Building the Empathy Machine, Pt. 2

This machine kills tear ducts

Welcome back to this two part look at the pipelines and processes behind the story crafting and writing of narrative games. Last time, I focused on Life is Strange: Before the Storm and Life is Strage: True Colors. This time, I will continue on with the next two games I shipped at Deck Nine: The Expanse and Life is Strange: Reunion.

With the finish line for True Colors in sight, studio leadership was plotting out what would come next. This time, they aimed for two simultaneous projects rather than one. A main reason for this has to do with a unique feature of narrative games: the concept phase.

Simply put, the concept phase is where the project leadership figures out what story they want to tell. Every game starts from a core concept. If it’s a gameplay concept— “Let’s make a stealth extraction shooter”— teams can begin prototyping right away. Design can build a gym to run gameplay ideas, environment can greybox levels that fit the design, animation can start work on a move set for the core gameplay, etc.

But if it’s a narrative game, the story needs to be fleshed out enough first for the early work to have any value. With True Colors, for instance, we knew we wanted to make a game about empathy. Cool, but that doesn’t tell anyone what gameplay loop we’re exploring, or give the art team anything tangible to concept, or let animation start testing things. For that, we needed to invent Alex, her power, Haven Springs, etc. And that takes time.

So what do other teams do while the story leadership is busy figuring out what game they’re making? Truthfully, they still do a lot. There’s systems to refine, new tech to explore, asset libraries to build, etc. But because they’re not fully tasked on the project yet, publishers can be reticent about paying for the entire team during concept phase. And even if they do, it creates a massive time pressure on narrative to lock in decisions that ends up causing problems down the line.

Which brings us back to running two projects simultaneously. This way, the concept phase for each project only needs to carry half of the studio’s personnel (assuming an even split) allowing for longer runways before ramping up into production. And if you get the timing just right, you could conceivably avoid layoffs between projects by shifting staff gradually from wrapping one game to pre-production on the next.

This is the theory, anyway. In practice it is very, very difficult. But aside from having a massive hit game that funds all future projects by continuing to generate revenue for the studio, or raising more funds from outside investors, it is about the only way to ensure staffing continuity between projects.

And so that was the backdrop for splitting the studio into two teams and beginning work on what eventually became…

The Expanse

Story Concepting: Initially, the project was another IP entirely, which fell away at the last moment and required an aggressive pivot. Many Deck Niners had been suggesting The Expanse show to me for years, and if I'd listened to them I might have mitigated some of the frantic energy in what transpired over the next month of concepting: binging six seasons of the show and nine novels, plus crash courses on space travel, 0-gravity physics, terraforming, and basic astronomy.

We briefly considered a number of possible stories and player characters—Amos was a top contender—but we pretty quickly landed on telling a Camina Drummer genesis story aboard a belter scavenging ship. Like most good ideas, it was a choice that felt obvious and inevitable right after we made it, even though it took a lot of work to get there.

Team Makeup: 3 writers. We brought on a new Junior Writer and an experienced Contract Writer. We also had a brilliant new Game Director who, though he didn’t write scripts himself, was integral to the story crafting and editorial process.

Pre-Production: As mentioned in Pt. 1, we knew that our success going forward would depend on improving how we forecasted production scope from the script. This became even more pressing for a smaller team with a tighter budget and many first time leads in key roles.

To combat this, we made two significant pivots:

1. Integrate scoping directly into the scripting tool set, so at the click of a button we could determine the “cinematic minute” impact of any given scene.

2. Write a rough version of the entire game script first, rather than a Narrative Arc. By having the scenes scripted earlier, we could make more accurate projections, not just for cinematic minute scope but also character count, environments, VFX, and so forth. Even if the scenes changed drastically, which they often did, these initial projections proved more worthwhile than those from the Narrative Arc.

Story Breaking: As a consequence, our story breaking work was much different than previous games. Instead of a Narrative Arc, we moved ahead with a story summary, a 5-page outline, and character bios. But because the writing process began so much earlier, we were able to explore ideas on the page and make drastic changes without halting production. Rather than treating story breaking as a distinct stage, the story emerged from the writing and the critical consideration of what we had written.

Content Pipeline: For the first time, we were working with a lead performance capture actor from out of state, and a bonafide TV star taboot, whose travel determined our shooting schedule. We were excited to do it, of course, because we were massive fans of Cara Gee’s performance and excited to feature it in the game… and then we were absolutely floored to discover the kind of human being and professional she is. It’s impossible to imagine how we could have made the game without her— her infectious energy, her stamina, her rare combination of integrity to her character and openness to new ideas. She was a human cheat code.

The shooting cadence with Cara determined the flow of our content pipeline which in turn dictated our writing schedule. Our Performance Director, also new to the studio and also brilliant, helped orchestrate all the pieces. Unlike True Colors and Before the Storm, but very much like television and film productions, we had to adopt a rigidity to our script locking process, knowing that we only had Cara for a finite time for each shoot and only a finite number of shoots for the game.

Take Aways: Many principles of TV writers room practices and production cadence had been present since Before the Storm. They’re baked right into the narrative toolset, the talent, and the culture. But it was not until The Expanse that we fully emulated a professional writers room with rigid lock schedules, thorough production scoping, table reads and other kinds of mocap prep, and highly efficient (yet still really fun) shooting days.

I’ve mentioned how we shifted our process to write a draft of the entire script during pre-production, which was key. But there was an involuntary factor that affected us just as much: the pandemic. The Expanse was our first fully remote project. Many of us were adopting to the new normal of slack messages instead of cubicle visits, zoom calls instead of meeting rooms. Some new members of the team never stepped foot in the office once. We found we needed to be far more deliberate about our culture and our communication than in any prior projects.

There’s an interesting tension at the heart of a cinematic narrative game development. The way that games want to be developed, like any software, is iteratively. Because the player is the ultimate variable, you want maximal time to test, learn, and implement refinements before shipping. But film and television work in the opposite direction. Because shooting time is massively more expensive than the work preceding it (and often finite for other reasons, like location availability and actors’ schedules) you want maximal preparation. That’s why you might spend years writing and rewriting a script, followed by casting, scouting, storyboarding, prepping every detail of production… and then shoot the whole thing in a few weeks.

In The Expanse, we sought a sweet spot of iteration on key gameplay features like 0-G navigation, while bringing an unprecedented level of preparation to all of the cinematic content. In our next project, we’d have to leverage every ounce of both to have any hope of shipping the game.

Life is Strange: Reunion

Unfortunately, the two-project plan for sustained continuity did not last beyond the initial split. Neither The Expanse nor Life is Strange: Double Exposure were as financially successful as our previous two games. Even worse, the industry had just entered a period of prolonged contraction.

There is a lot that happened between the end of The Expanse and the beginning of Life is Strange: Reunion, and a lot that didn’t happen. But like so many other game studios all playing the world’s shittiest game of musical chairs, we were left with our butts in the wind when the music stopped.

Layoffs followed. Work was halted. Excellent devs were let go. It was hard to see what the future looked like—for us, for our partners at Square Enix, for Life is Strange as a franchise. Except… there was still a portion of the remaining budget that Square had already allocated to develop the next Life is Strange game. And an opportunity, however small, to turn it toward a new concept that could become our next project.

Story Concepting: The first story outline for what would become Reunion was written in a travel notebook that I keep in my backpack. I wrote it while flying to Colorado just days after learning that we were going to pitch a new story to Square.

In truth, I had never stopped thinking about Life is Strange, even when I was working on The Expanse. And now we had the benefit of new collaborators from that team to bring fresh perspectives to the franchise. We met for a weeklong intensive story breaking session followed by a pitch to Square and a series of meetings to refine the concept.

Thanks to many new and better ideas from our story work with Square, the concept would change significantly from that initial outline. But when I look at it now, I am amazed at how many core pillars made it from that flight to the final game.

Team Makeup: 2 writers. I wrote roughly half the game while co-directing the project along with the Design Director. The writer who we hired for The Expanse now took on much greater responsibility as Senior Writer, writing half the game and sharing a ton of the story leadership responsibilities.

Pre-Production: After our experience writing The Expanse, we could now confidently say that writing the full script during pre-production was the best approach for our pipeline… and that we couldn’t do it with the schedule we had.

But we did leverage many learnings from The Expanse, including scoping. Every single creative decision was measured by production and the content teams and held up against the seemingly impossible deadline and budget. It was painful, like a never ending audit. It was demoralizing. And it was absolutely critical.

Because we were reconcepting from a previous version, we needed to utilize everything we possibly could from what had already been produced: repurposing partially built characters, finding new uses for environments, and refining Max’s rewind mechanic. This became the ethos for our story work: leverage every pixel that’s been produced already, be judicious about anything that hasn’t, and invent like crazy within those boundaries.

Story Breaking: The game directors and writer spent our holiday break working on the story each day which we then quickly developed into a Narrative Arc.

One of the challenges on previous projects was how soon the Narrative Arc would be made obsolete by subsequent story changes, requiring either constant revisions or simply abandoning the document during production. But thanks to how fast we had to move on Reunion, this was far less of an issue. We maybe revised the Narrative Arc twice during production, and each time it stood up for months as an accurate projection of the game we were building.

It would be emblematic of a notable feature of Reunion’s breakneck development, and the reason the game ended up shipping at all: tremendous efficiency.

Content Pipeline: Take script locking. On previous games, we might have multiple rounds of back and forths with various stakeholders for every scene in the scripting process. One round to pitch a new idea to the Narrative Director and refine it to their notes, then another round to surface it to the Game Director and refine it further, then to disseminate it to content teams who may throw a wrench that requires a story meeting for us to reconsider and develop a new plan, all before floating it to the publisher, receiving and implement their notes, re-disseminating the new version, working through any disagreements, maybe then getting notes from a different team on the publisher side leading us to scrap the intitial idea and try something new at the 11th hour… and so on.

On Reunion, we had virtually none of that. There were no added layers of Narrative Director or Game Director—I had either written the script myself or worked closely with the writer who did. We then shared the scripts with the team during the lock stage to generate their feedback and worked with an executive at Square, a writer and savvy storyteller himself, who consolidated the feedback on their side. Then we carried all those notes into a final revision before locking and… that was it.

Every team on Reunion worked just this way, leveraging their years of experience and trust in one another to cut out layers of review, back and forth, director/publisher intercession, etc., and just deliver at tremendous efficiency. We were blessed with an unprecedented level of talent on both the developer and publisher sides and we required every ounce of it.

Of course there is a downside to that way of working—less iteration, less oversight, no safety net—but the upside is that we were able to ship the game in seemingly impossible circumstances without killing the team.

Take Aways: In my post after Reunion’s release I described it as a game about miracles that is itself a miracle. It is a sad irony, though perhaps fitting, that the miracle could only occur one time.

In a different world, the Reunion pipeline could have been a blueprint for how to overcome the horrid circumstances of the current game industry by leveraging the efficiency of a small and talented team possessed of great tools and institutional knowledge. But I will always be grateful to the game we have to show for it, a fitting send off.

Add Ons

• Though I singled out Cara because of the unique circumstances on The Expanse, I should mention that we were similarly blessed with incredible leads in Erika Mori, Rihanna DeVries, and Hannah Telle. It’s hard to overstate how much our entire productions rested on their shoulders and all of them turned out to be not just brilliant talents but ideal coworkers and collaborators as well.

• I’m almost glad I had this Part 2 to write so I avoided diving into the Mixtape discourse. Lost in all the silly nonsense that is just rinse and repeat from so many narrative games —“is it even a game tho???”—is some really thoughtful critique. No surprise that my friend Robert was among those, check it out:

  Mixtape, Mixed Feelings