Procedural Content: Eighth Generation Hope or Hype?
by Josh Snyder
Gaming can be an expensive hobby, so to combat this, I make a trip every holiday season to my local retail outlet, where I purchase copies of video games (always new copies, if possible) I missed at release, but at a lower price. Therefore, I always spend this time of year playing two year old games, or catching up on downloadable content (DLC) I missed when it was released. Recently, I wrapped up developer Arkane Studios’s story DLC for their hit title Dishonored – The Knife of Dunwall and The Brigmore Witches. I also used this as an excuse to replay through the main storyline one more time.
When revisiting Dunwall and its charming rat plague, two things jumped out at me – I had played this campaign plenty of times before, and I was still finding new methods of infiltration, which is astonishing. At the same time, when I had to crack a safe or get a combination for a door, I knew exactly where to go to get the exact same code I had used numerous times before (only one safe in the game has 10 possible combinations – the rest only have one). There was a part of me that, for the sake of time, just wanted to turn to the internet to get the code for the safe and be done with it, but that felt like cheating, like an even bigger waste of my time. So I went through the motions, wondering if there was a better way this mechanic could be handled.
Implementing replayability into any game is a tricky task, one that developers have been struggling with for over two decades. One method has been procedural content – a system that generates levels, enemies and other elements randomly. There have been varying degrees of success with procedural content, and developers are still trying to strike a balance between a defined story and world, player involvement and impact on the world, and random content. But after my fourth playthrough of Dishonored, a game that is otherwise spot-on in its execution of stealth and combat, I wondered if the eighth generation is the time to implement procedural content into standard game design and theory.
62 Square Miles
Procedural content was first introduced as a way to solve a problem – how to make huge games with limited technology and limited resources. One of the most famous examples is Bethesda’s Elder Scrolls II: Daggerfall. While later games in the series like Elder Scrolls IV: Oblivion had open worlds roughly 16 square miles in size, Daggerfall was absolutely massive, with a world roughly 62 square miles for gamers to explore. The catch was that those 62 square miles were randomly generated – towns, dungeons, enemies were all random, giving each game, in theory, a unique experience. However, this led to a new problem – because only a handful of towns and dungeons were initially created for use in this system, gamers ended up experiencing a lot of repetition, the exact problem Bethesda was trying to avoid.
Still, the concept was one that many developers found attractive, especially on the indie scene. It’s an efficient way for a small team to build large-scale games. And in 2010, the concept took off. Games like Minecraft and Terraria used random content generation for each new world players created and explored. The success of these games has led to a whole host of developers trying their hand at procedural content.
But it’s important that the lessons Bethesda learned are still recognized today, along with the lessons both Mojang and Re-logic learned in successfully implementing procedural content. Depending on the type of game a developer is making, the tried and true method might work fine, or end up alienating gamers.
A New Sandbox, Every Time
The one common theme both Minecraft and Terraria have is that they are, at heart, sandbox games. There are traditional gameplay elements to each, but the reality is that each game gives the player the tools to reshape the world as they see fit. A quick search on YouTube turns up hundreds of real and fictional destinations, painstakingly recreated block for block in Minecraft. And with Terraria, players can construct huge mansions, filled with numerous non-playable characters who provide any number of services to the player.
For many gamers, the fact that each of these games offers procedural content is an afterthought – this element is not what draws in players. Think of each Minecraft map as a blank canvas – sure, the landscape has unique features, but those features are not the attraction. Doing whatever you want is the attraction. Procedural content was used to help players accomplish this by giving them enough in-game resources to build Minas Tirith or Hoth.
Procedural content works in sandbox games because it is a mean unto an end – creating a fully functional, customizable sandbox. It makes sense that in newer titles, such as Starbound and No Man’s Sky, that this method of content creation is being touted and praised – it works here, because the end goal is not the randomly generated content, but what players do with it.
Where Daggerfall Stumbled
As Bethesda learned, the problem with procedural content is that it cannot be the main attraction. Unfortunately, outside of sandbox games, few games have used this system to a positive effect.
Daggerfall simply repeated itself – a problem Bethesda continued to struggle with through Oblivion. After enough hours of gameplay, every town looked the same, every cave looked the same. Sure this cave might have trolls and a tunnel that goes to the left, whereas that cave might have goblins and a tunnel that veers to the right, but ultimately, it all felt the same.
A possible fix some developers found was to simply create more content for the system to use, lessening the issue of repetition. But the core problem still remains – procedurally generated worlds, in the long run, alienate the player. Players not only enjoy games that tell them a story, but they enjoy games that put them in a unique setting. There is a reason the world of Red Dead: Redemption is often praised – the attention to detail invites the player in, encourages them to explore the world. You never know what might lie over that hill, because everything in the world is unique and different. But if it’s randomly generated, even from a decent size selection of possible assets, you might end up running over that hill to see a town that looks awfully similar to the one you were just in. The sense of exploration is actually lessened when players realize that this is just a randomly cobbled-together collection of wood and bricks – it feels empty, sterile, as if it were manufactured in a factory, and not crafted by hand by artists.
The issue is that the way developers currently use procedural content will almost never work outside of the sandbox genre. Major AAA titles require a more defined structure to support both plot and character development. These are elements that cannot be left up to chance, or be randomized. But the counter to that is to simply hand-craft each locale and enemy encounter. Bethesda ultimately took this route with Fallout 3 and Elder Scrolls V: Skyrim, and these games are often cited for their level design and immersive world. Of course, Bethesda can afford to do this because of the resources at their disposal, leaving this as an option many other developers, even those working on AAA titles, simply cannot afford.
A Light at the End of the Tunnel
But the allure of procedural content is strong – for instance, take my example from Dishonored. Imagine if I knew what the city of Dunwall was going to look like, if I knew the major locales and plot points. But what if, on my first mission, the layout of Coldridge Prison changed slightly from the first time I played it? What if a door that was unlocked before suddenly requires a key? It would force me to think on my toes, to really utilize the skills I have obtained after playing through the game. Building on that, any game that offers even light puzzle elements could benefit drastically. This is already evident in Dishonored with the numerous ways you can infiltrate a building and eliminate your target. Granted, this isn’t a case of procedural content, but in theory it works the same way – depending on certain variables, certain approaches might be favorable over others. The replayability of these games would increase by a staggering degree.
Procedural content on a smaller scale can actually have a bigger impact in non-sandbox games. The upcoming survival horror game Daylight, from Zombie Studios, promises to offer just that – targeted, refined procedural content. The game takes place in an abandoned mental institution, and the developers have said that the layout of the building, along with the location of several paranormal scares, will be randomly generated upon each playthrough. This strategy could pay off, given the survival horror setting as one drawback to that genre is that, upon repeat playthroughs, the scares pack a lot less punch. With random scares, Daylight could be just as frightening on it’s fourth playthrough as it was on its first.
More Bang for Your Buck
Of course, the reason developers should be interested in using procedural content, and be motivated to implement it in new ways, is that it might solve the problem publishers have had with funding AAA games – the used game market. Granted, publishers arguments may be incorrect or ill informed, but that’s beside the point. Publishers still write the checks, and if they feel that the used game market is hurting their bottom line, right or wrong, they will act accordingly.
The way in which developers prolong the experience of a single player game is not through forced multiplayer or ill-conceived online passes, but with a game that changes each time it is played, offering a new look at the same story. With the accumulated knowledge of game design and theory at developers disposal, and with increasingly more powerful technology, procedural content should become a staple of the modern day developer. And it very well could be the first innovation that truly defines the eighth generation.