The six axes
Each axis runs from one valid preference to another. Neither end is treated as better.
How the 12 questions are balanced
Every axis is measured by exactly two questions. Each answer affects one axis only, and every axis has the same possible range. This avoids giving extra influence to an archetype simply because it appears more often in the quiz.
How city ranking works
Your answers are converted into six values between 0 and 1. Each city uses the same six-value structure. The quiz calculates the average absolute distance between your profile and every city profile, then ranks the smallest distances first.
fit = 1 - average(|your axis - city axis|)An internal numeric value is used for ordering, but the public result uses plain-language bands: Strong resonance, Good resonance, or Interesting contrast. If the top two cities are very close, the result says so instead of forcing certainty.
Archetypes are labels, not the ranking engine
Your archetype is a compact way to describe the overall pattern in your answers. It does not choose or reorder your city results.
How explanations are selected
The three closest axes become the main reasons for the match. The largest difference becomes the tension. This is why the result includes both resonance and a tradeoff.
How city profiles are reviewed
City profiles combine an editorial rubric with documented sources where practical data is used. Each profile must be reviewed, versioned, and tested as part of the full city set. A city is not a single experience, and neighborhood-level differences are outside this validation version.
The current 16-city vectors are editorial preview data. Practical fields display as unavailable until their source and review requirements are complete.
Soul Match and Reality Check are separate
Soul Match describes lifestyle resonance. The optional Reality Check looks only at the limited practical fields currently supported: budget band, climate preference, and connectivity. It never changes the city ranking.
Visa, tax, personal safety, healthcare, schools, housing availability, and employment are not evaluated. Missing or stale practical data is shown as unavailable rather than estimated.
Privacy by design
Your raw quiz answers and preference values remain in your browser in Validation V0. If anonymous product analytics is enabled, events may include the quiz step, completion, result city identifier, archetype identifier, match band, and share or premium-interest action. Raw answers and practical preferences are excluded.
Known limits
- The quiz measures stated preferences, not future happiness.
- City profiles simplify places that contain many neighborhoods and communities.
- Editorial judgments can change as the rubric and evidence improve.
- A match is an invitation to explore, not a promise or recommendation.
Versions and updates
Each result carries an algorithm version, question-set version, city-data version, and content version. Material scoring changes require a new version and regression tests.