How we build what we publish.

Calculator and paid reports

The calculator uses method pythagorean-components-v2. Its score and four labels are editorial rules for symbolic reflection. They are not a validated measure of compatibility, spiritual connection, relationship safety, or future outcomes.

For names, accented Latin letters are normalised to their base letters and punctuation and spaces are ignored. A–I map to 1–9, J–R repeat 1–9, and S–Z repeat 1–8. Sum the supplied letters and reduce by digit sums, retaining 11, 22, or 33. Unsupported scripts require a Latin spelling. The result describes the name supplied; the usual full Expression convention uses the complete birth name.

For life paths, reduce the birth month, day, and year separately, retaining 11, 22, and 33. Add those results and reduce again. This follows a published component-reduction convention; other schools use a flat digit sum and can disagree on master numbers. Example: November 29, 1984 gives 11 + 11 + 22 = 44 → 8. We use roots 2, 4, and 6 to compare master numbers 11, 22, and 33.

The following house table supplies both name and life-path pattern points. Rows and columns are the two single-digit roots. A matching root receives 100 points by our rule; it does not imply perfect compatibility. These point values are not prescribed by the practitioner sources.

Astrology uses a fixed tropical Sun-sign date table, without birth time, time zone, location, or an ephemeris. Signs are approximate, especially near a sign change. Element points are: same sign 90; different signs with the same element 80; Fire/Air or Earth/Water 70; Fire/Earth or Air/Water 50; Fire/Water or Earth/Air 45. This is a house point table, not a synastry chart. Whole-sign geometry in the PDF is explicitly distinguished from an aspect between actual planetary positions.

Score = round(name points × 0.4 + life-path points × 0.3 + element points × 0.3). Rounding happens once, at the end. The current tables give a range of 35–97 points. The labels, applied in order, are: Twin Flame at 85 or above plus an exact name-number or life-path match; otherwise Soulmate at 75 or above; otherwise Karmic Connection at 45 or above; otherwise Soul Companion. A score above 85 without the exact match can therefore be Soulmate. Labels have no established population prevalence.

Pattern flags record reversed two-digit days (01/10 or 12/21, excluding identical days), exact life-path matches, life paths of 11/22/33, or traditionally complementary elements. The flags do not add bonus points. They cannot demonstrate destiny or explain a person’s behaviour.

Free calculations run in your browser. Email requests and purchases send and store the supplied inputs and result. New records include the method and calculation working. Existing purchases retain their original snapshots; they are not silently recalculated. A legacy email reading without a complete snapshot is identified as recalculated if the current method is used.

Paid PDFs contain computed factual sections and optional AI-assisted reflection. Generated reflections receive limited symbolic themes rather than names or birth dates; reviewed copy replaces prose that fails checks. Automated checks reduce errors but cannot establish that all natural-language interpretation is accurate. Page count varies. A report cannot recover relationship history from names and dates.

Practitioner references: Numerology.com on life paths, Hans Decoz on component reduction, master-number comparison roots, and Expression names. These establish the conventions we chose, not scientific validity.

Quiz scoring

All three quizzes use our own questions and cutoffs. None has been clinically validated. Background reading on attachment, limerence, and abusive relationships does not validate our weights or spiritual categories. Results cannot establish or exclude a diagnosis, abuse, or a metaphysical bond.

The long quiz gives weighted points to five themes. Each displayed percentage is that theme’s points divided by all primary points awarded. It is not a probability; rounding can make displayed shares differ from 100%. Exact ties use this order: True Twin Flame, False Twin Flame, Karmic Connection, Companion Soul, Soulmate. The fixed order is a software rule, not additional evidence.

The long quiz ends after 20 answers if a category has at least 50% of primary points, otherwise after all 33. This measures concentration of points, not statistical confidence. Each reflection flag appears at 60% of its maximum possible points on the questions actually answered. A short run uses a smaller denominator; an absent flag cannot rule out a concern.

The ten-question short quiz counts every category assigned to a selected answer. The highest count supplies the label; ties use False Twin Flame, Karmic Connection, Limerence, then True Twin Flame.

The limerence reflection tool adds twelve answer values from 0 to 3, for a total from 0 to 36. Editorial bands are 0–9, 10–18, 19–27, and 28–36. These are not Tennov’s diagnostic cutoffs or validated severity categories. Review individual answers and their impact on daily life regardless of the total.

Historical partner dataset

The aggregate supplied by tarostarot.com contains 7,533 calculator submissions from November 2025 through April 2026: 4,998 Karmic Connection, 1,468 Soulmate, 547 Companion Soul, 408 Twin Flame, and 112 Soul Companion. The share 408 ÷ 7,533 × 100 is approximately 5.4%.

Those are submissions, not confirmed distinct people or couples. We have aggregate counts rather than independently verified relationship outcomes. The historical calculator used five categories and must not be treated as the current four-category model. Its output distribution does not establish how selective or accurate this site’s current calculator is, or the prevalence of any relationship type.

Journal summaries

Counts describe logged entries with completed interpretations, not every event in a person’s life. Time and weekday summaries use UTC logging timestamps, not the time an event occurred. Weekly summaries use the completed week’s cutoff; their 30- and 90-day context remains a longer window. A newly recurring value has at least two logs in the recent seven days and none in the ninety days before them.

Display thresholds are descriptive house rules: at least five entries and a 30% share for a four-hour time bucket; at least seven entries and 1.5 times the mean weekday count for a weekday flag; a 20 percentage-point type-share change with at least three entries in each comparison window; and mood lift of at least two with three co-occurrences among at least ten mood-tagged entries. Mood lift compares a value’s share within that mood with its share across mood-tagged entries. Missing moods are excluded from both denominators. These are not significance tests, causal findings, or evidence of unusual coincidence rates.

AI reflections can make mistakes. They cannot infer another person’s feelings or predict reunion from logged patterns. User-selected stages are context, not verified assessments.