How Wabiro tests are built and scored

Nothing in a Wabiro result is hidden or random. Each test has its own method page; this page describes what they all have in common.

Each test, its own method

• Purity Test — 50 yes/no questions, six weighted categories, a score that starts at 100 and drops with every yes.
• Am I a Red Flag? — 24 questions, eight weighted dimensions, a 0–100 red flag index and a main pattern.
• Dating Archetype — 16 statements, eight archetypes, the highest percentage wins.
• Are We Compatible? — two separate self-descriptions compared dimension by dimension; an alignment, not a compatibility percentage.
• Flirting Style — 14 statements, seven styles, one percentage each.
• Attraction Vibe — 15 statements, six behavioural signals, a vibe score and a signature signal. No photos.
• Social Battery — 12 statements, three dimensions, reversed recovery items, a 0–100 level.
• How Delusional Are You? — 14 statements, four areas, an entertainment level; never a clinical term.
• What Are My Beige Flags? — twelve harmless habits matched to one of six playful patterns.
• What’s My Yellow Flag? — sixteen statements, eight equally weighted relationship habits, a 0–100 index and a deterministic archetype.
• Is My Partner a Red Flag? — the same eight dimensions read from the other side, six high-severity behaviours weighted 1.5, plus five unscored safety questions.
• Why Am I Still Single? — 24 frequency statements, eight patterns including single by choice.

Shared principles

Stable questions. Every question has a permanent identifier, a dimension, a weight, a direction (risk or protective) and translations. Nothing is identified by its wording alone, so a translation change can never alter a score.

One answer scale. Answers are worth 0 to 3 points — yes/no items count 0 or 3. Protective items are reversed so that a higher number always means "more of the pattern" the test measures — a red flag, a yes on the purity test, a full battery.

Deterministic scoring. The engine is a pure function, covered by regression fixtures for extremes, thresholds, reversed items, ties, incomplete answers, each main pattern and safety flags. The same answers give the same result on any device, in any language.

Versioning. Test content (questions, weights, levels, profile texts) is versioned. A result records the test version, the scoring version and the copy version it was computed with. Changing a rule creates a new version; results and purchased reports keep theirs and are never silently reinterpreted.

Local first. Answers stay in your browser. When you share or buy, only an aggregated result (score, level, pattern) leaves your device — never your answers. The compatibility test sends one mean per dimension, never an answer.

Safety screening never changes a score. Some questions only decide whether a calm notice with verified support resources is shown.

Percentiles only when real. A result compares you with other Wabiro users only once at least 200 people have completed that exact version, from an anonymous distribution of scores. Below that, you see your band and nothing invented.

No diagnosis. We never use the words "toxic", "narcissistic" or "abusive", never label a person, and never tell anyone to stay or to leave. Wabiro tests are entertainment and a structured mirror for reflection, not clinical instruments.