Here's my long overdue post mortem.
My game: Blank Page
Final results (Compo)
Overall: 15th (4.1 average from 47 ratings)
Fun: 38th (3.9 average from 47 ratings)
Innovation: 1st (4.656 average from 47 ratings)
Theme: 2nd (4.567 average from 47 ratings)
Graphics: 287th (3.2 average from 47 ratings)
Mood: 221st (3.226 average from 44 ratings)
I'm very happy about this, especially considering this was my first game jam.
In this post I'll mainly cover how I did ideation, design and development during the compo.
Before the compo
This was a good time of the year to have a compo because it was scheduled to start 8am local time at the beginning of a three day long weekend. I didn't schedule anything else that weekend.
I wanted my game to be web-first (playable in a web browser) and ideally be written in TypeScript because I like the type safety. I was considering using Phaser for the game engine.
The compo
At the start of the compo I began brainstorming ideas. Before long, I realised that the theme "you start with nothing" was applicable to nearly all games, so it was very open-ended. You could build a Minecraft/Factorio spinoff or some RPG and that would technically satisfy the theme. But I didn't want to build a regular game. Ideally, it should be something with a stronger-than-normal emphasis on starting with nothing.
I liked the idea of starting with literally nothing: a completely empty webpage. No UI and no instructions. Nothing. This reminded me of games such as A Dark Room and Candy Box 2 where you start with a minimal UI, so I decided to go in that direction. I also liked Candy Box 2's charming ascii art, so decided to do that too. It was also easier for me to draw because I hadn't practiced creating my own game art in ages.
But what should happen next? I spent a long time trying to think of ideas. It was like the "blank sheet of paper" problem described in the rules page. I found that linked Wikipedia page relevant, so I made it the next step in the game progression. From there I brainstormed ways to manipulate the characters to make stuff happen and unlock more things. It was turning into a very weird and abstract game.
But an overly weird game can be annoying or even unplayable. I knew I could make it less weird by giving it internal consistency. I figured that a minimalistic, grid-based puzzle game would be suitable for this. So it became a game where you start with with knowing nothing of what it's about or even how to play. You had to figure it out. Additionally, I liked the idea of exploration: starting with nothing visible, then that small part of the page, then gradually uncovering more of the map. (More adherence to the theme.)
It was almost three hours in when I began setting up the project. By then I had decided that I didn't need a graphics library. It was enough to use just plain HTML, CSS and JavaScript (but transpiled from TypeScript).
From there (and throughout the rest of the 48 hours), I tried thinking of game mechanics – ways to explore, ways to progress, things to unlock and things that would add challenge to a puzzle game. First I thought of the click-to-remove letters and the expanding dot (.). Then the expanding +. Then the "beacons" (those numbers you need to light up), which were useful for dividing the map into separate challenges (or levels).
Due to the exploratory aspect of the game, I challenged myself to fit the entire thing on one grid. I also wanted to make it impossible to get into an unwinnable state. I added "powers", initially as a way to undo moves. Later, I made it a design rule that each (non-passive) power should add some new possibility to the existing space. For example, the slider (@) makes you use the walls differently.
I found it quite difficult to think of suitable new mechanics or powers. The slider idea happened at the nine hour mark. Twelve hours in, I came up with the idea of an "anti-beacon" (0) that you had to avoid, but I didn't have many ideas for new mechanics after that. Towards the end I thought of a weird tower defence idea, but didn't have enough time to implement it. [See Appendix 1]
Apart from coming up with ideas, I spent most of the time writing code and doing level design. My file structure was a mess. I didn't modularise it very well; one of the files was 700 lines of code + hacks. But it had types and only needed to be maintainable for two days. Level design took a lot of effort too. Due to my constraints of one map and no unwinnable states, the map became very complicated to design. I wanted the progression to be increasing in difficulty, and had to be careful to not make parts unintentionally skippable. I found it helpful to make a separate annotated version of the map for my own reference.
Regarding progression, I wanted to introduce the player to new mechanics gradually. In accordance with the theme, I aimed to do it with minimal UI – without an explicit tutorial or instruction manual on how to play. (I was inspired by certain games that did this well. [See Appendix 2]) Even the entry page was deliberately terse.
Overall, it was an interesting challenge that encouraged me to push my limits. I had some doubts along the way but managed to produce something playable and complete enough to be a game.
Thanks to everyone who gave my game a chance. And thank you for reading.
Appendix
1: Crazy tower defence / race idea:
* Each race has a endpoint, which is a tile on the map. There are also several start points for allied and/or enemy racers.
* The player is able to prepare the map (by manipulating walls) and then start the race.
* When the race starts, all the racers attempt to reach the endpoint via the shortest path clear of walls.
* So the player should prepare by arranging walls in a way that gives allied racers an advantage and enemy racers a disadvantage.
2: A game that's good at conveying the rules with less words: Mega Man X
https://youtu.be/8FpigqfcvlM
https://evepto.itch.io/the-slime-slayer
Orange highlights categories which we placed top 100 or better. Medals show top 3 status. Our best game yet was 
This line chart shows raw scores out of 5 per category per game. We tend to not focus on humor, resulting in some low scores there.
When taking into account the total participants, the results are more insightful. Our game 
