I am super stoked about the procedural generation! Games like FTL get a lot of replay value by being different every time you start over, and it should make the development process quite enjoyable since we will be playing the game A LOT.
There are two parts to this (I think): terrain generation and dungeon generation. In Binding of Isaac, dungeons seemed to be generated by first building the level graph and then occupying each room with a pre-set template. This is easy but is limited to the number of room templates you have.
In Minecraft, terrain is generated using Perlin noise such that the entire world doesn't have to be generated at once but you still get consistent terrain. Will we be generating everything at once or go for this sort of generate-it-as-you-need-it approach?
There are tons of other issues, like how one incorporates puzzles and keys such that the game is solve-able (block pushing, anyone?), item/monster spawning, etc. I am curious as to whether we will be applying the combinatorics to monster generation as well, maybe we can get random bosses with weird traits: Evil-ice-zombie-dragon or lightning-leopard-goldfish.
Is it too early to talk about some of these things? What do you all think?
There are two parts to this (I think): terrain generation and dungeon generation. In Binding of Isaac, dungeons seemed to be generated by first building the level graph and then occupying each room with a pre-set template. This is easy but is limited to the number of room templates you have.
In Minecraft, terrain is generated using Perlin noise such that the entire world doesn't have to be generated at once but you still get consistent terrain. Will we be generating everything at once or go for this sort of generate-it-as-you-need-it approach?
There are tons of other issues, like how one incorporates puzzles and keys such that the game is solve-able (block pushing, anyone?), item/monster spawning, etc. I am curious as to whether we will be applying the combinatorics to monster generation as well, maybe we can get random bosses with weird traits: Evil-ice-zombie-dragon or lightning-leopard-goldfish.
Is it too early to talk about some of these things? What do you all think?
Same here! I used to play angband for countless hours many years ago and that was simple ascii graphics. I'd like to see perma-death in this game, but that might be too hard for casual players.
The tech driving handmade hero should be awesome from the bits and pieces Casey has mentioned so far.
One thing with procedural and random generated games is they can be more fun for the programmer to make rather than for the player since the coder knows all the cool stuff that's happening behind the scenes, but the gameplay ends up being kind of boring.
Its early days still but I've got my fingers crossed this should turn out pretty cool!
The tech driving handmade hero should be awesome from the bits and pieces Casey has mentioned so far.
One thing with procedural and random generated games is they can be more fun for the programmer to make rather than for the player since the coder knows all the cool stuff that's happening behind the scenes, but the gameplay ends up being kind of boring.
Its early days still but I've got my fingers crossed this should turn out pretty cool!
One thing to look at is the kinds of procedural generation used by successful games.
I think most platformers and specific-short-term-goal-oriented games use the cobble-together-tons-of-authored-blocks method: Binding of Isaac, Spelunky, FTL, I think Crypt of the Necrodancer does this, etc.
I think far less authored are things like Minecraft, Terraria, No man's guy, et al., but exploration is a huge part of those games and there's a lot less pressure to find specific things and accomplish specific tasks in a short-term timeframe.
So I think the style has to be conducive to the gameplay in that manner. Interestingly, I think Zelda-style games a la HH can be both. Sprawling, exploration-heavy overworlds with focused gem-like dungeons that are authored enough to contain puzzles and clever satisfying objectives is a niche waiting to be filled (if there isn't a game that has already done so (which I would love to hear about)).
I think most platformers and specific-short-term-goal-oriented games use the cobble-together-tons-of-authored-blocks method: Binding of Isaac, Spelunky, FTL, I think Crypt of the Necrodancer does this, etc.
I think far less authored are things like Minecraft, Terraria, No man's guy, et al., but exploration is a huge part of those games and there's a lot less pressure to find specific things and accomplish specific tasks in a short-term timeframe.
So I think the style has to be conducive to the gameplay in that manner. Interestingly, I think Zelda-style games a la HH can be both. Sprawling, exploration-heavy overworlds with focused gem-like dungeons that are authored enough to contain puzzles and clever satisfying objectives is a niche waiting to be filled (if there isn't a game that has already done so (which I would love to hear about)).
Edited by theinternetftw
on
Reason: strategic hyphen deployment
theinternetftw
I think far less authored are things like Minecraft, Terraria, No man's guy, et al., but exploration is a huge part of those games and there's a lot less pressure to find specific things and accomplish specific tasks in a short-term timeframe.
So I think the style has to be conducive to the gameplay in that manner. Interestingly, I think Zelda-style games a la HH can be both. Sprawling, exploration-heavy overworlds with focused gem-like dungeons that are authored enough to contain puzzles and clever satisfying objectives is a niche waiting to be filled (if there isn't a game that has already done so (which I would love to hear about)).
The interesting thing about your "less authored" examples is that, if you went and took a look at the game's source, you actually would find parts, at least, that were authored. Minecraft has its strongholds and villages, Terraria its dungeons, and No Man's Sky its ship components, animal templates, plant templates, etc (that last one is guesswork, but I'm betting its not all done on the fly! It would probably be the highest in procedural-to-authored ratio, though). I get your point, though -- all those games emphasize their procedural aspects much more strongly and the authored portions are small segments within.
As for your counterexamples, I can think of two. The first: Starbound -- planets are generated randomly, but occasionally you can come across one with authored dungeon segments that work similarly to Binding of Isaac (abutting room templates). The second: Spore. It didn't start out with them, but the Galactic Adventures expansion allowed for planets to contain content authored by other players, which your "captain" (a concept that barely existed before then) could get out and explore. I can't say it pulled it off particularly well, but it at least had the right idea.
That being said, I think there's still plenty of opportunity for a game to come along and pull it off really well, with a Skyrim-level overworld that varies and Skyrim-level (or better, let's aim a little higher than that) dungeons. I, for one, would be anxious to try such a game.
I too am psyched for this. Procedural content always reminds me of like improvisational jazz.
They're both perpetually unique.
Yet they both run the risk of feeling forever the same.
The way Casey put it (I think), roughly, 'things made by algorithm tend to look like they were made by an algorithm.'
So you run into the problem of having an infinite minecraft world, but little reason to leave your spawn, since it's all the same every way you go (they've addressed this a bit, but last time I played those solutions still felt a bit like band-aids on a cracked foundation).
-
The second big pitfall I usually think of re procedural content is the risk that players will feel they need to keep playing for some certain output from the RNG gods...
This was an epiphany I had playing FTL: At some point (regrettably, I had to cave and look up how to unlock those last few ships) I found myself replaying the game over and over hoping to get the right combination of systems and random events to unlock those last couple ships (a new ship, after all, is new content.. one of the biggest rewards you can offer a player..) and I recognized an important and sharp distinction between the feeling of "playing more to get better" and "playing more to obtain a certain random output."
Personally, when I find myself in the category of the latter, I know it's time to quit that game: I've gotten everything out of it I will. At that point, it's just a glorified slot machine.
This maybe isn't so much related to the procedural content as much as the way the systems around that content function... but I thought I'd mention it. Also it's just my own personal preference when it comes to games (after all, some people really enjoy slot machines!).
-
In case anyone hasn't seen it, check out this for a neat interactive explanation of Spelunky's random level generation. I thought Derek Yu had an article on this himself, but this was the closest thing I could find, maybe it's what I had in mind. He mentions briefly how it works.
They're both perpetually unique.
Yet they both run the risk of feeling forever the same.
The way Casey put it (I think), roughly, 'things made by algorithm tend to look like they were made by an algorithm.'
So you run into the problem of having an infinite minecraft world, but little reason to leave your spawn, since it's all the same every way you go (they've addressed this a bit, but last time I played those solutions still felt a bit like band-aids on a cracked foundation).
-
The second big pitfall I usually think of re procedural content is the risk that players will feel they need to keep playing for some certain output from the RNG gods...
This was an epiphany I had playing FTL: At some point (regrettably, I had to cave and look up how to unlock those last few ships) I found myself replaying the game over and over hoping to get the right combination of systems and random events to unlock those last couple ships (a new ship, after all, is new content.. one of the biggest rewards you can offer a player..) and I recognized an important and sharp distinction between the feeling of "playing more to get better" and "playing more to obtain a certain random output."
Personally, when I find myself in the category of the latter, I know it's time to quit that game: I've gotten everything out of it I will. At that point, it's just a glorified slot machine.
This maybe isn't so much related to the procedural content as much as the way the systems around that content function... but I thought I'd mention it. Also it's just my own personal preference when it comes to games (after all, some people really enjoy slot machines!).
-
In case anyone hasn't seen it, check out this for a neat interactive explanation of Spelunky's random level generation. I thought Derek Yu had an article on this himself, but this was the closest thing I could find, maybe it's what I had in mind. He mentions briefly how it works.