Take the Reins

Business Administration, Management

September 29th, 2026

“Take The Reins”

Before talking to Claude at all, I brainstormed the kind of game I wanted to make. I don't have much experience playing video games, so I looked instead at board games I enjoy and hobbies I could translate into a small arcade setting. Horseback riding came to mind immediately since I've competed my whole life, and a horse-jumping game used to be my favorite thing to play on my mom's phone as a kid. I wanted the game to be about the feeling that's almost impossible to explain to someone who hasn't ridden: the last few strides before a jump, when you're measuring your distance to speed as well as how your own balance can throw you off completely. That's what I wanted the game to emulate. Adrenaline, a little stress, and the specific satisfaction of getting a distance right, the same way real riding felt to me as a kid. Claude suggested the name of the game to be “The Spot”, which I changed to “Take the Reins” (my own idea).

After the first version, I honestly thought about changing topics completely and went back to the drawing board. The tracking system was so imprecise that I couldn't accurately judge distance to the jump, and I was constantly hitting poles and racking up faults. There was also no ending as the only choices were "next course" or "restart career". The setting never changed, just the same blue sky and green grass every time. It didn't feel like a real competition, and it wasn't fun enough to make me want to keep playing, let alone anyone else.

So I went back with specific fixes rooted in what actually felt wrong. I asked for a real countdown mechanism so the approach to each jump was readable instead of a guess. I asked how many courses the game ran and how a win condition could be implemented. I also pushed for an actual ending, as well as taking the duration from unlimited to 3 courses each round. I asked for fake competitors so there was something to actually beat and for the setting to change per course. And like real competitions, I asked for ribbons to be awarded, so a clean round meant something and left the player wanting to come back. Where the collaboration was real: when I asked for "fake people" and "new settings," Claude gave me actual named competitors and named locations: Sunny Schooling Ring, Rainy Schooling Show, Indoor Arena, which were details that felt oddly nostalgic and more specific than what I'd asked for. Where it wasn't: the entire first build. It technically worked but completely missed the point, and no amount of prompting could have fixed that without me actually playing it, feeling that it was boring, and pushing back using real human experience as a sounding board.

I would not say I authored this game. I didn't write a line of code. But I don't think I just gave feedback on someone else's work either as the ideas that made the game good were mine, rooted in twenty years of actually knowing what this feels like from horseback, and the first version proves Claude couldn't have gotten there alone.

Some of my real prompts:

●​ "i dont love this game - i keep knocking poles and im not sure theres a good story behind it. im trying to brainstorm how to make it better or another game to come up with"

●​ "For starters, make the lead up to jump more like a countdown, i couldnt tell when to jump and knocked the poles every time"

●​ "how many courses does it run and is there a win? Also the setting should change between courses (inside/outside/rainy/etc)"

AI Disclosure:

I used both Claude Code and Claude Desktop in the making of this assignment. I started by using Claude Desktop to help brainstorm and as a soundboard for some of the ideas I had for games and how many lines of code roughly they would require so that I could stay within the limits of the assignment. I used Claude Code to write the code and create the game, with roughly four iterations (per the below screenshot). The idea itself to make a horseback riding jumping game was mine, and each iteration was rooted in my feelings while testing the rounds. I also gave Claude Desktop the original form of my Artist Statement, with the prompt: “Please check for grammar and run a word count on this, suggesting repetitive areas that can be cut down without changing any of my language”. Since I was still below the word count, I simply aimed to reduce repetition between my fourth and fifth paragraphs and then fixed some punctuation issues that Claude pointed out.