use this formula to calculate the chance of a group of twenty-three people containing at least two people with the same birthday (about 50%).

        But P factorial in this case is such a large number that you probably couldn’t even write it down if you used every molecule in the universe as ink.

        • MrGuy says:

          This assumes that the planets which get visited are selected completely randomly and independently, just as the birthday problem assumes birthdays are randomly distributed throughout the year.

          The solution to the birthday problem is quite different if the dates are biased towards certain values.

          The fact that two players met on the same day is excellent evidence that the PRNG used to generate the seed values are in fact biased towards certain values. Which makes it hard to say anything statistical with any degree of certainty about the game…

          • Matt Downie says:

            True. Note that any kind of bias – like all players starting in the same galaxy, or most successful sportsmen being of the right age to be the biggest in their school year – tends to increase the ‘birthday paradox’ effect (situations where the chances of a match are much higher than the average person would think) even further.

            So while we can’t state with any certainty what the odds are, we know it’s not going to be less than the amount calculated by that method.

          • Tsi says:

            I was thinking that since players have the same objective of going inwards to the center of the galaxy then it would become more likely for them to find an already visited system while systems on the outside of the known universe would be largely unvisited.

        • Zak McKracken says:

          Luckily, a large part of P! cancels out with (P-V)! :
          P! / (P-V)! is equal to (P-V+1)*(P-V+2) … *P
          That’s a lot easier to compute, although I will not try to apply it to the NMS problem now since I still need my computer today …

          Also “P” is a bad choice for a variable name. Players? Planets? Or maybe Probability, as is common in most of stochastics…

          Also also: I’m getting ridiculous numbers with your equation. Either I’m getting it all wrong or you have the wrong equation. Have no time to check, unfortunately.

          * For those wondering about the exclamation marks: they are factorials, i.e. the cumulative product of all numbers up to the one before the exclamation mark. That’s 1*2*3*4*5… and so on until you reach P.

          • Matt Downie says:

            Yeah, I think that equation I found on the internet was wrong. The correct answer is probably one minus something that looks a bit like that.

          • Zak McKracken says:

            okay, I got the calculation all wrong … today’s a bad day for me to do math, and your equation is almost correct. It is actually correct, except it gives the probability for no two people having visited the same planet, the inverse of what we’re actually after.

            The real probability is 1-(your equation)

    • Leonick says:

      That is a question that can’t simply be answered by knowing the number of planets.

      I’ve seen people claim that as people were seeing things discovered by other on the first day then clearly the devs must have lied about the number of planets.

      They fail to account for the fact that we don’t know how the number of planets is distributed over the game world and how the game world is structured or more importantly how large the “spawning” area is.

      Just because there is 18 quintillion planets doesn’t mean you can start on any one of them, starting planets could be and definitely is a limited number of the total.

      • Dork Angel says:

        Since we all start at the edge of the universe, there will be more planets towards the centre than the outside. As we move in the number of planets left between us and the centre will reduce increasing the odds we will hit one someone else has named. At some point you will pass the point where there are more planets behind us than in front of us (note I’m basing this on us stopping at the centre and continuing through it and heading back out again. I guess the real question is if you follow the line to the centre how many systems will you actually hit before you get there and will this path cross anyone else’s path to the centre

  4. Ranneko says:

    “According to this site my LG G2 phone is about 13.9mm wide.”

    …wait what? I think that decimal point is in the wrong place. According to that site your phone is 138.5mm wide. 13.9mm would be smaller than my smartwatch screen. Heck it is smaller than even the narrow watch band option for it.

  5. MichaelGC says:

    You feed the galaxy generator a seed number based on which galaxy the player is in.

    Just curious: do we know how many galaxies there are in NMS? Approximately, I mean – is it, like, seven, or multiple trillion?

    • Ninety-Three says:

      It’s more like nine than 10^9. A precise answer is impossible without spoilers, and it’s impossible to explain why it’s a spoiler without spoilers. So, spoilers.

      When you reach the center of the galaxy, you warp to a new galaxy and do New Game+. If you hit the center of that galaxy, you get a new galaxy, and so on. No one can know if there’s a limit to the number of galaxies, but players have reportedly managed to NG+ about a dozen times and they’re still finding new galaxies.

      If there’s no cap, let’s assume it takes five hours for a fully equipped NG+ player to blitz the center and find the next galaxy. If you were to play No Man’s Sky like this 16 hours per day, every day for ten years, you’d barrel through approximately 10,000 galaxies.

      Of course that’s all assuming these galaxies contain unique planets. As Shamus pointed out, 18 quintillion is precisely 8 bytes, which means there might only be 18 quintillion possible planets the generator spits out, and the new galaxies are just placing the same planets in different places, which no one will ever notice because there’s far too many to catch a repeat

      • MichaelGC says:

        I see, thanks! Yikes, now my head hurts a little. Too many —illions…

      • Leonick says:

        So that’s why everyone started in the same galaxy… Funny how they talked about getting the the center of the universe as a goal but its actually just the center of the galaxy and getting there is a restart button…

        Wonder why they kept talking about 18,446,744,073,709,551,616 (18 quintillion or whatever they said) planets then. Is that the total and there is a defined maybe not all too large set of galaxies they’re all distributed across (maybe with an actual end) or is that the amount per galaxy? If it’s per galaxy they could have claimed (effectively) infinite worlds as they could have used both the galaxy ID combined with the planet ID for the seed.

        Of course, they could get away with having less than a million planet seeds and calling it infinite, no on would notice as they aren’t all that unique anyway…

      • Tsi says:

        Well, I’m not sure what happens (if it’s really a new galaxy or the same with a different start) but in the case it really is a new one, which would defeat the purpose of any multiplayer experience (unless you can later fly from galaxy to galaxy), then the game probably just increments the galaxy seed by the number of times you reached the center.

        Basically, the game generates the first galaxy with some “New Game +X” variable where X = 0 for the first play-through.

    • Narkis says:

      Every player first spawns in the same galaxy, named “Euclid”. We know there are two others, “Hilbert” and “Calypso”, but not how many there are in total.

  6. Ninety-Three says:

    But having visited about 50 planets so far, I have gotten enough of a sense of things that I can say that the smallest planet[2] is still larger than all of FUEL.

    Some internet sleuth did their best to approximate the size of a particular planet and came away with 17 km in diameter (but they made a math error doubling diameter for no reason so it’s really 8.5 km), for a total area of a mere 227 km^2.

    Assuming that’s vaguely representative of NMS planets, you’ve put their planets down as two orders of magnitude bigger than they really are. But “18 quintillion” is such a huge number that two orders of magnitude ends up looking like a rounding error.

  7. Content Consumer says:

    The landmass of No Man’s Sky is broken into many planets and so there is no single map that can show the whole thing.

    But I wonder if we could convince Randall Munroe to take a shot at it…

    Actually, this whole article feels like something he would write. :)

  8. John says:

    Math pedantry ahoy! (I’m sorry.) The average of 1, 2, 3, and 4 is not 2 but 2.5. The average of the first n natural numbers works out to n/2 + 0.5 when n is even.

    This is, incidentally, the reason that, say, a 2D6 weapon is better (loosely speaking) than a 1D12 weapon in D&D. The 2D6 weapon does 2(3.5) = 7 points of damage on average while the 1D12 weapon does just 6.5.

    • Syal says:

      You beat me to it. 2 average is from 0 to 4. Which I suppose we could use if we wanted to, there’s probably some non-planet space in there.

    • Shamus says:

      Doh!

      To avoid a never-ending stream of people posting the same correction again and again, I’ve just changed the largest planets to FUELx3. Doesn’t matter, since that was a hand-wave number anyway.

    • Daemian Lucifer says:

      I hate you now.

    • MrGuy says:

      Meh. I’d argue the main 2D6 weapon is better because mainly it has much lower variance than the 1D12 weapon.

      With a 1D12, all values are equally likely, so you’re as likely to roll a 1 as a 12, making the outcome of each swing highly unpredictable.

      With a 2D6, the total is heavily biased towards the middle of the range. There are 6 ways to roll a 7 but only 1 way to roll a 12 or a 2 (and, of course, no way to roll a 1).

      The 2D6 will roll a 6,7, or 8 almost half (45%) of the time. Sure, you’re less likely to get an unusually high number, but you’re also less likely to get an unusually low number. Knowing how much damage you’re likely to do is very important when deciding whether to attack or run away.

      3D6 (average damage 10.5) is IMO much preferable to 1D20 (also average damage 10.5). Predictable matters.

      • Not to mention the certainty that if you hit, you’re dealing at least 3 damage instead of the person who might just deal 1…when you start factoring in all the DRs D&D has, that 2 damage could be the difference between tickling an enemy and managing to kill them.

        • Decius says:

          Most importantly, if you’re doing a large fraction of your damage with the “2d6” portion of your damage roll rather than the “+47”, you’re at a low enough level that you can’t pick the weapon you want.

        • John says:

          Damage reduction is a tricky business. According to my calculations, a 1D12 weapon is better than a 2D6 weapon for creatures with damage reduction of 5 or more. The difference is quite small–in the ballpark of 0.25 hit points at the most–but real. For high damage reduction cases, the larger variance of the 1D12 weapon is an advantage, since high outcomes are more likely with 1D12 than with 2D6 and only high outcomes do damage.

          For those of an inquisitive bent, my Java code:

          int[] DAMAGE_REDUCTION = {0,1,2,3,4,5,6,7,8,9};
          double[][] expectation = new double[DAMAGE_REDUCTION.length][2];
          for (int i = 0; i < DAMAGE_REDUCTION.length; i++) {
          // Expected damage for a 1D12 weapon
          expectation[i][0] = 0.0;
          for (int j = 1; j <= 12; j++) {
          expectation[i][0]+=Math.max(0,j – DAMAGE_REDUCTION[i]);
          }
          expectation[i][0]/=12.0;
          // Expected damage for a 2D6 weapon
          expectation[i][1] = 0.0;
          for (int j = 1; j <= 6; j++) {
          for (int k = 1; k <= 6; k++) {
          expectation[i][1]+=Math.max(0,j + k – DAMAGE_REDUCTION[i]);
          }
          }
          expectation[i][1]/=36.0;
          }

          Edited to add: My apologies for the formatting. I apparently haven’t guessed the right html tag yet.

          • To bounce off of Decius, by the time you’re fighting something with DR 10/anything, you should either have another weapon or two to bypass it or a caster to avoid it entirely; that’s the reason why Spoony suggested you carry more than one weapon so you’re not trying to fight a skeleton with a longsword.

            I even designed a half-orc Sorcerer around that idea, who would both not dress like a mage but have a light crossbow, a Morningstar, (changing this as I go to avoid stupid ideas) and a sickle, so that way he’s got Bludgeoning, Slashing, and Piercing weapons when the spells run out.

      • djw says:

        The larger variance of the D12 can be an advantage too. If you really need an above average roll to get past damage resistance D12 is better than 2D6 (though you are also more likely to get a really low number on D12).

      • ehlijen says:

        3d6 may be mostly better than d20, but try rolling a nat 20 crit when they only go up to 18 :p

  9. Daemian Lucifer says:

    Aside from the smart-asses who won’t be able to resist posting YouTube-style nonsense ironically. I know you were already working on your comment in your head while reading this. This ain’t my first rodeo.

    Ha!Youre wrong,Im gonna be a smartass about something else!

    Let’s be extremely conservative and say the smallest planet is FUEL×1 in terms of area and that the largest is FUEL×4. And let’s further assume that these sizes are evenly distributed, so that the average planet size is FUEL×2.

    This is wrong,because if the planets are evenly distributed,then the average size of a planet is fuel*2,5.So your whole calculation from here on is completely wrong,and needs to be fixed.

    It’s half a pixel.

    Its actually two fifths of a pixel.

    (((mic drop)))

  10. Daemian Lucifer says:

    “How long can the average player explore before they feel like they’ve seen enough?”

    This is the crucial sentence,because there is no average player.You,for example,have played literally weeks of minecraft,while I have played maaaybe two,and thats a generous guesstimate.Because to me,such a thing gets boring really fast.Yet both you and I have enjoyed plenty of games equally.

    Heck,even if we focus this down a bit,to for example “an average minecraft player”,therell still be a huge divide between various types of builders,streamers,diggers,explorers,….

  11. 4th Dimension says:

    For all their size the NMS planets still massively fails at feeling like an actual set space. It’s primarily because there is no way to properly orient yourself so game might as well randomly generate terrain whenever you step over a limit no matter if you already explored that sector.
    When you are flying over NMS planet it feels like you are flying over completely random terrain generated just there and then. This makes all planets feel small since you will have experienced everything a planet has to offer in minutes. Hell the terrain is so similar that if you tried flying arounf the planet I’m not sure if I could figure out when I completed my task.

    Adding to the confusion of the player the game doesn’t even try to properly display the planet from space and most of the time gives you a random marble that quite likely has nothing in common with the actuall planet. A blue green world looking world can easily be a blasted hellscape.

    Hell they might as well could have been randomizing the map the system and the planets and their terrain whenever the player loads them and most would not notice that since I never got the feel of the space and where I was in it.
    One reason for it is that what NMS has is not a PLANET generator but a BIOME generator. It would have been much better if in the first step the game generated an actual planet LAAARGE scale map of the biomes (continents, oceans, thick vegetation, sparse vegetation, cold/snowy regions etc) and then when the player enters it the actual terrain that belonged to the biome should have been generated. Also when you are generating the biomemap you could probably generate a low LOD model of the planet so I’m not surprised that a blue planet is actually a desert and so I could know “I visited this island, let’s see what is on that continent).

    This makes NMS’s planets MUCH poorer content vise than procedurally generated maps like Fuel one let alone to hand crafted JUUGE ones like let’s say Just Cause 2 (and that was a LAAAARGE map (larger than WoW I think)) one.
    On the other hand the actual goal of the game does not need this level of detail. Planets are there only as stopover points where to resupply your ship for future voyage and you are not supposed to hang around a single planet for long if you mean to get to the center of the galaxy. On teh gripping hand the exploration is MUCH more fun that the ostensibly primary goal.

  12. Falterfire says:

    I really think the correct answer to the question “Which game has the biggest world?” is “The question is wrong”, for all the reasons you describe in your article.

    I think it’s interesting that it feels wrong to classify the massive array of variations on the layout for a given Diablo stage as additional surface area while still classifying every No Man’s Sky planet as additional surface area, because of the difference in how they’re presented within the game.

    There’s no reason you couldn’t recast the Diablo setup as being an infinite series of dungeons oriented on a large map where you just pick one that matches the right level, and the difference would be in the interface rather than the actual generation.

    I guess what I’m trying to get across is that I don’t think the actual size of No Man’s Sky is as important as the impact the appearance of size has on the player. No Man’s Sky needs the player to feel that they are exploring somewhere unexplored, but also that they are exploring somewhere that exists, hence the coordinate system and the more visible seeding making the planet more ‘real’ than just a random convergence of procedural generation.

    • Syal says:

      Part of it with Diablo is that while each area is randomized, they almost always connect in the same order, and there is a clear linearity to it all. If you go back, the layout will be different but the exits will take you to the same places.

    • 4th Dimension says:

      For me the NMS planets don’t really feel like proper planets since there is no sense in where you have landed on a planet. Once you land on it and explore the immediate square kilometer of terrain you have basically seen everything that planet has to offer.

    • WJS says:

      I think that’s pretty important. Diablo may have many different possible layouts for the area, but each time you play the game presents it as the layout of the area. That is why you shouldn’t count it as a larger area, not because of the different gameplay. With Minecraft, you count the size of the world made by one seed, you don’t then multiply this by the number of possible worlds. In NMS, however, the conceit is that you could (given infinite playing time) visit all of the different worlds in one playthrough.

  13. MrGuy says:

    This ain’t my first rodeo.

    That’s irrelevant because this guy has been riding bulls professionally for 34 years, and you’re nowhere near that. ((mic drop))

  14. Wide And Nerdy® says:

    I’m curious if players are allowed to do anything on these worlds that leaves a meaningful impact (blast a crater, build a house, leave a crate). If so, will other players see it? It would be quite a challenge to store all of that.

    • 4th Dimension says:

      Considering that there are mounds of resources that can be mined out and flattened, but if you would look at one that has been mined out from your spaceship while flying above I sincerely doubt that they have enough storage to record every scuff mark the player makes on the maps and the terrain.

      • Wide And Nerdy® says:

        I wouldn’t consider that level of detail necessary. The natural processes of the planet can wear down minor changes like that quickly so its easy enough to cheat there without breaking immersion, especially on planets with higher gravity and thicker atmospheres.

        But if I left a “treasure” would it still be there? If I mined all of a particular resource, would it remember?

        I think there’s a real opportunity there for players to add to the character of the universe if enough information could be tracked.

        • I’ve been trying to leave patterns in the ore deposits for the planets I visit (I attempt a smiley face, but they look a little creepy tbh).

          I doubt anyone would see them even if it does store it, because the planets are so vast there’s only a tiny chance that someone else will visit that star system, that planet, that particular location.

          I would assume they would reset for other users (I’ve not tried going back to see if the places I mined would still be gone, especially since it’s so difficult to really know where you are and you can’t leave any markers and there’s no in-game maps of any sort).

    • Ninety-Three says:

      If you mine a specific iron node or whatever, then that node will stay mined forever, but only on your instance of the game, other players will see the planet as untouched. This has been confirmed by reading the game’s network traffic, it isn’t making any attempt to send the “This iron node got mined out” data to the multiplayer servers.

      It would require a multiplayer server with many terabytes of storage, plus the regular work that goes into making any netcode work, but it’s entirely feasible. The bigger problem is that planets are so big that most players will probably never set foot on the same ground as another player, so it would be a decent amount of work for very rare payoffs.

      • Sunshine says:

        How does it remember that specific iron nodes in specific places on specific planets are mined out, even just for you? I thought all this data would be thrown out when you go to hyperdrive, to free up memory for creating another star system. Is there an ever-growing file of all your minute changes?

        • 4th Dimension says:

          They could be remmebering by recording the locations at which the resources are mined out. All the nodes have their positions, and if during planet/terrain/resource generation they see certain position is mined out it’s easy not to place resources there any more. It would still be a pretty big file, but I don’t think it would be hitting gigabytes.

          On the other hand what they are most likely doing is that they are not recording anything on the disk and they trash all of the info once you leave the system/planet.

          One moment I can check that . . .

          I placed a waypoint on a location on a planet (by hacking one of those nodes that can point you towards locations) went there DID NOT GO ANYWHERE NEAR the waypoint so I don’t clear it. Mined out the Plutonium next to it and a prominent mushroom. Went outside of the flat area of the base and put some holes into some Herbidium.
          Went into space and came back and mined out resources were not there so it does persist things through leaving and reentering planet. Went to another outpost, saved there and came back and the resources were still gone. Went to the space station, came back -> same. Also the plant in the Space station are still mined out so I have to conclude that they do in fact keep a track of things you mined out at least while you are in system. I have not gone out of the first system yet so I can not confirm. Also I went into my savegame folder and NMS is currently using about ~3MB for my save.

        • Ninety-Three says:

          I can’t speak to the technical details of how NMS does it, but the standard approach is to save your changes as a delta, a simple list that says “Load the planet exactly according to the default procgen, except the player mined an iron node at 106.0254,154.127 and cut down a tree at 106.0261,154.129, so don’t load those two objects, and the player threw a grenade at 106.0263,154.120 so put a crater there.”

  15. Mephane says:

    I had to scroll up again to check that this post was not actually written by Randall Munroe. Hint: that was a compliment. :)

    • baud001 says:

      That’s not the same style, not enough jokes, stick drawing and weird assumptions (at least compared to his ‘what if’ series).

      But I agree with you that the article is great, even if it’s a little bit too defensive.

  16. Ciennas says:

    So I’m reading your article, and O can’t help but feel that Youtube traumatized you a little bit. Every picture in the article has the same caption about how this is the best your machine can do.

    Normally, you put some additional commentary about the game, or some kind of witty joke. Here you’ve used the same comment throughout.

    We’re not the Youtube comment system. At least I’m not. I’m reasonably sure about the rest of the commenters here too.

    I always liked to read those captions.

  17. Mersadeon says:

    I haven’t looked into NMS much (mostly because I didn’t want to end up disappointed), but from what I can tell, the planets look reasonably interesting – but I just can’t shake the feeling of them being higher-res, upscaled SPORE planets.

  18. Zak McKracken says:

    If you want to display that map properly, I suggest you upgrade to a “better” phone.

    Personally, I feel like the G2 has already way more resolution than I’ll ever appreciate (I blame my eyes) but with the Sony Xperia Z5 Premium, you get twice as many pixels per edge length (i.e. 4 times as many per unit area). If we put together a display of those suckers, we’d get about a billion pixels per square meter, which means we’d need 18.5 billion square meters. That’d be a circle with 76.8km radius, or 95 miles diameter (or a square of 170 miles edge length).

    However, if we went from that to a decent printer with 1200ppi (that’s planets per inch — good printers have more resolution which means we could even put some detail in!), that would then shrink to … 51km radius. So if you’re careful you could get to the galactic center with just a little more than a marathon, on feet!

  19. Zak McKracken says:

    What I do wonder about (and what I believe they need the servers for) is keeping a tally of all the changes wrought on the world by players. All names, killed animals, blasted asteroids harvested resources …
    I suppose there will be an increasing amount of storage on the servers dedicated to these things. But I also expect that they’re reasonably efficiently stored, and there’s an upper limit to how much a single player (or the entire community of players can change the universe per time, which I suppose is huge compared to what a single PC can store but easily handled by adding on hard drives to the servers as time progresses. And any single player will only ever see a very tiny fraction of these changes.

  20. Josh says:

    BUT SHAMUS, MINECRAFT ALSO HAS 2^64 POSSIBLE SEEDS AND EACH MINECRAFT WORLD IS 102,400,000,000 KM^2 SO THE SIZE OF ALL POSSIBLE MINECRAFT WORLDS IS BIGGER THAN ALL POSSIBLE NO MAN’S SKY PLANETS HAH I WIN YOU LOSE ((MIC DROP))

    • Paul Spooner says:

      I knew there was a reason I liked Minecraft.

    • djw says:

      To many words. You sound to educated to use the mic drop, even with the caps.

    • Daemian Lucifer says:

      Shouldnt that be cubed,since in minecraft you can dig,and in no man sky you are just surface hopping?

      • Ninety-Three says:

        Minecraft has both a ceiling and a floor, and they’re measured in meters (technically you go infinitely high/low [or at least infinite within the bounds of the variable storing your Y-axis data] if you have a modded jetpack or something, but you can only build in a very small range). Measuring by volume would favour NMS.

      • 4th Dimension says:

        You can technically excavate terrain by deforming it with plasma grenades. It is on the other hand not efficient way of getting anywhere and you will most likely use the feature to remove some terrain you keep getting stuck on, for digging yourself out of a cave in which you are lost or to create rest points on a cliff face where you can recharge your jetpack.

    • Zak McKracken says:

      Ah, so you’re taking the combined size of all possible Minecraft worlds? In that case I guess you’d need to compare to all possible NMS universes.

      The size you give for a single Minecraft map is 202 times the surface area of earth, or a bit less than twice the surface area of Jupiter, so it’s actually possible that at least some (real) planets are larger than that, although I don’t know about NMS planets.

      Assuming for the sake of argument that NMS planets have about 1/10 of the earth’s surface, that means that 2000 different random seeds for the NMS universe will do what 2^64 achieve in Minecraft.

      I’m not quite sure how many universes NMS has gone through so far but players of the leaked pre-release game have already seen a different one than the current, and I’d be a bit surprised if there weren’t more of these in the future.

      …in terms of content per square meter, Minecraft also loses, big-time. The whole landscape is resolved with 1 meter resolution, with a small number of materials or objects filling those “voxels” — so if you take resolution into account, the one NMS universe should already be close to any Minecraft universe.

    • Philadelphus says:

      And don’t forget The Nether and The End, which provide additional space for each world. Though you could argue they’re separate worlds since they’re technically different dimensions, but each one is linked to the same seed as the Overworld.

  21. Mephane says:

    I find it always odd that Shamus never mentions Elite Dangerous in this context; afaik he has never played it, fair enough, but it contains a full-scale replica of the Milky Way, procedurally filling in the blanks.

    • PinkCoder says:

      Agreed. I would love to see an analysis of where Elite Dangerous falls into this comparison.

      It may not have as many total possible planets, but I love how it is scientifically more accurate. Also, much more visually appealing in my opinion.

  22. Jabrwock says:

    While it’s very impressive to me that you can get a pretty much identical world by handing it the same seed, which makes the planet list massive for a small footprint, for me it comes down to “is exploring going to net me anything new, novel or interesting?”

    When Minecraft first came out, the idea of just continuously exploring was pretty cool. I sunk WAY too much time into spelunking, carving out little way stations along to the way to have safe places to sleep and recuperate, or prepare for the next outing.

    But after a while, it became same-y. Yeah it was a new mine or cave, but it wasn’t unique enough from the previous one to drive me to explore it.

    So while NMS looks impressive, I can see (for me) the exploration factor netting diminishing returns for me as time goes on. There needs to be a reason to continue exploring (after the initial exploration high wears off) beyond just “oh cool a new place I wonder what’s there.”

    • Zak McKracken says:

      I’m not on NMS (yet) so I don’t know about he gameplay but I always expected them to do it in a similar way to Minecraft: First you publish the bare-bones world with a couple of things to do, then you introduce more things to do, watch what the players do and ask for, have more ideas, and eventually there will be completely different types of games within the NMS world.

      I’m kinda hoping that eventually there’ll be a multiplayer mode where you start together with a bunch of other people and build up a base / defend your planet(s) / create a trade empire / form a military alliance … I could imagine the NMS universe being the backdrop for an EVE-like strategy game, economy/ecology simulation, PvP or PvE action shooter, or any form of second-world-y social gaming (even MMOs, if you must). Some of these things could happen in the same persistent universe, others independent … the sheer possibilities!

      Here’s hoping they get a grip on the current technical issues and then manage to provide new and interesting updates quicker than players tire of them…

      • Sunshine says:

        A blog post of the NMS site (I think) says that, more or less. (Also, I think it was a point that Rutskarn made on the Diecast about nerds becoming enthralled by new features in what they’re making, at the cost of ever making a finished product.) There are all sorts of ideas that they put in the “It would be cool if you could…” box until they had the core of the game ready to ship, like base building and player-owned capital ships. So once the technical issues are dealt with, they’ll probably be on those.

  23. Geebs says:

    I guess the way to get around the nonsense size comparisons is to work it out as total size / spatial resolution; so a 4 X 4 km space at 10 m per unit is 1.6 kevins (1 kevin = 1 km2 m-1) and a 2 X 2 m space at 1 mm per unit is 4.0 kevins.

    As an aside, as someone who has played a few gigs in my time, (((mic drops))) make me wince. Those things are bloody expensive.

  24. Adeon says:

    Really pedantic comment: the Inner Cloister was one of the levels in Diablo 2 that was NOT randomly generated, it was the same layout each time.

  25. Jokerman says:

    *makes joke about you posting lower quality screenshots and being unfair to the game*

  26. Ramsus says:

    Youtube Hi. [Please Insert Your Own Insult Here]

  27. Phantos says:

    To this day the video is still getting idiotic objections from people who are literally arguing with the title while ignoring the content, timing, and premise of the thing. Usually they do so by typing a one-word comment, which is the name of some game they think will “gotcha” my video. Often, they do this in ALL CAPS. If they’re a really special snowflake, maybe they’ll follow-up with the words “(((mic drop)))”, indicating they believe theirs to be the full and final end of the comment thread, ignoring the several hundred other people who already gave the exact same idiotically wrong answer.

    UHH, YOU FORGOT TEH “J” IN IDIOTIC

    (Youtube comments really are remarkable. People never cease to find new depths in being incoherently wrong.)

  28. Retsam says:

    You think that’s the Most Annoying Comment Thread on YouTube?

    Political videos. ((micdrop))

    It’s called the “largest on any CONSOLE” for a reason, which is that this qualifier rules out stuff like Daggerfall.

    Yeah, but you don’t mention that in the title, and it’s really not fair of you to expect people to actually watch a video before leaving a comment that totally disproves it.

  29. Steve C says:

    The mouseover text reads like butts on my machine, which means these snippets aren’t as interesting as they ought. I’m fine with the repetition, but I know some people will accuse me of being ‘unfair’ for pointing it out. Sorry. I’m a dick and couldn’t help myself.

    (Yes I read the disclosure above. This was supposed to be a reply to that. But internetting is hard.)

  30. Aaron says:

    well i was gonna wait until the price went down a bit, but this article caught me at the right time (fallout 4 isnt really interesting anymore and nothing else jumping at me) and shamus described it like the best ice cream you could ever possibly want in the desert…thus i had to buy it before it melted?

  31. Scott Blias wrote an interesting paper a while ago about using single-precision floats for the coordinate system in a game with a very large world, (Dungeon Siege) where float precision becomes a big problem: http://scottbilas.com/files/2003/gdc_san_jose/continuous_world_paper.html

    • 4th Dimension says:

      That same problem was one of the problems that had to be solved by KSP folks. Only in their case it was much harder since they needed to support LAAARGE range of scales from centimeter/milimeter precision needed for spaceships to star system scale where distances are measured in thousands of kilometers. And you need a seamless transition between them since you need to be able to land on planets where the terrain will be detailed and traverse astronomical distances between planets.

      I think this video goes into the detail of challenges involved. https://www.youtube.com/watch?v=mXTxQko-JH0

  32. Nessus says:

    What I’d like to see is this kind of tech used to create enemies and/or NPCs instead of giant worlds. As in, each NPC has the procedural complexity of an entire planet in NMS (critters and all).

    Imagine a version of Dark Souls in which every single enemy is as visually unique as the bosses. No more 10,000 identical skellies or low-level knights. Every. Single. Enemy. Or, if you like, enemies that are of distinct classes, but which are individuals: every one of those 1000 knights you kill throughout the game has completely unique armor, implying each was an individual noble with a backstory. Every one of those 1000 purple-headed cane choads is unique: maybe not enough to stand out, but enough to eliminate the clone-stamp awareness.

    Imagine an Elder Scrolls (or Fallout, Saints Row, GTA, Watch Dogs, Assassin’s Creed: take your pick) game where the NPC’s actually look like different people, rather than the same 3 face maps stretched over 1000 morphs of the same head geometry. Now apply the same to the body types and clothing.

    We already have facegen style systems, but they don’t really work as well as they’re supposed to, as they have too little granularity and variety. You still end up with worlds full of NPC’s that look like clones despite each being slightly different, because the differences are very shallow. Crunch the same level of procedural granularity a planet has in NMS down into a single NPC, and you could start having a world that felt populated by actual people, or monsters that felt like a living species instead of just game tokens.

    Keep the world small (for a given value of “small”) and deliberately designed, but use the breakthrough high-efficiency procedural algorithms to give that world the kind of fractal detail that isn’t possible with the more standard systems.

    Maybe use it to generate procedural shaders and height maps for a fractal LOD effect, so you don’t need bigger textures to make the ground right under you (or on the opposite hillside) look as good as the ground 4 meters away. Kill the scourge of tiling once and for all.

    I feel like spreading this level of granularity out across an entire universe of star systems is what makes it look less impressive on screen than it technically is. If you were to use it strategically on a more compressed scale, it could be a real sea-change, but by spreading it out so much, you spread it thin, so the bits you actually see up close only have the granularity of a common facegen system.

    • Ciennas says:

      That would be so cool, but could you do it with this system?

      Planets can look any way you like, but we are very very very good at determining people shapes. Not to mention clipping.

      Maybe instead of a galaxy, you make a planet. With procedurally generated cities and dungeons and caves.

      And actually to scale, an earth sized planet.

    • Geebs says:

      You might start to see more of that as DX12 and Vulkan adoption becomes more common. Most of the stuff they’re supposed to be able to do in terms of increased efficiency should be taken with a huge grain of salt, but the thing they’re actually good at is reducing the CPU cost of “binding” a new buffer – like, for example, a new character model. That would offset some of the cost of switching NPC models vs just drawing the old one with a bunch of different animation states.

      I don’t know for certain, but it’s highly likely that NMS gets a lot of mileage out of “instancing” – taking one model and drawing it a bunch of times each frame. That gets a lot of detail of of a small number of draw calls, and saves on the bandwidth cost of sending new geometry to the GPU.

      To a certain extent though, the scale of the world still doesn’t matter so much as the number of polygons on the screen at the same time, and the number of models and textures being shuttled to the GPU.

      • WJS says:

        Haven’t we kinda already hit “peak polygon” by this point? Modern graphics cards are so powerful that poly counts are rarely a bottleneck any more, and individual polygons haven’t been visible for a while on high-end games – why bother with more of them?

    • Phantos says:

      For some reason I could see that happening in a game where you fight robots, kind of like how a lot of Wall*E’s lesser robot characters are mixed and matched pieces.

      But with organic stuff, or creatures that aren’t just interchangeable parts or very simple geometric shapes, I imagine it would start to get trickier to randomize that stuff and still look good.

      Although it would probably be a goldmine for horrifyingly stupid glitched-out attempts at generating humanoid enemies. “Leg-Face, The Distorted Meme”.

      • Nessus says:

        What you’re describing is actually what lots of games already do (including NMS, which is part of why NMS doesn’t look as good as it was hyped to), and exactly what I’m talking about replacing/avoiding.

        If you walk down a street in Saints Row III, all the people you walk past are made of random combos of clothing and hair meshes, randomly textured, and with random body/face morphs applied. There are constraints applied to the random generator to make sure that things don’t get too uncoordinated, but it’s basically that. Same with Skyrim, and pretty much any game in the past 10 years that deals with lots of non-uniform NPCs. IIRC this kind of tech debuted with Half-Life 2 (though being the very early example, it’s so minimal you might not notice it). Other games do this with environments. Warframe, for example, creates sprawling indoor complexes lego-style out of a library of prefab rooms, then populates them with a library of containers and other prefab clutter meshes.

        … And of course, NMS does this with creature anatomy (and plants, and rocks), which is why it doesn’t look on screen like the technical leap it is under the hood. It’s supposed to have this fancy-pantsu new procedural tech under the hood, but the results just look like SR-III’s crowd gen system applied to critter parts instead of clothes, because that generating power is being spread across the entire environment instead of focused on the critter. That critter’s entire species is getting, proportionally speaking, only the amount of attention that a single NPC’s hairstyle gets in SR-III.

        What I’m talking about is more like taking the way fractal terrain generators work, and applying that to the topography of a creature or a human head instead of a landscape or planet.

        Basically, I’m picturing something like this:
        Every NPC or creature starts out as the same metaball-like blob. Then the generator adds a rig, which can be made of a library of pre-built components, or procedurally generated according to certain rules (the constraints don’t have to be a check-and-correct thing, they can be built into the seed or algorithm itself). Then you add procedural hightmaps to sculpt the surface (using the rig as one of the seed values), and layer in procedural shaders atop that (using the heightmap as one of the seed values).

        The thing to remember is that “random” doesn’t mean “EVERYTHING is random”. In fact, in a procedural engine, very little is random apart from the seed: the seed “grows” according to the logic of the algorithm, so you can actually control almost everything by changing the algorithm.

        Historically, doing this to create creatures or NPCs would have several big problems:

        1- The procedural generation would be very CPU expensive, especially with multiple people/creatures on screen. This is what I’m hoping the new super algorithms could alleviate, but even that likely isn’t enough to allow for crowds.

        2- You’d have to have a good way of either creating procedural animations, or modifying pre-built animations on the fly, physics engine style. I suspect the former is impossible with current tech, but the latter seems like a more complex extension of some things that are already done in games today (like modifying the foot and leg placements of an animation according to the level of the ground). Either way, it’ll add to your CPU-debt (the former much more than the latter).

        3- Some of these things would take a huge amount of R&D to create a mathematical model and development pipeline for. Once that was done, others could build on it as cheaply as with current tech, but someone would have to make that initial investment, and it could very, very easily be too expensive for a single game production. More like something that’d have to be done as a pure research project, or as part of developing a new licensed engine.

        4- Current GPU architectures are all designed around the common linear texture and mesh methods, so a system like this wold take a big efficiency hit on current cards. You’d need a new gen of GPUs to make the most of this method, not because current tech isn’t advanced enough, but because with current architectures you’d be sending an apple to do an orange’s job.

        The total is something that probably will have to wait at least for the next gen of hardware to come to fruition, but it’s something I’ve been thinking about for years, and the promise of new maths that could make this less CPU intensive makes me more hopeful of seeing it sooner rather than later.

        • WJS says:

          Didn’t Force Unleashed have procedural animations? I’m sure I saw a tech video about it at one point. I would hardly say that a game from eight years ago is “beyond our capabilities”, assuming I’m remembering it right.
          EDIT: Depending on how strict you are with the definition, you could easily call head-tracking in Quake Procedural animation.

        • WJS says:

          Also, is there in fact any evidence that there is “new super-algorithms” “under the hood”? From this post, it sounds a lot like “You can’t see it, but it’s there! Honest!”, which is not convincing, especially coming from Hello Games.

    • Felblood says:

      Warframe has been taking baby steps in that direction for a few years.

      A some of the newer enemy types have multiple versions of each body part, and are assembled like Lego men, from random bits from the list of possible parts for that enemy type.

      Too bad the individual parts seem to have come from a design school run by Michael Bay and Games workshop, so all the bits are buried in so many greebles and spikes, that you can’t tell the new guys apart.

      There’s one of those new guys over here!

      Is it the one with [weird special ability]?

      I’m not sure. He has a spiky halo on the back of his helmet.

      Is it a complete ring, or is the top part open?

      I shot him to death. Wait here comes a different one! Wow he’s got a lot of health!

      Does he have a rifle or a heavy rocket launcher.

      He’s got this spiky thing. He holds it like a roc- IT IS DEFINITELY A ROCKET LAUNCHER!

      Omw. Saryn, can you cover my side for a minute? Use your 4.*

      *Why yes, I am a disgruntled Saryn player. How could you tell?

  33. Abnaxis says:

    Indeed, the game seems to be afraid of numbers in general.

    Dammit, now I want absolutely nothing to do with this game. I HATE it when games refuse to give the information I need to decide pros/cons of the equipment I am using.

  34. I generally enjoy exploration style games, but I’m reluctant to pick up No Man’s Sky because I also generally want some kind of purpose to pursue. The size of the game is irrelevant to me if it’s BORING, and that’s basically impossible to quantify.

  35. Zaxares says:

    This game sounds very technologically impressive (that is, the technology behind the scenes that enables it all is amazing), but I’ve heard that the game itself has been meeting rather lackluster first impressions from gamers. Is there anybody here who’s played NMS and can give more insight into the situation?

    Speaking for myself, I really can’t play games like FUEL or NMS or Minecraft. I’m the type of gamer who likes to go everywhere and do everything. Complete every side-quest. Talk to every named NPC. Collect all the most powerful items and gear my party with enough WMDs to make the final boss weep in terror. And I simply can’t do that in games like NMS where the sheer scope of the things you can do is enough to keep you playing for the rest of your life. (Even games like Oblivion and Skyrim are pushing it.)

    So I’m content to just watch from the sidelines and marvel at what game designers can produce, even if it’s not a game I’d enjoy playing myself.

  36. Joakim says:

    This discussion of how big games can be reminds me of a mathematical insight I came to, though given the discussions going on others have come to similar insights using other games:

    In solitaire, the precise order the cards are stacked determines how the experience will play out. A deck of cards can be stacked in 52 factorial different ways.

    Of course, that’s not fair, since if ace of hearts and four of spades are card #1 and #2 on top of the drawing deck that won’t make a large difference as if they would be in the other order, i.e. four of spades before ace of hearts. The precise order in the drawing deck will not make a difference, so one must divide by the amount of orders these 24 cards can be in, 24 factorial.

    That leaves us with ~1.3 * 10^44 possible starting boards, or game worlds.

    Since such large numbers are hard to wrap your head around: the precise age of the universe is unknown, but most estimates fall shy of 20 billion years. 20 billion years corresponds to ~6.3*10^17 seconds.

    So if one imagines that a supernatural being starting playing through all starting sets of solitaire at the Big Bang and each game takes one second, he or she would have played less then 4.85 * 10^-27 of all possible starting set ups at the present.

    Yet I am tired of the game solitaire, despite there still being games of solitaires that have yet to be played.

    • Syal says:

      The precise order in the drawing deck will not make a difference,

      That’s assuming you’re only drawing one card at a time. If you’re drawing three at a time those go back in.

      Although the number’s a bit lower for things like “ace of hearts on top of ace of [anything]/ 2 of hearts”; there’s eleven available slots for cards to move so some combinations aren’t very unique.

  37. The Seed Bismuth says:

    [comment retracted]

  38. MadTinkerer says:

    Oh please, I bet if you picked some random Indie game with procedurally generated levels and added up all the real estate it generated since launch, I bet it would be bigger than WoW or FUEL.

    So basically:

    Good Robot ((mic drop))

  39. Pyrrhic Gades says:

    Hold the mic…

    Diablo II was procedurally generated? Never noticed that since I only played it at internet cafes, with multiple week breaks inbetween (although everytime I did play it, I started a new game). I just assumed that I had a poor memory for map layouts.;

    • Daemian Lucifer says:

      Even the first one was procedurally generated.And in fact,it went far beyond the boring blandness of many modern procedurally generated games,because you would never encounter all of the monsters in a single playthrough.Thats why restarting if you got a tough enemy combination was a thing back then.

Leave a Reply

Comments are moderated and may not be posted immediately. Required fields are marked *

*
*

Thanks for joining the discussion. Be nice, don't post angry, and enjoy yourself. This is supposed to be fun.

You can enclose spoilers in <strike> tags like so:
<strike>Darth Vader is Luke's father!</strike>

You can make things italics like this:
Can you imagine having Darth Vader as your <i>father</i>?

You can make things bold like this:
I'm <b>very</b> glad Darth Vader isn't my father.

You can make links like this:
I'm reading about <a href="http://en.wikipedia.org/wiki/Darth_Vader">Darth Vader</a> on Wikipedia!

You can quote someone like this:
Darth Vader said <blockquote>Luke, I am your father.</blockquote>