Game #3 of my one game a month challenge was going to be an endless runner, but I tripped at the first hurdle.

Here's how it went down:

Week 1: Pride Before a Fall

My first idea was Flappy Bird, but it felt too simple after Loco Lift Rush. I wanted it to be more impressive, and that's where I strayed from the true goal of the challenge: practice finishing, not impressing.

Next I tried an endless Frogger concept. Jumping into prototyping the procedural generation of an endless stream of roads and rivers was a fun distraction from the niggling feeling of dissatisfaction with an unoriginal idea.

After falling down various procgen rabbit holes, I forced myself to reset and come up with a plan and a schedule - the lessons I was supposed to have learned from participating in Pyweek.

Week 2: Scope Creep

While writing up the GDD, I decided that Frogger was too staccato. Usually in an endless runner the player is constantly moving forwards, but Frogger forces the player to stop and wait for a gap in the traffic or for a log to float by. Sometimes they might even have to go backwards.

This is how I landed on Eterniski. The player can only go downhill, and the only reason to stop is crashing into a tree. Going through consecutive slalom gates builds score multipliers. I liked this idea and it brought back fond memories of Horace Goes Skiing (which coincidentally starts with a Frogger level).

But then I wanted the controls to feel right. I wanted the satisfaction of long carving turns and quick slaloms. And suddenly I was in unfamiliar territory.

I realised the limitation of my control system, which only handled digital input (on/off) and not analog input (like joystick tilt). I spent some time refactoring my control system to handle this, but by this point I was eating into the time I had scheduled for composing the soundtrack.

Week 3: Procrastination

Music is daunting for me. With no background in theory, I've only ever dabbled. I procrastinated on the last game's soundtrack, and it suffered. I had made it a priority this time, but here I was again, doing anything but composing. I put a few notes together in Bosca Coeil, in the delusion that messing about would lead to a soundtrack. But it didn't. I started looking at tutorials on Strudel, thinking that I might have more luck "coding" music, but that was just more procrastination.

By the end of week 3 I was behind schedule and demoralised. I had a stub of a game that I didn't feel excited about, and didn't have enough time to make it better and learn how to make a soundtrack too, so I started avoiding working on it at all. I wrote a post on the control system instead, and even ended up working on my day-job in my spare time, which I try very hard not to do as a rule.

Week 4: Admitting Defeat

So here we are, on the last day of the month, and I have to admit defeat.

I've published the incomplete game on Github and you can play it here:

Play Eterniski

Lessons Learned

1. I need to finish, not impress. My ego was still buzzing from Pyweek, and I wanted to make something that stood out. I dismissed Flappy Bird as too simple, but that simplicity would have let me spend more time on finishing the game.

2. Don't let prototyping become procrastination. Jumping into code feels like progress. It's a developer's comfort-zone. But it is a trap. Messing about with procedural generation was fun, but it didn't get me any closer to a finished game.

3. Just enough process. I hate to admit it, but I need a plan to stay on track. A schedule and a GDD are the guardrails that keep me from driving off the cliff. So I'm making templates that I can fill in first thing each month.

4. I need to practice music. I have decades of coding experience, I've been drawing pictures since I was a child. Music composition experience? A few hours. I have to start being deliberate about learning.

Pride comes before a fall. This time I wiped out, but I've picked myself up and I'll be back next month with a finished game.