Posted December 09, 2025 by Twinbeard
A couple of players asked me about my creative process, so decided to get into it a bit here. Consider this a companion piece to the It 2 dev log, which goes into more technical detail.
Thematically, the game came together fairly quickly. I chose "It" as source material mostly because I liked the joke of inverting the "clown in the storm drain" trope by having all seven kids poking their heads out of the storm drain and alarming passers by. I started skimming TV Tropes for joke and gameplay ideas, and I hit on the idea of, if It feeds on people's fear, other monsters might feed on other emotions such as awkwardness. The new monster had to have a pronoun for a name -- in my headcanon there are a whole slew of these creatures that each have different pronoun names and feed on different emotions. And of course the game takes place in the aftermath of a breakup, with the quest of recovering It's record collection, because I think the idea of inserting Lovecraftian horrors into mundane situations is funny. It's a gag I've done before, e.g. in SPAXRIS.
Gameplay was a bit harder to come by. My game design process is to take a game idea that seems like it'll work, start implementing it, then while I'm implementing it I usually have a better idea and start doing that instead. Usually this keeps happening, and usually the ideas are related enough that the work so far isn't wasted. This process works for me most of the time! To start off, though, I need what seems like a good idea. (Although honestly I don't know if this is a psychological barrier. Like, if I started making a dull game idea, would it also lead to a better idea I could pivot to?)
I ended up thinking about gameplay for most of October, mostly because I was really excited about PiCoSteveMo. I've got the personal constraint that I always try to follow, of not including combat in my games. (I've got ethical reasons for this, but also, the world doesn't need more combat-oriented games, because we have so many of them. It's been done so much. It's boring.) I wanted to do an open world puzzle game, like Look Who's The Shining Two. I started trying to put together a set of complementary abilities for the seven protagonists, and a set of puzzle mechanics that they might interact well with. I've got a couple pages of notes like this:
Kid with bicycle, kid who smokes, kid who is very tall, you get the idea. I had an idea for an analog sokoban where you're pushing around balloons that aren't tied to a grid, with realistic physics. I had an idea for a game where you're managing water pressure levels by snapping together tiles like in Pipe Dream. The design problem I kept coming back to was how to let the player control seven protagonists without navigation becoming a chore, making them select and move around seven characters individually. I ended up pulling in a movement mechanic from the Frog Fractions Hat DLC, where two characters navigate a tile-based digging minigame together.
The group follows whoever's in the lead, and gravity affects the entire group, which can take various shapes based on who's moved where. I prototyped the idea with seven characters and it felt much different from two, much more interesting, so I decided to run with it -- still not totally sure what these characters would actually be doing other than moving around the space. One player described this game as on the continuum between Snakebird and Lost Vikings, which I think is apt, but neither one was a direct inspiration. Over on the Lost Vikings side of things, my direct inspiration was the Berenstain Bears story about the Spooky Old Tree, where three kids go an adventure: one with a light, one with a stick, one with a rope.
The process I've hit on for level design is to implement a game feature, then add some level design to the map that tutorializes or takes advantage of that feature, then repeat. This has the advantage that areas are naturally themed by mechanic. I built the top half of the map in roughly the order that you collect the protagonists, and I added each protagonist's special ability in roughly the order that I thought of them. I put balloons in the world and implemented an interaction where you can push them around or pop them basically just because it made sense.
Thinking about the kind of puzzle gameplay I was envisioning, I knew I needed a way to reset the game state. In a combat-driven game, there are a lot of examples to pull from, but typically the world resets to a checkpoint when you lose a fight. This game has no death, no lose state, and I needed to make sure you couldn't get stuck. (When I was considering the "analog sokoban with balloons" idea, I was going to take the scene where balloons pour out of a minifridge in the It miniseries, and add a "reset fridge" to each puzzle.) I had to address the same issue in Look Who's The Shining Two, which used the "rewind yourself" and "rewind the world" whiskeys to serve that function. Similarly to that game, I wanted these abilities to have puzzle solving function and not just a utility function, so I started thinking about how to design puzzles around them. I also knew I had to give you the "reset room" and "return to checkpoint" abilities first, and early on, because the puzzle design before you got them had to be very constrained to prevent soft locks. After that, character abilities could flow more loosely and freely.
For the digging guy, I pulled in the water flowing mechanic from the Hat DLC digging minigame, thinking I could expand it into puzzles where you're kiting huge pools of water across the entire world to solve puzzles. This turned out to be over-ambitious from both a technical and a design standpoint, and I quickly scoped it down to water only flowing within the current room. For the swimming mechanic, where everyone floats to the surface unless the designated swimmer is part of your group, I accidentally stumbled on the thing where you can brace yourself against the ceiling to force the group partially underwater, and leaned into it to make a few puzzles.
After designing the part of the game where you collect all the protagonists, I wasn't sure how to proceed. In my previous Zelda-inspired open world games, Gordy and the Monster Moon and Look Who's The Shining Two, the player has a lot of leeway in the order they collect new abilities, and they collect new abilities over the course of the entire game. This game, possibly because of the nature of the abilities or possibly because the size of your group constrains movement a lot, I had a hard time coming up with a level design that would support collecting your little guys in any order, and I also had a hard time imagining how to design areas that supported differently sized groups or different subsets of abilities.
Maybe I could've done it with some more thought and iteration, but in practice I ended up with a first half of the game where you're collecting guys, and then a second half of the game, where the puzzles get tougher and demand splitting up your group, more interaction between abilities, etc. Playtesters didn't seem to feel railroaded by the fixed order of things you have to do in the first half of the game, so I think it worked out fine, but it's not the game I set out intending to make.
Before I set out designing the second half of the game, I took a couple of loose mechanics that were just floating around -- pushing around and popping balloons, and poking your head out of storm drains -- and formalized them into the structure of the game, with doors that only open once you've popped a certain number of balloons, and unlocking the ending when you've reached a certain number of storm drains.
I started making a list of the kinds of puzzles and ability combinations I was missing, partly based on playtester feedback and partly just thinking about combining the various mechanics. I started building those out, and I also started expanding my thoughts about level design past pure puzzles, into what kind of experiences make going on an adventure interesting, e.g. climbing a sheer cliff by digging into its surface, or digging into a stalactite of dirt suspended from the ceiling, peering out and seeing nothingness in all directions, and taking a leap of faith anyway, or finding a toilet at the bottom of an underground reservoir and choosing to flush yourself down it. Exploring the sewers beneath your town should feel like an adventure.
Around this time I realized that my HUD layout didn't really make sense. One odd thing about making room-based games in Pico-8 is that screen space is tight enough that you have to lock in your HUD shape before you do your level design, so you know how big rooms can be. In Gordy, rooms were 16x14 tiles because I allocated the top and bottom rows for the HUD. In Shining 2, rooms were 15x15 tiles because I'd allocated the bottom row and the right column for the HUD. I borrowed the latter design for this game, and then, deep into development, realized that the right column isn't really adding much. The "right" thing to do would've been to redesign the hud, then re-lay out all the rooms, but I didn't have that kind of time, and I didn't have any great ideas for what to put in that column either, so I just let it be.
Once the game was playable start to finish I started running more playtests. I think the process of running a proper playtest is more widely understood now than it was ten or twenty years ago, but the short version is you have to sit a player down, you ask them for a running monologue about everything they're thinking and feeling -- some people are better at this than others -- and you tell them to ask questions, but that you can't answer any of them until the end. Then you sit back, shut up, and take notes. You don't explain what's going on if they seem stuck or confused, you just suffer. Players out in the wild won't have you explaining how to play over their shoulder. Your game needs to stand on its own. Think of the suffering as motivation to get out there and fix the confusing parts.
I suppose I should talk about art somewhere in here. My pixel art is much better than it was a few years ago, and one thing I can tell you for sure is that it's not because I've practiced a lot. It comes almost entirely from adopting the following process:
In Pico-8 you have a fixed, well-considered palette, so you don't need to worry that much about color. You're also mostly working with 8x8 tiles so getting roughly the right shape and then trying various tweaks almost at random until it looks good is extremely doable. There just aren't that many things you can try.
I had started writing the dialog early on. I enjoy writing dialog trees, so it's a fun motivational bit of work for me. I had written almost all the dialog in the game halfway through development, and I revisited it over the second half of development, polishing it, tweaking word choices, improving jokes, removing jokes that weren't good enough to justify the space they took up. I only hit on the final twist when it came to write the dialog tree for the ending and I realized I didn't really have an ending. "You got the record collection back, yay!" felt like it'd be a letdown. I took a couple days to think about it and ended up with an idea that seemed in-character for everyone, funny and horrifying at the same time.
The secret ending came very late. I had a bunch of unexpected room on the cartridge for additional art and I chose to spend it here. Not to spoil it, but it seemed like a satisfying reward for popping every balloon, and like a silly, simple way to subvert the original work that would make people happy to find.
I don't really have a conclusion here because this post is just a list of stuff, but I guess the common thread is: just get stuff out there and see how it works, then iterate?