My Obsession With Personality Tests Became a Game

This is a submission for the June Solstice Game Jam

What I Built

I have a confession. I am one of those people who will take any personality test that crosses my path. MBTI? Done it. Big Five? Done it twice. "Which bread are you?" Also done it, and I was sourdough, if you must know.

You take the test, you get your label, and that is it. You are an introvert. Cool. But am I an introvert who is becoming more social, or an introvert who is settling deeper into my shell? The test does not say. It took a photo and called it a portrait.

I wanted to know which direction I was heading.

That question became Arcus. It is a self-discovery game where you take interactive tests and find out where you stand and which way you are moving. Four tests, each inspired by a June Solstice Game Jam theme.

The Solstice Cycle maps your energy as a cycle with seasons and a direction. Are you gathering momentum or settling into rest? The solstice is the turning point. This test finds yours.

Passage of Time asks which part of time you live in. The past, the present, or the future. It also asks how time itself feels to you: a resource, a weight, a gift, or a mystery.

Modes of Mind is my love letter to Alan Turing. It asks the question he asked: does what a mind says about itself match what it does? You describe how you think, then you make decisions, and the test compares the two. The gap is the fun part.

Spectrum of Self explores how well you know yourself and how honestly you show it. Are you settled in who you are, or still figuring it out? Both are fine. This is Pride, and pride is the courage to be who you are.

The tests are grounded in real psychological research, but Arcus is for fun and self-exploration. It is a curiosity machine with good citations.

Take a test, follow your arc, and see what you discover.

Try it here: https://arcus.pows.workers.dev

Video Demo

The demo video walks through the experience from start to finish.

Code

powxenv / arcus

Welcome to your new TanStack Start app!

Getting Started

To run this application:

bun install
bun --bun run dev

Building For Production

To build this application for production:

bun --bun run build

Testing

This project uses Vitest for testing. You can run the tests with:

bun --bun run test

Styling

This project uses Tailwind CSS for styling.

Removing Tailwind CSS

If you prefer not to use Tailwind CSS:

1. Remove the demo pages in src/routes/demo/
2. Replace the Tailwind import in src/styles.css with your own styles
3. Remove tailwindcss() from the plugins array in vite.config.ts
4. Uninstall the packages: bun install @tailwindcss/vite tailwindcss -D

Deploy to Cloudflare Workers

This project uses the Cloudflare Vite plugin (configured in vite.config.ts) and wrangler.jsonc:

1. Install Wrangler: npm install -g wrangler
2. Authenticate: wrangler login
3. Deploy: npx wrangler deploy

For production env vars, run wrangler secret put MY_VAR for each secret listed in .env.example. Public (non-secret) vars…

View on GitHub

How I Built It

I could not write code and slap "personality test" on it. That is how you get a BuzzFeed quiz with delusions of grandeur. I wanted the numbers to mean something. So I did something unusual for a game jam. I started with research papers.

Finding the foundation

For each theme, I went looking for established measurement instruments in the academic literature. I needed constructs with decades of validation behind them, not vibes.

Russell's circumplex model was a perfect match for the solstice. A circular model of emotions for a cyclical theme. Cacioppo and Petty's Need for Cognition scale measured how much people enjoy hard thinking. Campbell's Self-Concept Clarity scale captured how well people know themselves. Zimbardo's Time Perspective Inventory described how people relate to past, present, and future.

Some ideas did not survive the vetting. Riding's cognitive style model looked great on paper until I found that Riding himself said it cannot be measured through self-report. The guy who invented the thing told me I was measuring it wrong. I replaced it with Epstein's Rational-Experiential Inventory, which was built for self-report from day one.

The template trap

My first version of the suite was gorgeous. Every test had two axes, two facets, four quadrants. Symmetrical. Elegant. I was proud of myself.

It was also wrong.

The Turing test broke first. Epstein showed that Need for Cognition and Faith in Intuition are independent. They correlate at about 0.08, which in research terms means "we have never met." Collapsing them into one bipolar axis meant a mid-range score could mean high in both, low in both, or anything in between. I was destroying what made the measurement work.

The Passage test confirmed the pattern. Its three factors are independent. Forcing them into two axes was like trying to fit a triangle into a line. You lose a dimension and call it simplification.

The template was the disease. Each test needed its own structure. Two axes where the construct was bipolar. Independent scales where it was orthogonal. Multiple factors where it was multi-dimensional. Letting go of my beautiful template made the code harder and the product better. A lesson I keep learning in every project.

Building each test

The Solstice test puts you on Russell's circular model using two axes: Solar Height (how activated you are) and Tidal Direction (outward vs inward). Each axis has two facets with eight items. The quadrant picks your season. But the solstice is about direction, so I added four trajectory items. The result gives you both a season and a direction: waxing, waning, or steady.

The Turing test keeps Need for Cognition and Faith in Intuition on separate scales because they are separate things. The combination picks one of four modes. Then comes the behavioral check: five override problems (the bat and ball, the lily pads, the classics) and ten decision scenarios that classify your actual decision strategy. The gap between what you said and what your choices showed is the Turing signature.

The Pride test mirrors the Solstice structure but measures identity clarity and authenticity. A commitment component asks whether your identity feels claimed or still forming.

The Passage test scores past, present, and future as three independent factors. Crossing a threshold on any combination gives you one of eight temporal types. Then a stance layer asks what time itself feels like. A resource, a weight, a gift, or a mystery.

You can explore each test in detail:

• https://arcus.pows.workers.dev/assessment/solstice
• https://arcus.pows.workers.dev/assessment/turing
• https://arcus.pows.workers.dev/assessment/pride
• https://arcus.pows.workers.dev/assessment/passage

Development process

The frontend is React with TanStack Start. HeroUI handles the components. Scoring and question banks are TypeScript, sourced from the research design docs. The scoring is deterministic and verified against a test suite. The AI reflection uses Google Gemma 4 31B. Users opt in before anything gets sent.

Development was assisted by PI Coding Agent with a GLM Coding Plan subscription and OpenCode Go using the DeepSeek model. The demo video narration was generated with ElevenLabs.

Prize Category

Best Ode to Alan Turing. The Modes of Mind test is not named after Turing for vibes. It is built around his actual idea: the Imitation Game. You say how you think. Your choices show how you decide. The test compares the two and shows you the gap. Can you tell what a mind is by looking at what it does? That is the question Turing asked, and it is the question this test asks you.

Best Google AI Usage. The results page offers an optional interpretation powered by Google Gemma 4 31B. Users opt in before their responses are sent. The AI builds on the deterministic result, the individual responses, and the test context without replacing the scoring. AI output is labeled and sits in its own section.

Why I Built This

I am fascinated by people. In the way where you sit across from a friend at dinner and think: how do you experience this conversation differently from me? How do you carry your energy? How do you decide things? How do you think about who you are?

Those questions sent me down a rabbit hole of psychology research. The deeper I went, the more I found frameworks that were rich, nuanced, and fascinating. The circumplex model. The independence of cognition and intuition. The multi-dimensionality of time perspective. These are not quiz fodder. They are careful, validated ways of describing how humans work.

But when most people encounter personality psychology, they get a four-letter code and a list of career suggestions. I wanted to do better than that. I wanted to treat the research with respect and still make something fun.

Building Arcus taught me things. Forcing every construct into the same shape destroys what makes it meaningful. The hardest part of a personality tool is deciding what to show and what to leave out. Research is a conversation.

I do not think Arcus will replace anyone's favorite personality test. That was never the plan. The plan was to build something that helps people see themselves a little more clearly, grounded in research I respect, and wrapped in an experience that is fun to use.

If you take a test and it tells you something you recognize, or makes you think about yourself in a way you had not before, then it worked.

Thanks for reading. Go take a test.