On Quora, a user asked a question about procedural generation in games. On the one hand, it gives the player the opportunity to explore huge locations, but on the other hand, the world created by the program can be empty or "cheap". A discussion broke out and, as is often the case, both players and developers joined it.
So is procedural level generation a way to save money on development or an opportunity to learn new technologies? Let's work it out with Quora users.
As a reminder, procedural generation is the automatic creation of levels, items, quests and characters in a random way using a pre-defined algorithm.
Game designer Mike Prinkle expressed the opinion that in his life he has never met such a procedurally generated game in which he would not be bored after 15 minutes. He believes that any such game will eventually become a "swamp" with a bunch of incomprehensible locations and quests. Only games in which game designers work through every aspect can generate interest and emotional response from the gamer. After all, artificially created content is unpleasant to interact with. Level generation is most likely just an excuse for not putting all your energy into creating your content.
In contrast, one of the developers of Lost Planet 3 and Far Cry 4 Marc Maretea spoke out. He argues that the lack of generation makes the game less "lively". He, like many players, does not like it when the game is 100% scripted. There is often dislike for such a project, when, for example, it turns out that the open world is actually linear, and the side quests and "exploration" are artificially built into it.
More precisely, the artist Andrew Bishop spoke, and we support his opinion. He says that it just needs to be applied to the place. For example, the generator of random cards in the Age of Empires strategy is quite appropriate thing. It is worse when procedural generation in games is inserted where it should not. An example of such unsuccessful practices is Spore and No Man's Sky. Everything looks different in them, but the gameplay features remain the same and predictable.
Let's stop at Spore. The gameplay is based on survival, and evolution is also a basic mechanic. When you are at the primitive stage, for example, you can meet different predators. Some may have a large body, a huge tail and a mouth the size of your head, the latter may be smaller than you, but with their dramatic differences in external, randomly generated form, they will attack you the same way, with the same force, depending on the type their fangs or paws.
Not cheap, but tasteful
"Cheap" doesn't always mean bad, says indie developer Daniel Super. Procedural generation actually helps create the game world with much less tools. As an example, he cites No Man's Sky, where the indie team Hello Games simply would not have pulled the world of Mass Effect.
It so happens that procedural generation is considered "cheap", as experiments with it are carried out most often in the field of indie games.
Fighting the routine
Travis Adder, former game developer for Storm 8 and Super Evil MegeCorp, believes procedural generation as such is not bad. Level creators simply have to rely less on it while developing the game, not to shift the main work to the algorithm. And so, it allows you to remove a large layer of routine from work.
If we compare the open worlds of ten years ago and today, then these are colossal differences. This became possible not only thanks to the cool hardware. It was the generation that made it possible to create intermediate content without any problems.
She can arrange trees or bushes herself, while the developer is busy creating things that are beyond the control of the program.
There are things in development that can be trusted to the algorithm, and some simply cannot be generated. As an example of this hypothesis, indie-developed Katrice McCloud recalls "Mega Man clones", in which character enhancements were artificially generated by enemies. This devalues the whole battle, since you spend a lot of energy, and you can get a small reward. Something similar happened recently with Anthem.
Some elements just don't work, if they are randomly generated, they can interfere with whole genres. But in special cases, they expand them. According to developer Edward Hughes, in some SteamWorld Dig, generation would help maintain interest in replaying the game a second time. Hollow Knight, on the other hand, would ruin it, because the environment there tells a story, and random generation in the construction of the dungeon is simply impossible.
Bloodborne has a particular problem - the game is great, but procedurally generated dungeons make it all to nothing. You start predicting where the monsters will come from. The atmosphere of the unknown is still hidden before our eyes.
To summarize. So is procedural level generation good or bad? Definitely a good thing in the right hands. Using it correctly along with the "manual" work, the game designer will be able to create the game in such a way as to save energy on the necessary things and improve the gameplay.
The Topic of Article: Procedural Generation in Games: Good or Bad?.