(This post is adapted from my micro talk "Teaching and Rethinking Level Design" at the GDC 2025 Educators Soapbox session. That's why it mentions "students" in the slide above.)
People want to do level design. They grow up playing games like Minecraft, Roblox, Fortnite — all 3D games with 3D worlds. And to create 3D worlds, supposedly you need this thing called “level design.” Then when you search YouTube, you'll be told that level design is about implanting secret lines that "guide the player" into walking down hallways. Such is the power of ARCHITECTURE!
But this is not how architecture works, nor how level design works. Imaginary invisible shapes cannot mind control players, and even if they could, no one needs to be mind controlled to walk down a hallway.
No one plays games like this, but why do we think we do? What's going on here?
It starts with a shred of truth: level design often has architectural aspects, and architecture can influence behavior.
Our main duties as working level designers often involve making game spaces by drawing layouts and building blockouts (simple rough draft 3D levels), which both resemble floor plans and architectural models. This also reflects the basic division of labor in most game studios, where level designers design and environment artists decorate.
So from there, it's reasonable to assume that good level design requires an architectural understanding. We're doing so much work that looks like architecture. A better layout drawing and a better blockout model = a better level?
Here the typical level designer skill progression tends to follow this viral meme image below:
Here the (supposedly) BEGINNER level designer spams disorganized shapes everywhere, while the (supposedly) INTERMEDIATE level designer organizes intentionally but adheres to an axis-aligned grid and room-hallway-room pattern. Then the (supposedly) ADVANCED level designer adopts minimalism and elegance, making efficient use of walls and verticality, to plan spaces that flow into each other. And so on.
This popular meme image represents a very real phenomenon in level design culture: ideally, your level resembles a plausible real-life building.
I know this because, dear reader, I am the one who made this meme.
But I now regret making it. Even if it's real, it's wrong. It's like some sort of secret imaginary invisible line that's mind controlling us, misleading and misguiding us along the wrong corridor...
For instance, the so-called “BEGINNER” level design works great in Stalker or Fortnite, where a systemic chaotic world matters more than individual room design. Maybe less organization facilitates more improvisation, like this gameplay is operating at a different scale.
The supposed “INTERMEDIATE” level design is great for Fallout vaults, which adopt boxy grids for a clean retro bunker kit aesthetic. If you want a space to feel more mechanical, more prefabricated, then a visible grid helps bring out that modular vibe.
And the “ADVANCED” level design makes sense for Halo / Destiny multiplayer maps with sleek elegant sci-fi spaces that push constant smooth flowy movement. (Conversely, I sort of hate these layouts in something like Stalker, the clean chokepoints often expose the AI's limitations.)
Don't let architectural envy do your thinking for you. Organization, realism, and minimalism are just tools. Instead, think critically about what your specific game experience needs, and design for that.
Because that's the true material that level designers are actually working with: game experience, possibility, and space. Not floors or walls, but concepts and ideas and behaviors.
Why must level designers stop at doing bad architecture? We can also do bad economics, bad military tactics, bad wedding planning, and bad comedy! In levels, this looks like planning out quest arcs, pacing setpiece beats, balancing spreadsheets, and choreographing battles.
These non-architectural level design tools are often more effective than architecture at shaping player behavior.
What sounds more effective for pulling the player to the exit of a non-linear level?...
(a) scripting, pacing, combat, economy: have an NPC brief the player on how to find the level exit, then place a ranged enemy turret, valuable item, and side-quest destination near the level exit
(b) the raw power of architecture: rotate a rock, pray the player approaches that rock from a direction that coincidentally forms a secret imaginary invisible line that mind controls them into hallucinating a path toward one of three skybox mountains in the far distance
 |
| "Ceci n'est pas un niveau" ("This is not a level") is a play on "Ceci n'est pas une pipe" ("This is not a pipe") from the famous painting The Treachery of Images by Rene Magritte. |
I'll call this the architectural fallacy of level design: the over-emphasis of architectural aspects of level design (layouts, blockouts, view composition) over non-architectural aspects of level design (pacing, encounters, economy, scripting, storytelling).
Pulling back from architecture is the best way for level designers to refine our thinking -- and thus, ironically, become better architects.
Because when we emphasize an architectural end product, people seek fast shortcuts to that architectural end product, and then we end up with those "speed level design" videos on YouTube where people race to smash together assets in Unreal. These are fine as art tests or fun asset jams, but don't call it level design! Levels need context. You can't make levels for a game that doesn't exist.
I'm certainly guilty of perpetuating the architectural fallacy myself. I thought I could flush out fake architectural theory with better architectural theory, and I wrote about 100,000+ words about it. Honestly, I still sort of believe that, I still believe in learning from architecture... but maybe just... slowly, carefully... and eventually.
For now, let's step back. A level is a space. What is that space made of?
I don't know anymore. But I know this: space is not a wall.
***
(In my original GDC 2025 talk, this is where I start talking about The Level Design Book and its strengths / weaknesses. I'll save that for a future blog post.)
(For a lovely non-architectural level design talk that's contemporary with this one, and actually grounded in real level design examples, see Steve Lee's GDC 2025 talk -- Level Design Summit: The Unusual Level Design of 'Tactical Breach Wizards', free to watch with no paywall.)
