As part of the Independent Games Summit at GDC today, Steel Crate Games’ Ben Kane took the stage to share some insight into the design of Keep Talking and Nobody Explodes.
The game’s story begins at the dawn of 2014, when VR still seemed relatively novel and headsets were scarce. At one game jam Kane and his compatriots saw a room full of people watching one person playing a game in VR, and they thought it was ridiculous. What if they could make a VR game multiple people could play with one headset?
“We made our prototype at the Global Game Jam, and we also made a video of people playing it,” said Kane. It got a surprising amount of views, so of course, one of the team members quit his job and they went all in on developing the game full-time. In the summer of 2015, Keep Talking and Nobody Explodes made its debut on Gear VR.
“Now that we’re on a bunch of different [VR] platforms and a bunch of different stores: in total aggregate, very vague data, we’ve sold more than 200,000 copies,” said Kane. “For a small team maybe taking a gamble on VR too early, we’re really happy with this.”
The core of Keep Talking and Nobody Explodes' design is simple: one player puts a VR headset on and tries to defuse a bomb, another player (or players) serves as the “expert” who talks the defuser through their task, using actual manuals for reference.
Demos weren't just playtesting -- they were key to the design process
“Demoing was going to be this ongoing process, all throughout development,” said Kane. It wasn’t just marketing, either; “Demoing fundamentally shaped what our game would be.”
Here’s how they did it: when players came up to play a demo version of the game at events, Steel Crate would hand them a postcard with info about the game (Twitter, website, etc.) and, on the reverse, a little notecard for them to fill out before playing the demo. Who, for example, would be playing the defuser? What level of difficulty should their bomb have?
There was also a large area set aside for taking notes, which ensured players left their demo with a personalized souvenir and also gave the Steel Crate team a chance to snoop on how people wrote -- and talked about -- their experience playing the game
“Our game involves talking, so our job involved eavesdropping,” said Kane.
Steel Crate assumed VR early adopters would like this game, said Kane, and they were right; but they discovered kids also loved it. They liked playing with other kids, and they liked playing with parents because they could “boss their parents around,” according to Kane. The parents seemed to like it too, because they could play games with their kids.
“This sort of wide appeal, recognizing people were having fun just communicating with each other, led us to a surprising conclusion," said Kane. "We’re not making a bomb game at all; we’re making a communications-based party game, it just happens to have bombs in it,”
When they realized that, Steel Crate knew they had to strip out any serious bomb references from their game. It didn’t make sense to make it scary, or threatening -- it had to be lighthearted and fun, though nerve-wracking nonetheless.
“We didn’t now what a finished game would actually look like,” said Kane. “We had these two pretty different roles players could take on, but we needed to focus….we needed a single guiding principle.”
So they decided to focus on what, exactly, they wanted the game to be: an experience that fosters interesting communcation.
Don't focus on the game, focus on the experience you want players to have with the game
“The fun really lives between players," said Kane. "Interseting conversations are where players make mistakes...but it’s also where friendships get made, and relationships get tested.”
So how do you foster interseting conversation in your game?
“It starts with balance,” said Kane. You have to make sure both sides have interesting, unique pieces of information to trade back and forth. In Keep Talking and Nobody Explodes, this plays out as the defuser (the only player who can see the bomb) relates what they see to the experts, who are the only ones who know what the various bomb variations look like -- and how to defuse them.
To spice up the back-and-forth, Steel Crate designed a special tracking system that will trigger random events based on how the players are performing.
“These might seem totally random, but they’re actually not; under the hood, we have a little pacing system,” said Kane. The game will tally a rough “score” of how well the player is progressing, and if they’re seen to be doing well, the game is designed to try and trip them up by, say, setting off an unknown alarm clock or turning off the lights.
The goal is to solve the tricky design challenge of giving the most players possible the most exciting, nail-biting finale possible -- without making the game too hard and causing them to fail the defusal.
“We don’t want to accelerate a death spiral; we want people to get those really exciting moments, where they defuse the bomb with two seconds left,” said Kane. “That’s difficult to get.”
Ideally, players should see any moments where they fail in your game as their own fault, advises, Kane, and they have to see how they messed up. For example, Steel Crate was tempted to make the “video game” portion of technically challenging, with intricate input requirements, but declined to do so because they didn’t want players to feel as though they’d failed because they weren’t good at video games.
“We require simple and deliberate inputs” said Kane. “The decision-making is the hard part….this ensures that the conversation doesn’t stray from team thinking, which is what we want, to blaming a player for their execution.”
Also, the rules had to be made clear and unambiguous -- something Steel Crate occasionally strugged with.
“The rules have to be totally unambiguous,” said Kane. “If players are forced to guess, or feel forced to guess, they’re either going to be wrong and feel cheated, or be right and feel unsatisfied, which is just as bad.”
This may seem simple enough, says Kane, but when you’re working on a game very closely it can be hard to see when you’re designing a mechanic that abides by ambiguous or confusing rules.
For example, Steel Crate realized halfway through development that the font in Keep Talking and Nobody Explodes made the capital letter O looked like the number 0, giving rise to confusing ambiguity when players were quickly reading out serial codes. To solve that problem, they wound up removing all capital Os from serial numbers in the game.
In another example, one of the game’s instructions mandated players cut the second wire in a bomb panel -- but didn’t clearly explain whether it was the second from the bottom, or second from the top. Steel Crate had assumed it was obvious, but in practical testing they discovered it wasn’t.
“The worst part, or the best part, is that none of these groups ever told us this was a problem,” said Kane. “It was only because we were eavesdropping on them that we caught this.”
The team continues to develop the game for new platforms, and they’re continuing to learn new lessons. It’s in demand in corporate teambuilding and school scenarios, for example, and some people enjoy playing it competitively.
But in summing up what Steel Crate had learned thus far, Kane reiterated that they would never have discovered what Keep Talking and Nobody Explodes could or should be without demoing it at events -- and they only made the right design decisions when they focused on the desired player experience they wanted to foster, rather than the nuts and bolts of the game they wanted to make.