Welcome, Guest. Please login or register.

Login with username, password and session length

 
Advanced search

1411283 Posts in 69325 Topics- by 58380 Members - Latest Member: bob1029

March 29, 2024, 01:08:08 AM

Need hosting? Check out Digital Ocean
(more details in this thread)
TIGSource ForumsDeveloperDesignGameplay that suits procgen
Pages: [1]
Print
Author Topic: Gameplay that suits procgen  (Read 1689 times)
hmm
Level 2
**


View Profile WWW
« on: June 30, 2016, 11:44:23 AM »

I'm a bit of a procedural generation fiend when developing my own little personal projects, and I've been using it extensively at work as a way to expand the scope of what can be achieved with a small team.

One thing that interests me, and I was hoping to hear all your thoughts on this, was how procedural/random generation affects the way people play a game, their experience of the game, and the types of game mechanics and structures that work best when used in conjunction with procgen environments.

I feel as though there are a few emerging patterns among the games that feature procedural generation. Roguelikes/lites/likelikes are the obvious trend, which feature a focus on the player's mastery of the game systems and mechanics, and in some cases strong meta-game progression systems. But are there other ways of looking at procgen as a feature? Arcade games are another genre that use procgen extensively, but also focus on mastery.

So, to summarise the above as questions to y'all:
  • What game mechanics, or approaches to gameplay design do you think work well with procgen environment?
  • What meta game structures work well with proc gen? (Series of generated levels? Focus on player upgrade systems?)
  • How do you see procedural generation being used to create specific player experiences beyond mastery focussed games?

Very vague, I know. Just wondering what you think. Also, here's links to some of my procgen efforts.
Mt. - climb mountains generated from player input
Colourblind - odd roguelike/puzzle game featuring very random 'dungeons'
Kapsul Infinite - endless series of randomly generated levels in a physics action game
Logged

voidSkipper
Level 2
**


View Profile
« Reply #1 on: June 30, 2016, 05:41:15 PM »

I think that the majority of procgen games focus on mechanic mastery because that's about all they can achieve. Essentially, the game has a core mechanic, and then the procedurally generated aspects are just the game throwing different scenarios at the player in order to challenge that mechanic. Perfect mastery exists when you essentially overcome the proc-gen nature of the game by developing strategies that are effective regardless of the product of the procgen algorithms.

Less-explored procgen focuses would be exploration, discovery, and communication.

For example, a game where you find unique, uncharted land would be proc-gen exploration.

A game where you categorise procgen'd creatures would be a discovery game (think pokemon but where it actually makes sense for you to be filling out your pokedex).

A game where you simply live in a city and make friends and enemies with NPCs who have proc-gen's lives would be communication.

However all of the above is considerably harder to procedurally generate in a meaningful way, because the things that make fictional lands exciting, fictional creatures interesting, and fictional characters' lives enthralling is very hard to capture in an algorithm. It's very hard to procedurally generate 'inspired' content.

The more "hooks" (eg, human-designed aspects) the algorithm relies on, the less longevity the algorithm will have before it starts repeating itself and becomes, once again, a mechanic master game.

To clarify, I mean that for example, in the "living in a city game", the programmer is going to have to come up with various lists of words that can describe the NPC's situations, tastes, natures, etc. But no matter how long that list is, it's going to become homogenenous after enough gameplay. Before long, metagame discussions will emerge like "oh no, he has a 'flighty' nature, he's probably going to end up in an 'affair' scenario. I just finished one of those, so I'll talk to someone else".
Logged
s0
o
Level 10
*****


eurovision winner 2014


View Profile
« Reply #2 on: July 01, 2016, 04:54:55 AM »

Quote
For example, a game where you find unique, uncharted land would be proc-gen exploration.

literally every single roguelike ever. also every single space sim ever.

Quote
A game where you simply live in a city and make friends and enemies with NPCs who have proc-gen's lives would be communication.

the sims

 
Quote
Before long, metagame discussions will emerge like "oh no, he has a 'flighty' nature, he's probably going to end up in an 'affair' scenario. I just finished one of those, so I'll talk to someone else".

that sounds like real life to me actually
Logged
darkhog
Level 7
**


Dragon Agent


View Profile
« Reply #3 on: July 18, 2016, 10:36:52 AM »

Quote
A game where you simply live in a city and make friends and enemies with NPCs who have proc-gen's lives would be communication.

the sims

Lol, I'd actually love to play fully procedural The Sims where the city layout and families (names, look) are generated fully procedurally so you get different city each time you start a new save file. You can of course set a seed so you can replay specific city over and over again.

Also think about how much smaller it would be - heightmaps and level data (including data regarding sims themselves) take fair amount of space so the game could ship with more content instead.
Logged


Be a computer virus!


I cannot C well, so I stick with simpler languages.

There are no impossible things, there is only lack of skill.
s0
o
Level 10
*****


eurovision winner 2014


View Profile
« Reply #4 on: July 18, 2016, 11:03:46 AM »

actually why doesn't the sims have that option? wouldn't be difficult to do from a technical standpoint. there are plenty of precedents for world generation at this point.
Logged
darkhog
Level 7
**


Dragon Agent


View Profile
« Reply #5 on: July 18, 2016, 12:35:43 PM »

actually why doesn't the sims have that option? wouldn't be difficult to do from a technical standpoint. there are plenty of precedents for world generation at this point.

Probably because EA and they want to sell worlds too, darn it!

Seriously tho, I think because making any procgen is easy, but making one that gives both coherent results and doesn't suffer from "once you've seen one level, you've seen all" syndrome is hard as you need to constantly tweak the formula used to generate things. Even the best procedural games, like Enter the Gungeon, Binding of Isaac and Spelunky suffer from "once you've seen one level, you've seen all" syndrome. Heck, even Minecraft suffers from it to some degree as all villages and strongholds are practically clones.
Logged


Be a computer virus!


I cannot C well, so I stick with simpler languages.

There are no impossible things, there is only lack of skill.
hmm
Level 2
**


View Profile WWW
« Reply #6 on: July 26, 2016, 07:33:56 AM »

That Sims idea highlights something that could work particularly well with procedural generation. Since the behaviour of the characters is a simulation (rather than scripted) it could be interesting to just watch how the systems interact with one another.

Which actually brings me back to Age of Empires - you could play on randomly generated maps, and tweak the parameters yourself. It introduced this kind of play where you could explore how the game plays under different conditions quite easily. In that sense, there's a kind of meta gameplay with the procedural generation itself, where most games that feature procedural generation hide their inner workings from the player.

So, maybe that's something to explore? A kind of experimental gameplay where the player can create game modes for themselves by playing with the procgen system.
Logged

s0
o
Level 10
*****


eurovision winner 2014


View Profile
« Reply #7 on: July 26, 2016, 12:36:01 PM »

as far as im concerned, the sims is already a "procedural" game. almost everything happens dynamically. it wouldn't be that hard to randomize the initial town layout (like i said). and i don't think it would make that much of a difference.

imo, the difference between "procedural generation" and "handcrafted" design is kind of overstated in general. what procedural (level) generation is, is using a set of rules (i.e. a game mechanic) to determine the starting state of a game rather than using a fixed setup procedure. of course computers can use much more complex rules more easily, but designwise generating a planet in no man's sky is not fundamentally different from drawing a hand at the start of a poker round.

so in that light i think "procgen" benefits any game that benefits from a randomized or dynamic starting state.
« Last Edit: July 26, 2016, 06:43:12 PM by Silbereisen » Logged
quantumpotato
Quantum Potato
Level 10
*****



View Profile WWW
« Reply #8 on: July 27, 2016, 05:22:16 AM »

On what voidSkipper said about complexity & Silbereisen said about dynamic, I think anything where you are introducing varieties is a great spot for procedural.

https://boardgamegeek.com/geeklist/75022/shibumi-challenge-geeklist-winners-announced - challenge for humans to generate design rules on a 4x4 board with marbles of 3 colors that can be stacked on a 2x2 formation of marbles.

The human designed rules were transcribed into code for procedural generation of rule variants. The outcome was reviewed by humans and a set of rules were published.

Inspired by Kenta Cho's Defeat Me http://wonderfl.net/c/9ykQ I made a SHMUP where enemy movements are mirror copies of how you killed them. On Level 3 you fight a clone of yourself from Level 1 + Level 2. On Level 4 you fight a clone of yourself from Level 1 + Level 2 + Level 3, etc. www.quantumpilot.me

(+1 to what hmm said about starting conditions!)

This isn't traditionally viewed as "procedural" but I consider competitive multiplayer games to be procedural. As one strategy dominates, counter strategies evolve. All of the pieces (the "core algorithm") were there from the start -- all advanced strategies are theoretically available from the first publication of the game.

Personally I find the concept of designing a game intended to have its strategies evolve extremely interesting - if you know everything about the game, there's no surprise. Know too little and the game will spiral into a stagnant dominant strategy too fast. Patch too quickly and deeper strategies will never be generated.
Logged

matwek
Guest
« Reply #9 on: July 28, 2016, 06:31:15 AM »

I sometimes have a bit of a strange opinion on randomly generated content if I'm aware of how that content was created. Sometimes it can draw me out of being immersed within a game. For example...

I'm exploring a randomly generated world and I find a treasure chest. Within that chest I find an awesome new weapon thats perfect for tackling that boss.

The question I can't help asking myself is if the procedural content was designed to create world where the game director has hidden a chest containing the item I'm looking for and through exploration of the environment, I found it.

Or did the game director decide that I needed to find that weapon, so when I opened a chest (any chest) it decided to spawn that weapon within it.

The first option makes me feel like an adventurer/explorer who has discovered something unique, the second makes be feel like a factor in a numerical equation.

Logged
hmm
Level 2
**


View Profile WWW
« Reply #10 on: August 02, 2016, 05:04:07 AM »

I sometimes have a bit of a strange opinion on randomly generated content if I'm aware of how that content was created. Sometimes it can draw me out of being immersed within a game. For example...

I'm exploring a randomly generated world and I find a treasure chest. Within that chest I find an awesome new weapon thats perfect for tackling that boss.

The question I can't help asking myself is if the procedural content was designed to create world where the game director has hidden a chest containing the item I'm looking for and through exploration of the environment, I found it.

Or did the game director decide that I needed to find that weapon, so when I opened a chest (any chest) it decided to spawn that weapon within it.

The first option makes me feel like an adventurer/explorer who has discovered something unique, the second makes be feel like a factor in a numerical equation.

I know what you mean. I've been thinking lately about an approach where the focus is on the generator producing interesting or 'natural' results, and gameplay is designed around that. From the player's point of view though, this could just end up being terrible. Realistic terrain/dungeons/whatever aren't necessarily the most fun or interesting to explore.
Logged

hmm
Level 2
**


View Profile WWW
« Reply #11 on: August 10, 2016, 04:55:08 AM »

I've been reading a bit more about the gameplay in No Man's Sky now that its out, and it seems as though it opts for survival+exploration as its core game loop.

While pure exploration sounds great, and is something that a lot of people were excited about with NMS, it seems as though developers resort to survival mechanics as a kind of psuedo-challenge to make the exploration feel a bit more 'dangerous'. Its a bit disappointing, as the craft/survive loop feels more like an arbitrary obstacle.

That's just my opinion though. What are your thoughts on the survival loop commonly used in procgen/exploration games.
Logged

s0
o
Level 10
*****


eurovision winner 2014


View Profile
« Reply #12 on: August 10, 2016, 05:06:33 AM »

nothing wrong with making exploration feel dangerous. some of my favorite games like dark souls or la mulana do "dangerous exploration" extremely well and the danger actually enhances the exploration. but adding a progression loop to a sandbox exploration game almost always results in griiiiiiiiiiiiind

but regarding NMS, keep in mind that they're trying to market it to a mainstream audience. it's basically being marketed as a AAA game. i think they probably feared that a huge game where you "do nothing" a la noctis wouldn't be palatable enough (the game still had those criticisms leveled at it regardless). so they added mechanics their target audience is familiar with from minecraft and terraria. that's my impression without having played it anyway.

btw everyone excited about no man's sky should at least try noctis. i don't think the windows version runs on 64bit systems, but the dos version runs well in dosbox.
« Last Edit: August 10, 2016, 05:15:23 AM by Silbereisen » Logged
dao cowboy
Level 0
**


View Profile
« Reply #13 on: August 24, 2016, 12:41:51 PM »

A couple of quick thoughts on Procgen:

1)  I'd like to see a Roguelike or something similar where the progress you've made upgrading your character doesn't die.  Any interaction with the world, loot picked up, things built, etc., would die with the map, but not your character upgrades.  It'd be nice to if you had some "play" where you could adjust your build within certain limits if you want to try something or you think you made a mistake.

2)  I'd also like to see procgen where adjusting the parameters was part of the gameplay.  Say you have a fetch-quest heavy procgen JRPG.  As you beat the initial worlds, you unlock adjustments you can make to the next worldgen.  You might, eg, make the world at a higher difficulty in exchange for more XP.  Beating the entire game would be a matter of unlocking all the worldgen parameters.
Logged

I have books for sale on Amazon.  Search for "Peter van de Welde."
s0
o
Level 10
*****


eurovision winner 2014


View Profile
« Reply #14 on: August 24, 2016, 01:05:19 PM »

Quote
1)  I'd like to see a Roguelike or something similar where the progress you've made upgrading your character doesn't die.  Any interaction with the world, loot picked up, things built, etc., would die with the map, but not your character upgrades.  It'd be nice to if you had some "play" where you could adjust your build within certain limits if you want to try something or you think you made a mistake.

rogue legacy has that, and probably a bunch of other modern "roguelites" too. i personally don't like it because it always results in grind.


Quote
2)  I'd also like to see procgen where adjusting the parameters was part of the gameplay.  Say you have a fetch-quest heavy procgen JRPG.  As you beat the initial worlds, you unlock adjustments you can make to the next worldgen.  You might, eg, make the world at a higher difficulty in exchange for more XP.  Beating the entire game would be a matter of unlocking all the worldgen parameters.

games by soldak entertainment do exactly that.
Logged
saluk
Level 2
**


View Profile
« Reply #15 on: August 26, 2016, 10:16:13 AM »

I would like to see more procgen exploration in rulesets vs in the environment. I think that one of the things I enjoy most about games is that learning aspect - figuring out what the elements are in the game and how they interact. If the algorithms were constantly throwing new wrenches into the mechanics, I would be more likely to stick around longer, as once I have mastered the mechanics is precisely when the game becomes dull. No Man's Sky is a clear example of the opposite - you master the mechanics in the first hour, and then the rest of the game exploration of all of those crazy worlds is just not enough to carry the game, as there is not a lot more to learn.

The closest I think we've seen to this are loot systems in ARPGs or roguelikes. Every new piece of loot you have to analyze, as you just don't know what kinds of positives and negatives are going to be thrown at you. The best arpgs provide their fans with many many hours of entertainment, even ones that do not take place in a procedural environment.

Eventually even there you do approach the limit. Once the mind can sum up all of the data into an integral and see the equation behind it, then they have mastered the "meta" mechanics and there isn't any more to learn.

Another good example of the kind of procedural gameplay I'd like to see explored more is the nemesis system of Shadow of Morder. While I didn't like the game and the movesets in those types of action games don't hold much depth for me, it was interesting to not know where the bosses were going to be, what their weaknesses would be, or how they would attack. No one complained about the shallow algorithm there because it was such a light touch, but it was enough to change things up and keep players in the zone more.

I think finding gameplay that suits procgen is perhaps backwards. The best approach might be to see how procgen can enhance a particular game and give it more variety and longevity.
Logged

s0
o
Level 10
*****


eurovision winner 2014


View Profile
« Reply #16 on: August 26, 2016, 02:40:33 PM »

Quote
I think that one of the things I enjoy most about games is that learning aspect - figuring out what the elements are in the game and how they interact. If the algorithms were constantly throwing new wrenches into the mechanics, I would be more likely to stick around longer, as once I have mastered the mechanics is precisely when the game becomes dull.

well that's exactly what roguelikes are
Logged
JWK5
Level 9
****

A fool with a tool is an artist.


View Profile
« Reply #17 on: August 26, 2016, 02:46:37 PM »

I don't know if it has already been done, but taking the original Doom/Heretic/Hexen/Duke Nukem/Blake Stone/etc. games or something along those lines and having procedurally generated levels would be pretty awesome.

I know there is Tower of Guns and this other one on the PS4 I forget the name of where you play as a sorcerer (reminds me a lot of Heretic/Hexen), but I mean something closer to oldschool Doom.
« Last Edit: August 26, 2016, 02:52:20 PM by JWK5 » Logged

My Art Tutorials:
 Here

"Today is victory over yourself of yesterday, tomorrow is victory over lesser men." - Miyamoto Musashi
valrus
Level 3
***


View Profile
« Reply #18 on: August 26, 2016, 05:53:30 PM »

I like your Mt. game, and it reminds me that procgen is relatively untapped in sports games, specifically sports whose variety and interest is in environmental challenges like skiing, golf, skateboarding, etc.  It's a natural fit, because the core loop in all of these sports is "hey, here's a different combination of environmental elements, what do I do about it?"  There have, of course, been sports games with random courses all the way back to the first sports games, are there are some sports games where the procgen really is attempting to approach more humanlike level design, but it's not a big trend the way procedurally generating dungeon-crawlers and platformers has become.
Logged
hmm
Level 2
**


View Profile WWW
« Reply #19 on: September 06, 2016, 05:53:18 PM »

I don't know if it has already been done, but taking the original Doom/Heretic/Hexen/Duke Nukem/Blake Stone/etc. games or something along those lines and having procedurally generated levels would be pretty awesome.

I agree. The old school fps given the procgen treatment would be cool. The question is though, what would procgen add to the experience beyond infinite levels? Infinite isn't necessarily a good thing.

I like your Mt. game, and it reminds me that procgen is relatively untapped in sports games, specifically sports whose variety and interest is in environmental challenges like skiing, golf, skateboarding, etc.  It's a natural fit, because the core loop in all of these sports is "hey, here's a different combination of environmental elements, what do I do about it?"  There have, of course, been sports games with random courses all the way back to the first sports games, are there are some sports games where the procgen really is attempting to approach more humanlike level design, but it's not a big trend the way procedurally generating dungeon-crawlers and platformers has become.

Thanks! And yes! I've actually toyed with the idea of a procgen skiing game. I think sports games fall into the category of games that focus on the development of a skill. Procgen allows the player to constantly pit themselves against new challenges.

I've been playing a few short, narrative driven games recently (Abzu, Inside) and was wondering how procgen might apply to such experiences. Definitely not a natural fit - the focus of these games is to deliver a tight and finite experience so handcrafted levels / puzzles seem to be the obvious way to achieve that. However, I think there may be some interesting ideas to explore at the intersection of the two.
Logged

Pages: [1]
Print
Jump to:  

Theme orange-lt created by panic