See related article: Ascii Dreams: The Death of the Level Designer: Procedural Content Generation in Games - Part One

As for it being the death of the level designer?  Absolutely not, this guy is just trying to get views on his site.


Anyway this was something that I thought of years ago (tyvm), and I think it'll be interesting to see how gamers react when someone finally does it right.  "But Jacob, it's already been done right!  I've played games where the terrain was randomly generated, and where whole dungeons were created randomly which gave the game unlimited replayability."  I call BS, at least in the games that I've played, I'm hardly impressed with this random level generation.  Sure, if there's a game that randomly builds rooms/dungeons by reusing/rearranging cubes then technically the user will have a "unique" experience everytime they play, but it still feels like I'm playing the same levels.  I'm still encountering variations of the same twists and turns, textures and enemies and eventually (and it doesn't take long) but I start realizing I'm doing the same **** thing just slightly re-hashed.


As someone who's played video games like crazy since NES, I've become incredibly self aware of the tipping point at which playing a game becomes boring/monotonous.  I think we've all experienced the gamer's attachment lifecycle.  An upcoming or existing game piques our interest, it impregnates our minds and thoughts until the moment we get our hands on it, and if it's a good game, we binge like crazy for days, weeks, or months on that one game.  That is our game and every chance we get we play it for any number of compelling reasons (it's fun, new and engaging, there are weapons/items we are striving to attain, all our friends are still on it).  This is akin, in many ways, to "new couple love".  You were just going through life aimlessly sampling the playing field, getting your heart broken along the way until your night in shining armor (your new game) came along.  Now you've got a new "crush" and love is in the air.  Man, you've never seen a game with graphics like this, and every time you play it you fall in love all over gain.  And this goes on, for a while.  But at some point something happens...


All of a sudden the game doesn't offer the same high it once did.  The novelty is gone, playing is no longer an engaging experience, we don't give a **** about any new weapons/items we could get, and maybe too many of our friends have moved on.  Playing a game during this stage is like being in a relationship that's going nowhere.  In the beginning you two had loads of fun, everything was brand new, it was like a god damned fairy tale.  Now there's no ****.  Those graphics you once adored are hardly stimulating anymore, and in fact you're tempted by the graphics of other games (don't feel guilty, we're all human).  Lastly, every time you play it you're simply reminded of all the little things you hate about the game and would love to change.  It's time to move on, and you do, starting all over again.


What I mean about doing automatic/random level generation right, and the purpose for my whole ridiculous tangent, is that cookie cutter rearranging of the same stuff is not enough to "keep the love alive" or "rekindle that lost flame".  If it was done in a way that created the sensation of actually getting new content, then the replayability and value of a game would drastically increase.  This is of course a very complicated problem to solve algorithmically.  The level generator would need lots of sample content (and the ability to generate more?), and the architect of this system would need a thorough understanding the current level design and what the players are looking for.  This system if developed for one game, may be in fact so closely tied to that game that it may not be easily extended and used for others.  But I still think it's an idea worth toying with.


Imagine a game where it's shipped with NO levels.  Lets put it into the context of a first person shooter.  Everytime the players go to join a game, the level is generated right there on the spot.  No one knows where all the best spots are, and players will need to actually think strategically about the positioning of their team and the enemy.  Camping would actually be skill that would need to be learned, you couldn't just know that some building is great for posting up and mowing enemies down because everytime you've played that level you've seen someone else do it!  Spawn camping would be a thing of the past.


And what about real-time level generating in-game?  Ever play a game where you're moseying around the map and BAM, you've hit an invisible wall?  That's the edge of the map, how realistic.  What if you could just... keep going?  It'd certainly be a pain for someone chasing you who in the past might have known the level better and cornered you because you're new.  And wouldn't it be exhilarating darting off into uncharted territory with someone on your tail?


This sounds computationally expensive, and hard...  But being able to solve hard problems is something that can reasonably be expected of MSOE students, and new technologies emerge at a rate that I certainly have hard time keeping track of.    To me, nothing is out of the question, and if you're to compete with all of the indie game devs and industry powerhouses, you better bring something new to the table.