Since January, most of my time on Students Against Spooky Stuff hasn't gone into Unreal at all. It's gone into research and writing, and I want to walk through why, because it was less a single decision and more something I kept re-confirming was right every week I chose not to open the engine.
Where this started
Back in January, my first instinct was to dive straight into Unreal and start blocking something out. I opened it a handful of times, poked at a few tutorials, and then made myself close it again. Every time, the same thought stopped me: I didn't actually know what I was building yet. I had a premise and a vibe, not a story. Building on top of that felt like pouring a foundation before I'd decided what the house looked like.
I don't have a technical background. I want to be upfront about that, because it shapes everything else in this post. What I do have is years of watching stories land, or not land, with real players, back when I was community managing instead of making games. I know what it feels like when a player finishes reading something and actually cares. I have almost no instinct yet for what it feels like when a level finishes loading and actually works. Those are different muscles, and trying to build both of them during the same tired hour every night was never going to end well.
Leading with what I know
So instead of forcing myself to learn engine systems and narrative design at the same time, I did the thing that actually plays to my strengths first: I wrote Students Against Spooky Stuff like a book.
Every chapter, starting with "The Slushie Specter of Stop N Sip," now exists as a full short story before it exists as a level. Characters, dialogue, the shape of the scare, the joke that undercuts it right after. All of it on the page first, in a document, with no engine open at all. Some sessions I wrote for two hours straight and never touched Unreal, which a few months ago would have felt like I was falling behind. Now it feels like the opposite.
Why this order matters to me
If a chapter isn't fun to read, it's not going to be fun to play, and finding that out in a document is a lot cheaper than finding it out after a week of blocking out geometry. I already saw a small version of this play out, back when I sketched a level layout on paper around a joke that read fine in my head and fell completely flat the moment I actually wrote the dialogue down and read it back. Cutting a paragraph costs me twenty minutes. Cutting a level would cost me a week. That math alone justified the switch.
Writing first also means that when I do sit back down in the engine, I'm not making things up as I go. I'm building toward something that already works on paper, with beats I've already tested against myself as a reader. The uncertainty gets front-loaded into the cheapest possible medium, which is just words on a page, instead of getting discovered late, in the expensive medium, which is a half-built level I now have to tear up.
It's also just honest about where I am. I'm still learning 3D basics and still learning Unreal, and that's not going to change by April. Leaning into the writing while that catches up isn't a shortcut, it's using the tool I already have instead of pretending I don't have a gap in the other one. I think there's a version of me that would have felt embarrassed about that, like writing first is a workaround for not being a "real" technical dev yet. I've mostly let that feeling go. Every solo dev is compensating for something. This is just what I'm compensating with.
What the process actually looks like
For anyone curious about the mechanics: I start with a single paragraph describing the feeling I want the chapter to leave someone with, before I write a single line of dialogue. For Slushie Specter that paragraph was about a place that's supposed to be boring, a gas station at 2am, turning out to hide something both stupid and genuinely unsettling at the same time. Everything else gets built against that paragraph. If a scene doesn't serve the feeling, it gets cut, even if it's funny, even if I like it.
Then I write the whole chapter start to finish like a short story, prose and dialogue together, no formatting for engine purposes at all. Only after that draft feels right do I go back through and mark up which beats are dialogue, which are environmental, which need a specific mechanic to land. That second pass is basically translation work, taking something that already works as a story and figuring out what it needs to work as a space someone walks through.
What's next
I've got three chapters drafted this way and a fourth outlined. Once "Slushie Specter" is locked as a finished piece of writing, it moves into the engine for the first time as a real vertical slice, built against a script instead of a vague feeling. Book first, game second, in that order, for as long as this keeps working. If it stops working, I'll write about that too.
- Mikey