Methodology & Data Sources

This page documents every dataset, distribution, and formula behind the dating standards calculator. Nothing here is guessed. Each number the tool produces traces back to a named public dataset from the US Census Bureau, the CDC, or the Bureau of Labor Statistics, and the exact way the filters combine is written out below so the result is auditable and reproducible.

How the calculator works

Your filters are applied to the adult population (age 18 and over) of the selected gender. Each filter keeps only the share of people who meet that criterion, and the shares combine to give the size of your matching pool. The key point is that standards multiply, they do not add. Ask for the top 15% by income and you keep 15 in every 100. Add a height cut that keeps 1 in 7, and the two combine to roughly 2%, not 15%. Stack a third filter and the pool shrinks again.

The correlation adjustment

Treating every filter as fully independent would be wrong for traits that move together in the real world. Income and education are the clearest case: higher earners are far more likely to hold a degree, so multiplying their raw probabilities would double-count the overlap and understate your true pool. The engine treats income and education as one correlated group and anchors their joint probability to the more selective of the two, using a documented interpolation between full independence and full dependence:

jointGroup = p_min × product over the others of ( p_k ^ (1 - rho) )

Here p_min is the most restrictive marginal probability in the group, and rho is a correlation strength between 0 and 1. At rho = 0 the formula returns the plain independent product. At rho = 1 it returns the single most restrictive probability. We use rho = 0.55 for the income and education pair. The result always sits between the independent product and the minimum marginal, so it widens the estimate back toward reality rather than shrinking it. All other filters are treated as independent, because no reliable joint distribution is available for pairs such as height and income.

The delusion score

The delusion score rates how statistically rare your combined standards are, on a scale from 1 to 10. It scales with the base-10 rarity of your final pool:

score = 1 + 1.5 × log10( 1 / poolFraction ), clamped to the range 1 to 10.

Benchmarks: a pool of 1 in 2 scores near the bottom, 1 in 20 lands around 3, 1 in 1,000 sits near the middle at about 5.5, 1 in 10,000 reaches about 7, and 1 in a million or rarer reaches the maximum of 10. The score measures rarity only. It is not a judgment of any person.

Biggest constraint and privacy

The tool also reports which single filter is shrinking your pool the most, so you can relax one slider and watch the pool grow. Every calculation runs in your browser with JavaScript. Your inputs and results are never sent to a server, stored, or tracked.

Our data sources (United States)

The calculator computes on United States distributions. The table below lists the source, measurement type, and reference year for every input. Self-reported figures come from large government surveys; height and body-mass figures are physically measured by trained professionals, which is more reliable than self-report.

Key figures the engine uses

For full transparency, these are the anchor statistics the US tables reproduce. Where a figure differs slightly between published sources, we note the value used and its basis.

Reference data for other countries

The calculator computes on United States data. The figures shown on the calculator pages for the United Kingdom, Canada, Australia, and Germany are cited reference context, drawn from each country's own national statistics office. They let you read how the same standards land in another market, but they are not run through the scoring engine.

Source classification

Each source falls into one of three roles:

• Core rarity. Distributions used directly in the pool and score calculation: age, height, income, education, marital status, body composition, and ethnicity for the United States.
• Measured. Physical measurements taken by trained professionals, which are more reliable than self-report. This covers height and BMI from CDC NHANES.
• Reference and context only. Figures shown for information but not fed into the score, including the international country data and the desired-distance setting, which provides geographic context and does not change your rarity.

Ethnicity methodology

Ethnicity uses OR logic. Each option you select adds that group's share of the population to your pool, so choosing several categories widens your options rather than narrowing them, and selecting none applies no ethnicity filter at all. Shares come from ACS Table B03002, which reports Hispanic or Latino origin by race, following the common survey practice of treating Hispanic as a distinct category alongside single-race groups. The shares sum to approximately 100% of the adult population.

Limitations and disclaimers

• Results are statistical estimates for entertainment and educational use, not guarantees, and they do not predict your dating outcomes.
• Figures describe the national population, not the real people in your city, so a local pool can look very different from the national estimate.
• The single or available filter uses marital status as a proxy, which counts unmarried people who are cohabiting as available.
• Where no joint distribution exists, filters are combined as independent probabilities. The only modeled correlation is between income and education.
• Self-reported height and weight tend to overstate height and understate obesity. This is why German figures read tall, and why we prefer measured surveys for the United States.
• Datasets are the most recent available and may be one to three years old. Each table records its source and version in the code, and we update the numbers when new releases land.

All of the data referenced here is publicly available and free to access. We do not modify or alter the source figures beyond fitting them into the distributions described above.