Worldbuilding on a budget

Worldbuilding isn't just for AAA teams with the resources and staffing to create open-world blockbusters. In a panel discussion at the DICE Summit today, three successful indie developers--D-Pad Studio's Simon Stafsnes Andersen (Owlboy), Campo Santo's Sean Vanaman (Firewatch), and Vlambeer's Rami Ismail (Nuclear Throne)--discussed their approaches to worldbuilding on a budget.

The three games had different team sizes, genres, degrees of narrative focus, and development cycles (Owlboy took nine years to complete), and the creators had similarly differing strategies and pieces of advice for worldbuilding. For example, Campo Santo set Firewatch in a Wyoming park with a distinct time period because he didn't to have to waste time and resources creating a world from scratch. He actually marveled at how people working with fantasy worlds can focus on their stories instead of getting bogged down in trying to consider what every last object in the fiction should look like. For Andersen, the trick was to figure out what the world looked like, and then reverse engineer things around that.

"I'd drawn all these assets that we had for the backgrounds and everything," Andersen said. "And I'd start looking around the background, going, 'What does that mean? What is that symbol that's on there? Why are there a bunch of owls flying around? How does this make sense?' And that's how we built our lore. I don't think we actually figured out our story until maybe the last two years of development."

"Say you have one box that goes into another box. This sounds silly, but you can actually sit there and say to yourself, 'What is the motivation of this one box going into another one?'"

Simon Stafsnes Andersen

He added, "It's interpreting your own abstract work. And I actually recommend this to a lot of people when they make programmer art and that sort of thing. You can even do it with small boxes. Say you have one box that goes into another box. This sounds silly, but you can actually sit there and say to yourself, 'What is the motivation of this one box going into another one?' And just the process of doing that will give you some sort of story you can base the rest of your game on."

One guiding principle for Andersen in creating his story was "having something that feels like its own thing." He called it "The Ghostbusters Principle," but acknowledged that moniker has become less apt since the release of the Ghostbuster remake.

"When you watch that movie, it is that movie," Andersen said. "You're not comparing it to all these other movies. It just feels like its own thing. So I apply that to every aspect of the game I can... You just go for interesting elements and then the rest of it can be interpreted."

Ismail said Vlambeer took a nearly opposite tack with Nuclear Throne, which lifted many characters and elements from the developer's favorite bad sci-fi books, and simply bent them to fit the game's universe.

"All work is based on previous work," Andersen conceded. "That's just something you have to admit to yourself. The challenge is to make it feel like you're not just dragging something straight out of the page. And I feel like we all have that same approach, just in different ways."

One tactic Vanaman said Campo Santo uses is to generate marketing for the game they would like to have made. For Firewatch, that took the form of the game's original trailer, which they then showed off to everyone and set the bar they were then forced to live up to.

"In terms of communicating creative direction, it's incredibly efficient, because you have this product target," Vanaman said.

Even though Vlambeer used an open development model with Nuclear Throne, live streaming the development process every week and constantly accepting and responding to feedback from the community, the one thing that never really changed that much was the lore. Very early on in the game, the team had a 20-page lore document detailing 2,000 years of history in the game world. Much of that detail never surfaced in the game itself, but it was a helpful internal tool for the team.

"Your ability to create is dwarfed by any one player's imagination space. You give them a gap and then they fill it with infinite possibility"

Sean Vanaman

Much like Nuclear Throne, Owlboy's development was shaped by the community, even if the finished product didn't always reflect that. Andersen said after the original Owlboy demo was released about six years ago, they heard plenty of players speculating about features that would be in it. Thinking that the demo had set expectations for those features, the D-Pad team then implemented some of those features, only to find out they were making the game boring, so they took them out.

"As we were expanding the game, we realized what the game was supposed to be for ourselves," Andersen said. "It feels strange to say this on stage, but outside influence hasn't had that much to say on what goes into the game for us. We sort of figured that out for ourselves, what this has to be. And then let's present that in the best possible way for others to find out. So a lot of our game development has been making ways for people to interpret what our actual official story is. So it's all about leaving clues for that."

Vanaman has embraced that approach at Campo Santo, saying, "You could spend another nine years making Owlboy, but your ability to create is dwarfed by any one player's imagination space. You give them a gap and then they fill it with infinite possibility."

When people are doing traditional AAA worldbuilding, Vanaman said, they're answering too many questions and filling in every detail rather than supplying players with more questions to keep them interested.

Ismail cautioned that can backfire on developers, pointing to Vlambeer's experience with the dogfighting action game Luftrausers. Ismail said it was drawing heavily on World War II and Cold War fears of German and Russian superweapons, so they thought it would be cool for players if they could control those superweapons. They only intended players to think of the force they were on as "not the Allies," but the Nazi-inspired iconography of the game led many to assume they were intended to be playing as a Nazi.

"For us, it was a warning that if we were going to leave space empty, to at least consider what could be filled in there," Ismail said.

With Nuclear Throne, there were points in development where they realized that people could possibly interpret something they definitely didn't want, so they would refute those ideas in the game, sometimes with as simple a fix as a loading screen message.

Andersen pointed out that for Owlboy, the developers deliberately left out any hint of romantic relationships between the characters in the hopes that players would focus on the intended topics of the narrative.

"Now if you look at all the fan art and stuff people have written about the game, that's literally all it is," Andersen said. "It sounds like I'm going negative about that, but that's actually kind of incredible. Something we didn't intend for and didn't even plan for is what came out of this entire process and what people felt was important after this. It's something we couldn't have prepared for when we started."