REP COUNTER

The Art Of The Game Statement

10/06/2026

I wanted the game to reflect my challenges as a track athlete and my original idea was a day-in-the-life game where you juggle practice, class, treatment, and sleep. Every choice spends a limited energy bar, and the "win" is surviving the week with nothing completely failing but it was too complex in terms of the activity itself and how i felt it truly captured my student athlete experience so i decided to go with one challenging aspect of my athletic life which led me to rep counter in which this one specific exercise at lift be challenging me, single leg lunges holding two dumbells, it burns half way and to the end i have to hold form and push else i wouldnt get everything i can out of the exercise I started with a bigger idea, a day-in-the-life game where you juggle practice, class, treatment, and sleep on a limited energy bar, and the win is surviving the week with nothing completely failing. It was honest, but too complex, and I felt it didn't capture my experience as a Division I track athlete. So I picked one challenge from my training and built the game around it: single-leg lunges holding two dumbbells. Eight reps, three sets. It burns halfway through, and the point of the exercise is holding form and pushing through that burn. If I don't, I don't get everything I can out of the work.

That's the meaning underneath. Being intentional in every action matters all day, from practice to class to the weight room, and the last rep is where it's tested. It has to be patient and as consistent as the first one, not rushed because I want it over with.

Design decisions I defend

The shrinking timing window. When fatigue hits, the window to land a rep gets smaller, so staying precise gets harder exactly when the burn does. A bad rep makes the figure shake and lose form, which feels unstable, because that's what it feels like when I lose control mid-lunge.

The last rep changes rhythm. The final rep breaks the pattern and demands a patient extra push, which stands in for the rising pain across the set.

The setting. The game takes place outdoors, in bright colors, to reflect the island life I come from in the Caribbean, where finding innovative ways to work out is part of training. Where the collaboration was real, and where it wasn't

Claude gave me a playable game quickly, but it needed real correction. An early version gave the character only one leg, which I had to fix. The timing cue was nearly impossible to hit, so I added a visual indicator to make it doable. Then I went the other way and made it harder, thinning the ring, shrinking the timing window, and making the ring close faster. I did that about four times so the game stayed challenging. The cues make it slightly easier to know

what's next, but I think they make it less confusing. [Add one concrete example of something Claude did better than you asked.]

What I authored

I didn't write the code but i added, took away and edited the code. I decided what the game is about, what it should feel like to lose, and how hard it should be. Claude can only build what I can describe, and the description came from my own body and routine.

Prompts that guided development

1.​ the character has one leg, make it a lunge and give the last rep a cue"

2.​ "make the ring thinner and more challenging by like 20% as the reps and sets progress"

3.​ "make it thinner and increase the speed of the ring closing"

4.​ "let the ring close in faster on the second set and even faster on the last set and make the last rep of each set have the ring close in way fater"

5.​ "faster"

I left out "give me the link to attempt", since that was a request for the link and not a change to the game. Fix the "fater" typo in prompt 5 if you want clean quotes, or add [sic]. AI Disclosure (draft)

I used Claude in the claude.ai chat interface to build REP COUNTER as a single HTML file. The first prompt described the full game. I then went through five rounds of changes: ●​ The figure rendered with one visible leg, so I asked for a lunge, which gave it two legs. In the same round I asked for a visual cue on the last rep, because the audio pulse drops away there.

●​ I made the timing ring thinner and the windows tighter in two steps. The first was a gradual 20% tightening over the round, and the second was a further 15% reduction. ●​ I sped up the ring's closing animation three times. It started the same speed in every set, then became faster in each later set, with an even faster close on rep 8, then faster again.

I did not write the code by hand. I tested the game myself and decided what to change. [Add anything you abandoned, any other tools you used, and what you tested yourself.] I changed these from your draft:

1.​ Claude Code and Desktop: this was the web chat, so I removed "Claude Code to build

the game". Add it back only if you really used it elsewhere.

2.​ "Add one leg": it was a fix that turned the squat into a lunge with two legs, so I reworded it.

3.​ The cue: you described "a timing indicator because the cue was too hard to hit". In

this chat the approaching ring existed from the first build, and the new cue was added to rep 8 only.

4.​ "About four times": the windows changed twice and the ring speed three times, so I split them.

5.​ Round count: I counted five revision rounds after the first build. If you iterated more elsewhere, update the number.