Procedural Laziness

peep game son

peep da detail dat shine fam

nice job retard

Procedural generation reminds me menstrual blood paintings. You shit a lot of blood on a white picture and call those random blood splats an art.

Is the image supposed to invert colors?

It doesn't for me….

Procedural generation to complement things that are already premade to appear under consistent circumstances is entirely different from what NMS is doing and what devs who made that want to do (full procedural generation). Even though D2 has procedural generation there certain locations or areas that appear in the same position when attached to the level or always show up on the same level so you know exactly what to expect.

Those types of games are the only types where procedural generation generally works, Roguelikes work because you use your imagination the majoirty of the time.

lel no, procedurally generated games boils down to a very samey and average looking environment since it's all simplified and boiled down to very basic layouts and they never really change thus ensuring it all looks the same. Hand crafted the devs can just fuck around and make things work out in a way procedural generation can't. This is because what's generated is based on an algorithm, and the less you put in the more samey and predictable it all is so you'd need to pump in a lot of shit out for assets and how they'd go together but because it's algorithmically created it'll still have obvious patterns that players will notice.

When you craft it by hand you can get bored and fuck around, or even know when the most boring part is so you can just place a trap just to fuck with the player.

Procedural Generation is only okay when its used as a core game mechanic OR when it is presented in a way that is completely transparent (eg. Minecraft)

However, when a map is procedurally generated and then compiled into the working main map of a game that has nothing to do with procedural generation as a mechanic, then it is complete laziness.

It can work but you have to spend years to polish it. DF does it well. Heck, even Spore isn't so bad. No Man's Lie bit way too much and ended up leaving a boring and empty game. Spore might be a disappointment but it's still fun. NMS is dull and shite. I think you have to remember gameplay when you do procedural generation.

not true. procedural generation has a low floor, roll the dice and shit stuff on the map, but an extremely high ceiling.
a well made procedural system is an art piece in itself. that's why they're so damn rare.

i ain't play dis no man sky mess but it seem like dey jus made da assets n procedural generation shit n dey ain't bother w/ nothin else.

like dat shit can b cool if dey got sum ole handcrafted unique pockets that would warrant exploration every once in awhile but slime came out w/ jus dat walk aroun look at shit type bull

das ryt

it's an NBC conspiracy, they're tapping into the game market

Is it really?

How are you going to do galaxy simulation then? Or should, say, Space Engine be lit on fire and the only kind of space sims allowed are X-Wing clones?

Are you clinically retarded fam?

I want to give them the benefit of the doubt that they just needed more time but even then they also desperately needed to level with the public and dial back the hype. That retard would keep going into interviews refusing to confirm or deny anything, then is somehow surprised when people's imaginations go wild only to become angry when the game doesn't deliver.
Alternatively selling it for $20 would go immense lengths towards tempering expectations. People don't expect much from something they didn't pay a lot for

Reminder that there are people here who weren’t alive when ✡Seinfeld✡ was on the air.

...

It still is in sped-up reruns on cable. DVD and tapes of the original broadcast show the real thing, though.

It's too bad they couldn't procedurally generate some fun.

reminder that the berlin interpretation is a bunch of self-contradictory pretentious bullshit.

but look it's fine because it allows any and all games to be roguelikes!
It's hilarious because it's the most self defeating definition ever

Procedural generation is the future in the kind of game where the level desing doesnt matter

Hopefully never. DF is the best work in the history of the medium.

Until they become obscure again.

This analogy does not work.
Procedural generation would have more to do with the interactions of the bacteria in her menstrual discharge.

Procgen gives videogames a soul. Too bad so many videogame developers think that a lifeless boulder will behave any differently if you give it one.

I have plans to start development of a proper procgen survival game, but unfortunately I'm going to have to wait until the market desaturates on that front. My ballpark is that'll have to be at least 5 or so years.

Very unfortunate.

I'd also like to point out that Procedural Generation and Random Generation are often confused, No Man Sky did not use Procedural Generation except on the planet height maps, that's all it used it on, Everything else was built from randomly sticking multiple models together- this is why none of the actual creatures in the game are persistent. The reason they used Procedural Heightmaps is because it allows you to dynamically alter the heightmap at runtime. It actually does have some very cool uses, but developers under use it, similar goes for Tessellation which is really useful but only used in practice on Rocks and stone textures. Parallax Mapping also is the same too, very underused effect that has historically proven to work really well.

Simply put, developers are idiots that are using these technologies only for the buzzword and not for their actual practicality.

Lets be clear here, Procedural Generation should be treated more as a visual effect than an actual mechanic. Actually generating gameplay would be done more with Emergence (which I know is a buzzword but has a valid meaning which is, to discover gameplay where there is none. Or rather to get the AI to generate a design in a way it thinks the player will want it.).

Is an actual Game Design teacher. ;)

When gamers stop whining about 6-hour long games.

You can still create the underlying systems now and then make the game later.

I have a notebook where I write "concept clouds" and chunks of pseudocode when it hits me.

I'll be more clear on definitions:

ProcGen = When you create an art asset from individual lines of code specifying where everything should go on a model/texture and generating it out at runtime. (no drawcalls)

RandomGen = Taking pre-created art assets and randomly placing/using them in the game environment. (generates a draw for each texture and model)

The reason you're getting shit quality games has less to do about the effect and more to do with the developer thinking the effect gives them a free pass not to create content. Games that use random generated content can be good but it requires a LOT of playtesting where the developer is aware of the journey the player takes. The issue NMS has is that they very clearly didn't test that journey which is why its possible to get stuck and why the game has boring tedious and annoying parts to it, because its clear no one checked to see if the game was actually fun or not.

when gamers stop being retards

never

Okay, OP, I'm seeing your point. But imagine this for a moment:
A developer used a generation algorithm to create that city in your image there, and then they used that model to create assets and refine their world around. So in that case procedural generation has made an open-world city that was then refined by artists. With unique areas sprinkled in and given personality
The workload is reduced and the whole team now has substantially less work on their hands hand-crafting a vast cityscape. Of course I know AAA is too incompetent to do anything like that, but it's a nice thought.

Compare NMS to Minecraft - while NMS is nearly completely random, Minecraft has rules to it's procedural generation. Minecraft has a "canvas" of blocks to draw on, while NMS's canvas are predetermined patterns in which things can be arranged.

Minecraft uses mathematical functions to "draw" a world on a bitmap canvas, while NMS uses nothing but RNG to shove premade assets onto a dummy planet/station/star/asteroid. See the issue here?

Minecraft does something similar to this:

1) fill canvas with one type of block

2) paint it with randomly spread biomes

3) create heightmap based on biomes

4) paint layers of blocks based on their usual height restrictions

5) generate caves (calculus was never my strong suit. I have no fucking idea what's happening there. All I can say is that it uses functions to "worm" series of tunnels with occasional pockets of air. It's pretty sophisticated)

6) generate pockets of resources in patterns that are dictated by that particular resource

7) stuff manmade structures (some of them are procgen as well) into the world

8) draw rivers, lakes, ravines, etc.

9) plant trees, grass, etc.

NMS does the following:

1) create dummy planet with a hundred or so variables

2) RNG the fuck out of the variables

3) rinse and repeat with other objects in the "universe"

actually minecraft is all Random. It randomly places blocks. All of the textures are files generating draws, only the models are generated and they are from the basic cube setting I believe in Java's 3D APIs- and later DirectX for C++ version.

Pretty sure Minecrafts height maps are more translated into Z-position values rather than an actual displacement.

Same with Civs Map Generation, if you look at the code for Civ 4 which is open source you'll see it does map generation in layers but is only used to designate tile types which it references from art assets.

Procedural generation is straight up going from no predesignated file or data of any kind and generating art from a "procedure" of code- or basically telling the computer to create the art for you at runtime. What this means is that the actual texture and model are stored on the GPU rather than the hard drive which is vastly more efficient as it doesn't have to go through the front side bus which is occupied not only by the file you're calling (ie Draw Call) but also all the other crap Windows needs to operate without crashing. Using this form of Generation keeps the FBS clear because its stored entirely in the GPU- the limit would obviously be based on how many designated processes you can allocate- which on modern GPUs ranges into the thousands. This is why there was a lot of hype about ProcGen because it means you can literally create games with thousands of art assets. Whereas using files stored on the HDD generates a drawcall, which drawcalls by modern systems only go up to a few thousand. Its why its considered a good idea for Terrains as Terrains tend to look really badly detailed because they require so many unique assets to make not look shit. Of course than means nothing though if you simply just random generate the game as tiles from a single Atlas.

The benefit really is minimal when you get down to it. You will never want nor need to use that many art assets, and if you did you can simply atlas down to the point where the negative effects are minimal. The only time I could ever see a point to doing it is on maybe a mobile device, and even then, mobile tech is getting better every year, so eventually draws will be identical to most modern consoles.

TLDR; the old ways are still the best ways. Voxels will never replace polygons and creating an art asset and drawing it is superior from generating the art at runtime.

I am 99% you're either talking out of your ass or accidentally describing a different game.

Actually I am wrong, its not all random, it proc gens everything outside of the original map. I haven't played MC in a while and am thinking about how it worked in Beta- ie. Notch's shit implementation of it. Clearly the people he hired weren't as shit as I had thought because his original method was as I suggested. But yeah maybe I am mistaken as far as modern Minecraft is concerned, it has changed a lot.

Procedural generation can be fun.

It can work well if you take the time.

from the man himself:
So I switched the system over into a similar system based off 3D Perlin noise. Instead of sampling the “ground height”, I treated the noise value as the “density”, where anything lower than 0 would be air, and anything higher than or equal to 0 would be ground. To make sure the bottom layer is solid and the top isn’t, I just add the height (offset by the water level) to the sampled result.

So yeah he generates a perlin noise map, but this only allocates a Z-position to individual blocks, hence me being correct.

Later version of it used a proper displacement map that was generated as the game plays which is why they were able to do multiple realms that were as big as the original generated map.

Sorry for the confusion.

Thanks for reminding me about it with your picture, OP. Ironically enough, the picture you use is also one of the most realistic and effective city generators to date

why not both? have a story mode with hand crafted levels and some survival mode where everything is procedurally generated.

worked for age of wonders.

A game does not have to be art, it has to be fun. The art aspect is what has been taking all the fun out of video games. It's a notion started by pseudointellectuals that oppose GamerGate.

I'm getting paid to work on a procedural multiplayer rogue-like right now. Why're you such a hater, OP?

Glad to see someone in this thread who knows wtf theyre talking about. I was afraid of this happening when NMS came out, it gives people the wrong idea about ProcGen, even though it is unquestionably the future since it is capable of unlimited creativity that far surpasses hand-built shit.

yeah its really useful from a level designers perspective. Saves a lot of time at the expense of having to QA a lot more. The failures that exist primarily exist because QA is somewhat of a dead thing in our industry now, the industry got the "let the customer test our shit for us" bug that the IT industry had throughout the 90s.

So-called artists and pseudointellectuals in vidya really hate to be associated with fun (le bread and circuses pleb xD) or commercialism (fun games sell fucktons, because that's the whole point of games) and yet they are butthurt that their versions of gaymes don't sell well. Really, all they want is to make as much money as the other devs but they don't want to put as much effort in understanding what works in the medium, and then try to change everyone's standards to meet theirs. It's all a combination of laziness and envy, basically, and yet these kinds of fucktards are allowed to hold influence in vidya.

Are you some sort of retard? Pseudo-randomly throwing components together can't beat actual level design. I'd like to make ideafags obsolete as much as the next guy, but that's not going to happen.

There's nothing inherently wrong with procedural generation. The problem is that a lot of it happens to be pretty shitty.

So I see a lot anons say procedurally generated maps/items/events etc. aren't a bad thing, just a poorly executed idea. I would really appreciate a thorough explanation on what the perfect PG system would be. I think it entirely depends on the game; Isaac and Spelunky greatly benefit from PG content because their short to beat and were made to be replayable.

One of the bigger issues of Procgen is very easily understood even by a layman if explained properly. Do you know what "tiling" is?

It's a thing that happens when you a particular pattern in an image is repeated enough times for your eye to recognize it, and see it come through on all instances from there on out.

Same thing happens to Procgen. The believability and realism, the illusion of it being an actual "thing" falls apart the very second your brain understands the underlying pattern. It's why games like Minecraft are an incredible, insane amount of fun for the first 1000 hours but become very sharply boring after that.

It's why Procgen always has to be complex, convoluted and unpredictable, with lots of very unlikely things that come through once in a blue moon. You have to keep the illusion of novelty.

Binding of Isaac, for example, has thousands of item interactions. For that reason, its always something new for that first thousand hours until you learn all of them. How does Ludovico's Technique (floating controllable projectile) interact with Mom's Knife (knife melee instead of shooting tears)? Floating, controllable knife.

Procgen is a tool that has applications bordering on infinity, but is constantly used by hacks and wannabe game designers to create shallow box of coins, that all you can do with is shake it up and wonder how many heads vs tails you're going to get.

I need my fucking pills so I can actually start consistently working on a single idea that comes to fruition, I know way too much about this shit and output zero finished games, only unfinished prototypes and shitty proofs of concept.

OK, so, give us really good examples procgen in games.
Give us examples to understand it better.

What a shame

TBoI

Minecraft, Terraria, Dwarf Fortress, Binding of Isaac, wish I had better memory to take names off the top of my head. I'm more than certain there were plenty more that I played extensively.

Wew, I was in the middle of typing out my response and you posted that.

pardon the namefagging, I was just modding my board

You don't play vidya user? Anyone who've played vidya with any prominent procedural generation can easily understand what he is talking about because that's exactly what happens. And that also happened with NMS, and it inevitably happens with everything.

Thanks user.

This makes me proud more than it realistically should

(Checked)

ADoM is on Steam too.

good posts deserve recognition Well you opened my eyes to something I didn't really notice consciously until. I thought procgen was just randomly generate a map using limited resources. Different outcomes based on the same formula. I am thinking of making a game sometime in my future and I really love the idea of an endless content and possibility, epic adventure roguelike. Your post really helped understanding how to do roguelike properly.

This could also be why idiots are so easily impressed by procgen.
Pattern recognition is one of the keys to understanding complex concepts. The ability to reduce them from the complex to the simple. It's what we call intelligence.
Most people are stupid and either lack, or have a severely reduced ability to recognize patterns and are thus blown away by procgen since they can't see the underlying systems.
Since they never catch on to the patterns, idiots will always be blown away by procgen since they are literally unable to comprehend that the tools used are actually fairly simple.

That's a rather conceited way to look at it. I'd chalk it up to their lack of attention span rather than intelligence, to be honest. They get bored with it before it loses its charm.

I wouldn't be here if I wasn't a cynical asshole.

You too are the result of
Read about self-assembling macromolecules for example.

There's still a fuckton of normalfags playing Minecraft who haven't caught on to its patterns yet.

I'm the definition of a cynical cunt, i run >>>/sfg/ for fucks sake.

There's a difference between being cynical and being conceited. That difference lies in recognition that you're human, too.

Ha. Try Penrose tiling for non-repetitive patterns generated from simple rules.
stephencollins.net/penrose/
condellpark.com/kd/quasig.htm

But cynics, at least me, tend to start out being conceited, having their low opinions of others consitenly validated and end up bitter because being consistently unsurprised at the general ineptitude of mankind actually starts to hurt.

I know about games like Spelunky, Gungeon, Dwarf Fortress, Minecraft, TBoI, Terraria, The Flame in the Flood and some others. But why so many of them are "roguelike"? And why it's usually restricted to scenarios in most of the cases?

I think this list is useful store.steampowered.com/tag/en/Procedural Generation/#p=0&tab=NewReleases

Mankind is not a hivemind, it's a force of nature. Filter through enough shit and through comes a diamond.

Keep your expectations low but keep an eye out for those outliers.

Procedural level generation just doesn't fit a lot of genres, and never will.
Fags who say there's nothing wrong with it don't even play these games.

A huge problem with procgen is that it usually takes more work than simply making a good level, gets old as quickly as said level, and has no guarantee of actually being good.
Any sense of discovery while playing the game is lost when you realize that you're walking through a bunch of computer noise, only so that you can see more computer noise. Without a designer obligated to surprise you and keep you engaged, the interaction becomes one sided.
It's as if your DM was a drooling retard who decides every single fucking thing by dice rolls and doesn't care what happens next.

To be brutally honest, it's a lazy gimmick for idiots and a staple of low effort nu-games.

Opinion discarded. All depends on the complexity of the formula. Remember that the Mandelbrot set is also procedurally generated.

Ok……….but there's only one boss

and it was easy

One important thing to clarify is the necessity of meaningful variables. Alongside unpredictability it forms the foundation of good procgen. For example, NMS has some fairly unpredictable generation, like in creature appearances, but all these differences are shallow. Contrast this with Terraria, where a great deal of the worldgen is predictable, but very meaningful; the different kinds of dungeons and the locations where corruption is spawned have a significant impact on what the world can provide and how much work can end up on your plate.

Because Rogue started it and since most of them before Indies started moving over from sidescrolling to dungeon crawls rogelikes are simply just tiles or Ascii graphics so it's entirely up to your imagination what everything looked like. The new stuff claiming to be roguelike is more like Diablo in execution.

Daggerfall was procgen.

Could the problem be the dependence on procedural generation? As you've pointed out, you can use procedural generation for textures and such rather than tiling, but in games such as Minecraft and others where world procedural generation is the selling point, it has to be as variant as possible in order to remain interesting.

exactly, procedural generation is shit, I work 10 hours a month and spend the rest of the time shitposting on the internet

If you focus on a particular thing, you have to make it work well. That's not exclusive to procedural generation - it's true for all things.

And it's also boring as shit.

More interesting than your smug animu grill posting habit.

anime smugposting is the first sign of losing an argument. Don't be a retard, please. This thread was nice up until this point, so keep it that way.

Best way to use procedural generation is in combination of or in support for normal game design.

You have fixed levels and arrangements as the core and frame of your level design and then fill in blanks and low priority areas with procedurally generated content.

Not all games have the same requirements for a level.

Making a doom clone or making a realistic simulator for example, are two different things.
In the second example, you want the levels to make sense, while in the first you can create bases with insane layouts that no architect would ever make.

Or better yet, instead of letting the machine do everything let it logically combined handcrafted elements.

Procedural generation only really works when it is allowed to run within the confines of a well defined set of rules.

For what it's worth, Path of Exile is doing a good job with that.

Logically combining handcrafted elements is what every procgen does. The issue is the scale.

maybe there's a way for a machine to create "handcrafted" elements?
I can't imagine AI isn't far off from being able to create games.

QUIT TALKING LIKE A FUCKING NIGGER

Some games don't need/use procgen at all.

It's mainly because "Dungeon Crawler" has twice as many syllables as "Roguelike" and calling it a rogue-lite isn't hardcore enough. All anything has to have these days to be considered Roguelike is have permadeath and proc gen. It's lazy and wrong but it is what it is- if you don't like it, work toward changing that.

Also, there is absolutely nothing wrong with procedural generation as long as the game is designed and balanced around it. You'll have vastly divergent player experiences in contrast to a designed level where you're delivering the same pre-packaged experience to everyone. It's certainly a tradeoff that affects some games more but, as I said before, there's nothing wrong with that.

I've always thought in theory it should make the world something unique to you and more of something to explore and experience while enjoying the gameplay.

It NEVER works out like that, does it?

What the fuck kind of insult is this?

Holy fuck

What a run

Really? It works for some 4X strategy titles. Master of Magic with procedurally generated maps has more replay value than vanilla Age of Wonders 2 with only fixed maps.

This is actually an extremely good point. 99.9999…% of the Mandelbrot Set IS extremely boring. In fact, if you zoom in on any one place at random you will likely encounter a whole lotta nothing (and probably miss the points in the set and literally end up with nothing). Since it's infinite, though, there's some pretty cool shit in there, but ONLY if you know where to look.

So basically it's exactly like NMS. Infinite possibilities, and most of them suck. Some of them might be unbelievably beautiful, but the mathematical chances of you encountering them at random are pretty much zero. So why bother with the procgen in the first place? If you handcraft stuff to be unbelievably beautiful, you don't need the rest.

procedural generation can sometimes be good, but more often than not it is the hallmark of the lazy talentless hipster dev who doesn't want to be bothered with learning how to actually make an interesting, challenging, memorable level, because effort, and making their game procedurally generated gets them trendy meme points with all the other indieshits.

Wrong wording. Procedural generation is not sometimes good - it's an amazing tool. What sucks is that people use it wrongly.

Just the same as ambient occlusion, it's a wonderful thing that's used terribly wrong.

a hurtful one

but that's vaginal

He's talking out his ass, you can tell because of all the words. It's all procedural generation.

pcgbook.com/

You go play Morrowind or your average cRPG after the first run, there's a lot of shit you'll have memorized that you can just go get straigth away because you know it's there in very run.

I know that in NWN2, after I get Kelghar, it's always a good idea to visit the Lizard Cave because at the end there's the Phonix Bone mace, a pretty good mace for Kelghar that he will keep using for a long time and carry the game easily with it. And I get it in every playthrough because I know it's there and there's no reason to avoid it.
Morrowind does the same shit with some of the static loot where you know exactly where some of the best items are and you can rush them at lvl1 for a very easy early game.

Meanwhile, Oblivion had things like the Eidolon Edge (a very powerfull ring) drop in master chests at random. You actually had to explore dungeons until you eventually found it instead of getting it at lvl1 in the same location in every run you made.

It's still a bad example, but there's more like Elona that has a static overworld map but with randomly generated dungeons. You have a familiar place to visit every playthrough, towns are always the same, but the dungeons change, giving you different chalenges for you to face and explore.

Properly made procedural generation is exactly that: it gives you a new environment that you have to explore and react to, instead of the same scripted level in every run.
But in order to be good, it needs a very good set of rules that create levels that make sense and content that shows up and makes the game interesting.

Both No Man Sky and Space Rangers 2 have a randomly generated world, and yet the second one is a much better game because there are things happening in that world, making it a dynamic part of the game instead of a cute bakcground in which you do nothing because you can't interact meaningfully with the world.
And that's always the problem, you generate this wonderfull procedural generated world and then you can't do anything with it. Interaction stops at creation when it should instead be greatly expanded afterwards.

In this context, it means it has rules, i.e. a recipe to follow per the context.
Combining both terms, means the system follows rules, as to generate content per the context.

An additional goal here is to be able to reproduce the same content, as long as you give it the same inputs, and thus the same outputs.
That inherently means it's not random, as it has rules, and is deterministic.

ProcGen is far reaching, and doesn't necessarily have to replace traditional level design; nor does it specifically apply to level design in the strictest sense.
Also, an example of a sufficiently robust procGen alg relating to level design is one that takes as input all the data of a level designed by a dev, and reproduce said level w/variations within specified ranges (not going over a pseudo alg here, too lazy heh).
Potentially taking a seed for base variation inputs to methods used to alter layouts, rules to prevent certain oddities from occurring, limiting type or range of edits to specific types of objects in the level, or specific methods as inputs (using this noise function for this layout placement pattern for this type of object, etc etc); thus producing variations that could very well be worthwhile.
However, even better is to introduce a learning method into this, say, a neural network being trained with this good and bad level data (or an evolution algorithm, as to produce variations based on good sets of levels); which could produce variations, while becoming better over time.
Since that process would be offline, I'd imagine the overhead is negligible, and using something like GPGPU it would reduce computation times to mere ms.

It can also apply to many other things, and there are plenty of patterns elicited by physical systems irl that can be reproduced by procedural techniques.
However, the mainstream use of procGen isn't necessarily to add to complexity, improve quality of content w/quality control, or to create more possibilities for fun; it's to curb their laziness… so I understand where you're coming from.
Although, not all developers are that way, and imo to sufficiently judge procGen is to have a full understanding of its complexities, limitations, and potentials.

Though, there are times when proc-gen is necessary, and times when a personal touch is necessary.
It really depends… on context.

I don't recall which game actually did this, but it was a nice example of the way you describe ProcGen.

It was a co-op game similar to L4D where players had to traverse a level and reach the end, fighting whatever enemies they found on the way and using the supplies available to them as well.
However, the level layout wasn't the same in every playthrough, instead being composed of different tiles arranged in sequence till the end tile, which each tile having enemies, supplies, both or neither.

The interesting part was that the game prolonged the level accordingly to how well the players were doing and changed the amount of enemies and supplies generated according to the same rule as well. You'd have essentially levels generated for you that presented a chalenge for your group's skills as it was a system that generated levels based on feedback from your performance. Never too easy, never too hard.

To some extent, L4D did the same. Although the levels were the same, the supplies generated and the enemies spawned also changed accordingly to the player's performance. Do too well and you'll se more tanks and medikits being replaced by pills. That might count as ProcGen too.

Now the most important part in such a system is that it generates the levels based on your performance. It has a reason for doing so and a result it wants to achieve, that leads to levels that feel more interesting for every kind of player.
If you were to pick the usual indie way and just have it entirely and completely random, then there would be no point for the ProcGen.

I think L4D used an AI narrator, and this system is one having rules which produces content procedurally; hence qualifying it for procGen.
Agreed, it's a good use of procGen.
Especially in the sense that it can dynamically adjust for player skill, and will most likely improve a player's experience while increasing the possibilities for fun; in addition to having the sought after emergence of cool/fun moments.

Indeed, a mix of hand crafting (AI narrator) + procGen (changing level/content based on factors w/rules for generation) can go a long way compared to the usual lackluster method used.

When it becomes so good you don't notice it.

Procedural generation isn't a bad thing, they just need to figure out how to do it in an engaging way.

There was a programmer who made a program that could generate music compositions with the idiosyncrasies of Mozart. Every orchestra refused to record and album with it and he ultimately stopped trying after some time.

The right set of circumstances needs to happened for someone intelligent enough to do the same thing with game/level design.

There is a formula. It just needs to be simplified.

the fact that Tiling and Patterns were brought up in this discussion only proves Holla Forums is much smarter than the usual moron over at Reddit, you guys nailed it. Well done. So yeah the telling signs that an item is proc-generated is that it tends to repeat a lot, they try to mask it with noise but that most of the time looks really bad. Its the equivalent of using the Clouds filter in Photoshop which game artists are discouraged from using in favour of actual photo references. This is why its actually a bad thing because its only capable of producing art that a fairly average game artist could produce.

Correct, but only in parts. The dungeons are random-gen as they use modular pieces. The towns are random-gen for the same reason. The entire landscape is procedural as its generated from a dynamic height-map. Same applies with Elite 2 as well, all of the planets are proc generated, except for the cities and space stations which are randomly placed. Its why people get the two confused as the two are used in tandem with each other. The key to remember is one is generated at runtime (proc-gen) and the other is pre-cached before the game starts (random-gen).

To some extent even spore is a combination of both, as the creatures are made up of parts, but the body is sculpt-able clay which is generated at runtime. The planets are procedural with exception to vegetation and buildings which are all pre-built assets. That said Unreal 4 has started getting some procedural vegetation plugins so we might finally be at the stage where the trees and plants you see are all generated. But no game has actually done that yet. I believe it uses randomly placed splines and constrains the verts to them at runtime. Very impressive shit. Same could be done with Roads in a racing game if you can generate a spline and conform the road to it during runtime.

The key to proc-gen is that it has to be done at runtime, else its pre-built, which does not get the benefit of pro-gen being stored exclusively on the GPU.

Everyone wanted that screensaver planet generator to be a game. Shut up.

Design stuff to go on top of it. Walkways and buildings and such to make it work.

You know you like that too. Could even be a good enemy for you, talk to it, tell it sweet things, feel her breasts, no-one must know your secret.

What separates NMS and DF is width and depth.
In DF you have a multitude of different ways of achieve your goals, it doesn't matter what sort of fortress you want to do or what kind of adventure you want to have, the game gives you all the tools you need to do it. In NMS you have a blaster and a space ship that both do very limited actions.
In Dwarf Fortress you have lore that is completely optional but tells a story, literally hundreds of different stories. DF also rewards exploration and grinding, it gives you clear goals to work towards. In NMS there is virtually no story, no background info, all of the fluff is completely pointless and disconnected from everything else. It doesn't matter how many aliens you talk to or how many made up words you learn, it will change nothing. NMS also fucked up exploration and grinding. In the game there is literally no reason to explore, every planet has the same outposts, the same minerals, and animals that do the exact same thing. In NMS every upgrade feels completely pointless unlike in BF when you can notice the difference between a copper and steel sword.

Finally the biggest difference is that in Dwarf Fortress you can actually have fun instead of having an Indie Art House game where you look at retardedly randomly generated animals and rocks.

...

I get where you are coming from OP. It's like when you get games where you can design your own units. Developers are just using it as a crutch for lazy game design rather than putting any real thought into it. Probably cause they were too creatively bankrupt to design their own units.

i hope this trend continues but used in the right way, like on FTL or something and not like in other games where procedural generation is pointless like the animals in NMS

Actually, an interesting thing about Daggerfall is that, although the cities were RandomGen, the seed to generate each one is the same in all playthroughs, leading to precisely the same city being generated.
This allows the game to actually store the entire composition of a city in a single string of characters that can be invoked to visit the exact same city, which is an interesting way of saving an entire town and load it later in RandomGen.

Precisely. ProcGen excels at extending a game's width. As NMS shows, you can pick a very small game (you can describe the entire of NMS in just 5 or so planets) and extend it infinitely but only in with. ProcGen does not create actual content or expands further the gameplay at all, those are elements that the Dev still has to do on their own and if not present, the more you stretch the game, the more shallow it looks.

If you have a game where factions had to coexist, you could use ProcGen to create thousands of different factions, but if the only thing a faction can do is declare war or peace with any other faction, then that's all those thousand factions are gonna be able to do, which makes the game boring.

Honestly, if a game advertises itself as purely procedural generation, to me, it better have as much depth as DF or URR. Indie developers seem to think procedural generation is an excuse for them not to do level design or mistakenly believe it creates infinite replay value because they heard about roguelikes. What they don't realize is that roguelikes (good ones) use procedural generation as a means to make sure the player can't just memorize dungeon layouts, monster spawns, etc and will be presented with different situations that will test their level to think prepare and think in advance.

Thinking of it now, there's a lot of mechanics that roguelikes popularized that indie devs love to slap in their game for the sake of having it. Perma death being one of the big ones next to proc generation.

The problem with ProcGen is that it is greatly misused and devs are lazy fucktards for not putting in enough effort to make it good.

What I mean by misused is that ProcGen is VERY unsuitable for anything that is created with (human) intent. It's very good for shit that's made by nature, but anything made with intelligence falls flat on its face. Even then devs have to put in a fuck ton of content and textures for the ProcGen to actually make good terrains, fauna and anything else untouched by intelligent creatures.

Basically, devs are bad, lazy fucks and retards not because they use ProcGen, but because they use it for wrong shit and they don't put enough effort to make it good.

On the topic of No Man's Buy, I would add that it was always going to be bad because the premise was boring from the start. Space is boring, traveling through space alone is boring, uninhabited planets are boring, and (most) unintelligent alien animals are boring. The actual interesting thing about space SciFi is not space, but meeting intelligent species of other planets and watching their culture. People who bought NMS didn't stop and think through what they could find in the game that would make their time worthwhile and are thus retards.

Incorrect, you faggots simply have reading comprehension issues.

I meant flora, but fauna could work too, they are not that intelligent to matter at all.

Once, when playing Nethack, I zapped a wand of fireball at a Goblin. He had a shield and somehow managed to reflect the beam back at me. When I was struck with it, a Potion of Invisibility I was carrying broke and applied it's effect to me.

So there was this goblin looking at a crazy wizard that aimed a wand at him, he raised his shield and now there was someone shouting and cursing in pain as a mass of flames ran down the hallway burning an invisible wizard.

This kind of shit is what made those games interesting, the depth of the gameplay. The procedural generation just gives me more levels that I don't yet know to explore while experiencing all this.

Whenever Indie devs make their very basic and shallow game with none of those cool interactions between monsters, equipment and items, it's just like a boring maze being made even bigger to compensate.

Dwarf Fortress and Ultima Ratio are passion projects while shit like NMS are shitty cash ins. I want to see these fucking indie dev nu-males sit in front of a computer for hours on end programming a game about drunken dwarves or being thrown into a world where you have to learn the cultures on your own. It says a lot when a team consisting of just one guy can shit out a more complex game than ones with several people backed by a corporation.

According to the Berlin Interpretation Civ and Fallout 1/2 are roguelikes.

Sometimes its lazy and sometimes its the only way to do things; dwarf fortress for example.

If you replaced the graphics in fallout with ascii it would be pretty close.

Nethack is in perpetual development.

Why don't devs do this?

you wouldnt know level design if it cummed in your mouth

Pretty sure this was written by a bot.

A pretty good bot, then.

It depends how sophisticated your pseudo-random process is. When a human is designing a level they're still adding randomness - they're just doing it in a very careful and clever way. A sufficiently advanced algorithm could, in theory, surpass a human.

How do you figure?

Maybe. Assuming that the handcrafted stuff would be worth the effort. I prefer every 'natural' rock and grain of sand in a game to be the result of procgen instead of a human slaving away.

I actually just finished going on a rant about how faggot indie devs are giving procedural generation a bad name as a "lazy" method when in reality it's the most difficult, time-consuming method of content generation. I fucking love procedural generation and it is by far the the best method of content generation. If done correctly, that is. Most people just do like one shallow ass passthrough of generation and call it a day, which is why it gets such a bad name and that literally triggers me because procedural generation can be absolutely amazing when done properly.