How Supergiant hopes to avoid the sophomore slump with Transistor

It's been 3 years since Supergiant Games released Bastion in the summer of 2011 to remarkable critical and commercial success. The indie studio's debut title went on to sell over two million copies across all platforms, and Supergiant's Greg Kasavin tells me its success was both unexpected and affirming. It validated the team's decision to quit their various jobs in AAA development and essentially lock themselves in a house to work on something they were passionate about. It also set a high bar for Transistor, the follow-up to Bastion that Supergiant is releasing later this month. "Nobody knew who we were and we had nothing to lose. If no one liked [Bastion] it would have just come and gone," says Kasavin. "There are a lot of people with high expectations of us this time."

Act casual

I point out that lots of creators struggle a bit when following up a successful debut -- whether it's a second album or a second novel, the "sophomore slump" phenomenon isn't unique to game development. Kasavin admits that he feels a bit of pre-release pressure, but it's not so different from what it felt like to release Bastion -- though Transistor has the benefit of a bit more development time. "We didn't really have a preproduction phase at all on Bastion, whereas this game did," says Kasavin. He claims the team appreciated the extra time to work on fleshing out the game's setting, its core characters and its gameplay systems before formally revealing it, but he also makes it sound as if they floundered a bit without validation from the outside world that the concept was sound. "It actually wasn't until we revealed the game, in March of last year, that we felt like we were on the right track." It's a sentiment many developers can sympathize with: Kasavin claims he and the folks at Supergiant have been working on Transistor so closely, for so long, that they don't

really know if it's good any more -- they just know that they like it. Kasavin says Supergiant used all that time to challenge "every aspect" of their new project -- from the camera angle to the use of voiceover -- to try and avoid retreading the path blazed by Bastion. To hear him tell it, one of the tougher challenges was working out how to make the isometric real-time combat of Transistor satisfying over long periods of play. Combat in Transistor plays out in real time, but the protagonist Red can freeze time and activate a strategic planning mode when her energy bar fills up that allows the player to queue up a series of actions. Each action saps a bit of energy from the bar, and when time unfreezes she executes them all instantaneously. It took Supergiant a good bit of time to get that system working, but once they were happy with it the question became: what keeps that core combat loop engaging over the long haul? "We really wanted the game to encourage player experimentation," says Kasavin, who laments the fact that Bastion players tended to select a favorite combination of weapons and upgrades that they would stick with throughout the game. This time Supergiant wants to encourage Transistor players to experiment with their abilities, so they themselves experimented with a systems of drawing and discarding powers that were inspired by collectible card games. "The CCG analogy sounded cool, it held up in conversations, but in practice there was just too much contrivance in making players discard powers they were using," says Kasavin. "We tried a lot of systems, and nothing really felt right." Eventually the team landed on the "combinatorial" system that's in the game now, which is designed to encourage players to experiment with different power combinations by ensuring that powers can be combined with each other to affect the game world in interesting ways. Those development changes didn't require anyone to rewrite the game's master design document, because it doesn't have one. Neither did Bastion, for that matter. "We’re small enough as a team that we can iterate ad hoc, and we like to work on a more intuitive level," says Kasavin. "When Darren is creating a piece of music, for example, he's just gonna go off and make that music. No one is gonna micro-manage him on that."

Don't plan ahead

I press Kasavin a bit on how Supergiant manages to work together harmoniously; now that the studio has swelled to twelve people it seems to be stretching the limits of what we might call "small team" development. "We're pretty quiet; the word 'monastic' has been used to describe us," says Kasavin. "We don't over-schedule, and we plan in a tactical way; we have plans of what we're going to do over the next week or two, and that's pretty much it." It helps that Transistor is (mostly) built on the same tech that Supergiant used to

build Bastion. The engine -- which was originally developed to use XNA and later modified to use a fork of Monogames so they could run it on platforms like iOS and PlayStation 4 -- is a pure 2D affair with a bunch of custom stuff running on top, including a unique animation and lighting system. They also recently switched to using Lua for scripting and Fmod for audio work, including lots of dynamic music layering and what Kasavin excitedly refers to as "neat audio tricks." All that work has been self-funded. If Transistor doesn't do well for them, Supergiant will have to reevaluate whether it can continue to operate the same way. Kasavin tells me he likes to joke that the game's May 20th release date is like the end of the Mayan calendar for the studio: "there's nothing beyond that date. We have no plans for what comes next." It sounds dire, but when I tell Kasavin that he's quick to put a positive spin on the situation. "We didn't plan too far ahead before Bastion was out, and a lot of interesting opportunities came up after the fact that we never could have expected," he says. "If we'd had firm plans in place beforehand it actually would have handicapped us a bit....since we're small, I think we can move fast and take on stuff as it comes."