Transparent-ish math
How the estimate works.
The result is an entertainment estimate. It starts with country population and progressively applies broad demographic ranges. The app deliberately dampens most filters because real traits overlap and are not statistically independent.
1. Starting pool
The app uses the latest population value available from the World Bank API, with bundled fallback values for major countries. Working-age share, female population share, urban share, smoking prevalence, and tertiary-education indicators are used when available.
2. Core filters
Age, partner preference, search radius, height, education, lifestyle, children, faith, and relationship goal use public indicators where possible and broad regional ranges otherwise.
3. Optional preference model
Appearance, hair, eye color, personality, career, and relative income use broad regional priors and conservative dampening. These categories are not official country census counts and should not be interpreted as precise or as judgments about identity. “Income” means a relative level inside the selected market, not one global currency amount.
4. Mutual interest
The “make them like me back” button adds a deliberately conservative mutual-interest factor. Attraction and compatibility cannot be measured accurately from demographic data, so this is explicitly playful.
5. Humor
Jokes are chosen from localized language, country, preference, result-band, and trend fragments. Saudi visitors can optionally select Riyadh, Jeddah, the Eastern Province, or the south/west for extra local flavor. The humor layer never changes the math.
Important limitations
- Country indicators can be incomplete or from different reporting years.
- Traits overlap, definitions vary, and real populations change.
- Sex/gender statistics are often published in binary form; the “anyone” option intentionally stays inclusive but cannot repair gaps in source data.
- No result should be used to judge a person, identity group, or country.