The Develop Post-Mortem: The Banner Saga

When Stoic’s tactical RPG The Banner Saga arrived on PC and Mac in 2014, it was immediately striking.

Exquisitely well-drawn and beautifully hand-animated, the game by a team of former BioWare staff boasted a distinctly accessible turn-based combat mechanism and a branching narrative that deftly juggled a sense of player freedom with a tight, well-paced delivery.

Two years on, The Banner Saga 2 – the second part of a planned trilogy – has arrived, and it’s got the Stoic team in a reflective mood. After all, it is the work on the first game that allowed the team to push themselves further for the sequel. 

ON THE AIR 

Famously, The Banner Saga was created with a custom engine; a move that, perhaps surprisingly, was conceived to allow the team to move quickly and start creating content from the off. Except, of course, games development is never so simple, and even describing Stoic’s development platform as an ‘engine’ might be a little inaccurate.

“The word ‘engine’ has a lot of different meanings to a lot of different people,” offers Stoic co-owner and technical director John Watson (pictured).

“One thing to understand is that, really, if you’re going to build a game of any complexity, you’re going to be building an engine for your game, even if it’s on top of something like Unity.

“We had all these bespoke game systems like the tactical combat, unique ability system and our RPG class systems, and then the stage dialogue.”

Throughout the whole development, the focus was on content; asking how we get the game playable, get content in and get it so that we can see it, feel it and know what we’re doing.

John Watson, Stoic Games

Stoic was in a position many other developers will empathise with. The team had a strong idea of what they wanted to build and, as such, needed a technology solution optimised for their vision. At the same time, it wanted to get working on content as soon as possible. All of which meant there wasn’t the time or need to write a custom solution for the fundamentals of game development, such as rendering, audio, keyboard and mouse input and networking.

“Never build that yourself,” Watson advises other studios.

Ultimately, the decision was made to let Adobe Air handle the elementary foundation, on top of which Watson and his colleagues built a custom technology framework within which they forged their game.

Adobe Air isn’t the most obvious choice of game technology, but it worked for Stoic, and happened to perfectly compliment its approach to that sumptuous animation style; one that is so hard to capture without a human hand painting each frame.

FLASH BACK

Adobe, of course, is also behind Flash, a perennial platform across the creative sectors, loved and loathed in equal measure. 

“The way hand animation is done, in the majority of cases, is in Flash,” Watson explains. “If you watch Cartoon Network or animated television shows, they’re done in Flash. So the artists that were creating our art for us were creating it in that format. One nice thing about Adobe Air was that we could take that content, in the native artist format, and render it in the game.”

That approach let Stoic begin on its best foot, launching the team into a furiously productive opening development stage. By the end of The Banner Saga’s creation, Watson and his colleagues had changed their rendering pipeline so as to use efficient raster-based sprite sheets. However, using Adobe Air at the off allowed the team to delay time consuming work on a rendering pipeline, and focus on the development of content, quickly building a clear vision of their game world.

“Throughout the whole development, the focus was on content; asking how we get the game playable, get content in and get it so that we can see it, feel it and know what we’re doing,” confirms Watson. “We were going to worry about the performance and portability later. That’s what allowed us – as a really small team – to actually finish the game in two years, and do it as well as we did.”

That process is part of what John describes as a ‘technical debt’ approach, where the work is easy and productive upfront, at the expense of ease of porting and other work later. By the time Stoic took to porting The Banner Saga to console, it would feel the pinch of that debt. But the team remains confident taking out a technical debt let them make a successful game.

“If we hadn’t been willing to take on some technical debt then, we might have taken a lot longer to finish the game, and that would have been very bad for us, because we had limited budgets and limited time,” Watson asserts.

FORKED LIGHTNING

Another challenge – and triumph – of The Banner Saga is its branching narrative. It was written largely using Inklewriter, an interactive fiction tool by 80 Days outfit Inkle that was never expected to be used on such a polished game. In fact, Inklewriter is more commonly associated with classroom use.

“Using the Inkle tool is another example of technical debt, really, as it got us off the ground quickly,” states Watson. “It meant that I didn’t have to write a branching dialogue tool. Straight away we could immediately start writing content for the game.”

We had to leave a lot on the cutting room floor. That was difficult; very difficult. We did well on that by the end of Saga 1, but at the beginning we weren’t really sure what we could pull off.

Arnie Jorgensen, Stoic Games

Indeed, four years after beginning work with the tool, Inklewriter is still in use at Stoic, and in wielding it, the team have learned a great deal about the delicate art of building interactive narratives. Of course, with its heritage in BioWare’s history, Stoic has worked on projects with renowned branching game writing before. But the fact remains; for a small team, structuring a game with narrative forks is a momentous task.

“People want diverse endings and branches in games that would blow any studio’s budget,” suggests Drew McGee, freelance lead writer and former writer and designer at Stoic. “Every breath you take and every eyelash [your characters] bat makes your title a different game. That’s not something feasible, so what we learned and discussed early on for The Banner Saga 2 is that we write to the idea that the player is the leader of a caravan, not a god or leader of the world.

“That means what you do and the choices you make can have drastic impacts on the people around you and the way they see you, but not necessarily across the world as a whole.”

CLEAN CUT

Creating any game, of course, means challenges, and the Stoic team faced plenty. In particular, developing The Banner Saga offered a powerful lesson in understanding the editing process.

“We probably over-scoped on the first game at the start,” co-owner and art director Arnie Jorgensen muses. “We then cut right back, and we had to leave a lot on the cutting room floor. That was difficult; very difficult. We did well on that by the end of Saga 1, but at the beginning we weren’t really sure what we could pull off.”

That whole process, of course, has been hugely useful for the recent work on the sequel.

“On Saga 2, we have a lot better idea of what we can pull off,” Jorgensen says. “We came in right where we wanted to be with the length, and that is really tricky to do. At times it felt like we working on it blind, not knowing the true scope of it.

“On Saga 1 we bit off as much as we could chew, but that meant on Saga 2 we could absolutely bite off as much as we could chew again. But it wasn’t more than we could chew. That was important.”

Article originally published in the May 2016 issue of Develop.