Method, sources and limits

Every axis is built from a named source or clearly marked original items. Here is what each one is, how it is scored and what it cannot tell you.

Sources by test

Horoscope: The Astrology of Lasting Unions, Jan Jilecek, 2025 (blog.jilecek.cz). Public-figure marriages from Wikidata, born after 1940. The rising sign is computed from sidereal time (IAU 1982), the obliquity of the ecliptic and your latitude, with the time-zone database of your browser; the aspects between rising signs, Suns and descendants are the traditional whole-sign aspects.
For fun and curiosity. Sun signs are not scientifically established predictors of relationships.
Numerology: Pythagorean number reduction with the master numbers 11, 22 and 33; pair verdicts from the friend, neutral and enemy relations of the ruling planets in Vedic numerology (Parashari friendships).
For fun and curiosity. Sun signs are not scientifically established predictors of relationships.
MBTI-style type: Statements from mbti.jilecek.cz, scored as continuous scales. MBTI is a registered trademark of The Myers-Briggs Company; this is an independent, unofficial indicator.
For self-reflection only. This is not a clinical or diagnostic instrument.
Big Five: IPIP Big-Five Factor Markers, 50 items (Goldberg, 1992), International Personality Item Pool, public domain.
For self-reflection only. This is not a clinical or diagnostic instrument.
Archetypes: Hero Myth Index (Pearson and Marr), 72 items, via archetypes.jilecek.cz.
For self-reflection only. This is not a clinical or diagnostic instrument.
Character strengths: Original items on the six virtue groups of the VIA classification of character strengths (Peterson and Seligman, 2004), three per virtue.
For self-reflection only. This is not a clinical or diagnostic instrument.
Political outlook: Original, neutral items on four separate leanings (equality, liberty, tradition, order), after the two-dimensional model of political attitudes and moral-foundations research. No party is named.
Private and optional. Your answers are never shown in an invitation. Like every test, only your scores (never your answers) go to the AI writer for the written commentary. For reflection only, not advice.
Attachment style: attachement.jilecek.cz, adapted from measures related to Fraley, Waller and Brennan (2000).
For self-reflection only. This is not a clinical or diagnostic instrument.
Love styles: Original items on the six love styles of J. A. Lee (1973), in the tradition of the Love Attitudes Scale (Hendrick and Hendrick), three per style.
For self-reflection only. This is not a clinical or diagnostic instrument.
Love beliefs: Original items on four beliefs about love, after research on romantic beliefs (Sprecher and Metts, 1989), three per belief.
For self-reflection only. This is not a clinical or diagnostic instrument.
Love languages: Original items on the five love languages of Gary Chapman (1992), three per language.
For self-reflection only. This is not a clinical or diagnostic instrument.
Conflict style: Original items on the five conflict-handling modes (Thomas and Kilmann, 1974; the dual-concern model of Rahim), three per mode.
For self-reflection only. This is not a clinical or diagnostic instrument.
Everyday life: Original items on daily habits in couples, after the Big Five facets of orderliness, openness to new experiences and sociability, three per area.
For self-reflection only. This is not a clinical or diagnostic instrument.
Household chores: Original items on how couples divide household labour, three per area.
For self-reflection only. This is not a clinical or diagnostic instrument.
Intimacy attitudes: Original, non-explicit items on four attitudes to sex and intimacy, after research on sexual attitudes (Hendrick and Hendrick) and sociosexual orientation (Simpson and Gangestad), three per attitude.
Private and optional. Your answers are never shown in an invitation. Like every test, only your scores (never your answers) go to the AI writer for the written commentary. For reflection only, not advice.
Dark Triad: Short Dark Triad, SD3 (Jones and Paulhus, 2014), via tests.jilecek.cz.
Self-reported personality traits, not a diagnosis. Scores describe tendencies, not a person.
ADHD screener: WHO Adult ADHD Self-Report Scale, ASRS-v1.1 (Kessler et al.), via tests.jilecek.cz.
A screening aid, not a diagnosis. Only a qualified professional can make one.
Autism spectrum screener (AQ-10): Autism Spectrum Quotient, AQ-10 (Allison, Auyeung and Baron-Cohen, 2012), via tests.jilecek.cz.
A screening aid, not a diagnosis. Only a qualified professional can make one.
Empathy: Original items covering cognitive, affective and empathic-action facets.
For self-reflection only. This is not a clinical or diagnostic instrument.
Reasoning: Original reasoning items.
A short reasoning exercise, not an IQ test. Scores are a rough indication only.
General knowledge: Original questions.
For self-reflection only. This is not a clinical or diagnostic instrument.
Work motivation: Nine-motivator work motivation test, extended with Schein’s career anchors and the Seeker archetype career analysis (blog.jilecek.cz, 2020).
For self-reflection only. This is not a clinical or diagnostic instrument.
Interest types: Original items on John Holland's six interest types (RIASEC), after the public-domain IPIP interest markers, three per type.
For self-reflection only. This is not a clinical or diagnostic instrument.

How scores are calculated

Answers are scored on your device. Reverse-keyed items are flipped, then each dimension is scaled to 0–100. Only these scores, never your answers, are used to build a report.

How compatibility is calculated

Each axis has a small explainable model: similarity for values and rhythm, stacking for risk traits, weakest-link for empathy, a style matrix for attachment and a lookup table for horoscope pairs. The overall score is a weighted average of the axes you both completed. Weights favour axes with stronger evidence: attachment, Big Five, dark traits and empathy weigh most, horoscope and knowledge least.

What the AI does and does not do

A language model writes the commentary from your scores. It does not compute scores and cannot see your answers. Treat the text as a thoughtful summary, not as advice.

Where the Big Five career matches come from

If you complete the Big Five, your five trait scores are compared with the typical trait profile of 3,513 occupations from the vocations dataset behind bigfive.jilecek.cz. The closest ten are listed by Euclidean distance. The original site sums signed differences, which lets opposite differences cancel out, so this site does not. The matching runs on the server and only the ten results are stored with your report.

A note on MBTI scoring

Our MBTI test uses the same 72 statements as mbti.jilecek.cz but scores each dichotomy as a continuous 0–100 scale toward the first pole, so all 16 types are possible.

Limits

Self-report tests measure how you see yourself today. Short tests are less reliable than long ones. Clinical-style scales here are not diagnostic. Relationship outcomes depend on far more than any test can measure.