I’m Building a Game Backwards: The World Already Exists, So I’m Not Building It
I’ve been building a small experiment called River Racing League, and it has pushed me toward a different way of thinking about what actually needs to be “built.”
Instead of creating a digital world, programming its physics and deciding how objects behave, I start with a real river.
I put racers into it. People pick who they think will win. Then the river does the computation.
Current, rocks, obstacles, collisions, slow water, getting stuck, breaking free and random paths determine the actual outcome.
Here’s what the current version looks like:
https://www.youtube.com/shorts/Rm8pP2A97bk
Right now it’s deliberately primitive because I’m trying to avoid building complexity before I know what people actually respond to.
The build path I’m testing is basically:
real physical event → audience prediction → measure attention/interaction → add only the digital layers that prove useful
Eventually that could mean persistent racers, scoring, statistics, seasons, computer vision, AI commentary, predictions and other game systems.
But I don’t want to build all of that first.
The interesting question for me is whether the physical world can supply enough complexity that software only needs to organize and enhance the experience around it.
So this is genuinely being built in public. The simple version is running now, and what people actually respond to will influence what I build on top of it next.
I’d be interested in what other builders would add first—and what you’d deliberately leave out.