UPDATED SEPTEMBER 11, 2026

The map, the math,
the receipts.

The map brings three things together: women's reported model choices, your length and girth, and a simulated past-partner lineup. Each has a different source. This page explains how they are drawn and what the numbers mean.

The women’s preference data

Prause, Park, Leung & Miller (2015) recruited 75 adult women in California, screened for attraction to men. Participants chose from 33 physical models in different relationship scenarios. This small convenience sample does not represent all women, and simplified model selections do not measure satisfaction or compatibility.

We downloaded the authors’ original CSV and questionnaire (Figshare v4, CC BY 4.0). Question 22 is the one-night scenario; question 24 concerns a long-term partner. There are 60 complete one-night responses and 63 long-term responses. Both views use the same 60 participants who answered both questions, matching the article’s paired comparison. The other three long-term answers are excluded to keep the comparison consistent.

These means are recomputed from the raw data, not from rounded centimeter values in the abstract. Extra digits describe the arithmetic, not the precision of people’s preferences. The graph plots aggregate counts at the original paired coordinates; the tables show every selected model.

How to read the heat

In the smoothed view, warmer colors mean more selections nearby. Each response contributes a Gaussian kernel with a 0.4-inch bandwidth in both dimensions. That bandwidth is a display choice. Colors are scaled relative to each view’s peak and clipped to the original choice ranges (length 4–8.5 inches, girth 2.5–7 inches). Equal colors across different contexts do not necessarily mean equal absolute density.

Switch to Exact selections to see coordinates and counts without smoothing. Circle size increases with the count. The study offered only 33 of a possible 100 grid positions. Blank or unlisted areas can include models never offered: they do not mean “disliked.” Neither view is a probability of satisfaction, a universal ideal, or a confidence region.

The numbered color key shows relative concentration from 0 to 100. We divide the smoothed density by the peak on the display grid and multiply by 100, capped at 100. The chart and key use the same color function. A score of 70 does not mean 70% of women prefer that size. Hatched areas lie outside the study’s offered size range and are unscored.

Your A–F map band

The letter summary groups the displayed one-decimal concentration score into our own bands: A = 80–100, B = 60–79.9, C = 40–59.9, D = 20–39.9, E = 10–19.9, and F = 0–9.9. These thresholds are a presentation choice. The study did not give people or models letter grades. A letter is not a satisfaction score or an assessment of sexual performance.

The result shows your letter and the number of simulated partners who scored each letter. Score ranges and the study explanation sit under “What do the letters mean?” Outside-range positions have no letter and are counted separately. The six band counts add up to the number of in-range positions, including when your own position is unscored. Context changes can change the bands; changing units, pronouns, or orientation does not.

How many land in a more popular spot?

We count simulated partners whose positions have a strictly higher smoothed choice density than yours in the selected context. This describes concentration of the study’s selections, not sexual performance or an individual partner’s taste. We compare unrounded densities, so small differences can count even when displayed scores round to the same number.

Simulated partners outside the offered length/circumference range are excluded from this comparison and its denominator. Their excluded count is shown. If your own measurements fall outside that range, this comparison is unavailable. Switching long-term/one-night changes the preference comparison while retaining the simulated measurements.

How we estimate past partners from a count

A count alone cannot reveal anyone’s measurements. We draw that many hypothetical people from a two-dimensional normal model. The eggplant markers are one possible lineup, not reconstructed history. Redraw creates a fresh sample without changing the analytical odds.

Eggplants have fictional names set by your partner and orientation selections. She/her + straight uses male names; she/her + gay/lesbian uses female names. He/him + straight uses female names; he/him + gay uses male names. Bisexual and pansexual choices use a mixed pool. For they/them, ze/zir, or a custom name with straight/gay orientation, or for another/undisclosed orientation, choose the fictional name mix directly. Names remain stable across preference views and unit changes. They do not establish anyone’s identity, anatomy, or actual history. Crowded plots show names where space allows; every name remains in the table and point inspector.

The marginal inputs come from Veale et al. (2015):

The review’s maximum overall sample of 15,521 includes other measurement types; it is not the sample size for erect dimensions. These marginal estimates come from different pooled samples, not a single jointly measured cohort.

The length–girth correlation is an assumption

The reference does not provide a compatible covariance estimate for the joint model. We therefore default to independent length and girth (ρ = 0). You can explore assumed correlations of −0.3, 0.3, and 0.6. These are scenarios, not measured population correlations. The displayed 0-to-0.6 comparison is a sensitivity illustration, not a confidence interval.

If U and V are independent standard normal draws, a paired draw is:

L = 13.12 + 1.66 U
G = 11.66 + 1.10 (ρ U + √(1 − ρ²) V)

Each hypothetical person is an independent draw. Actual dating pools may be selected and correlated in ways this model cannot describe. Gender, pronouns, orientation, and relationship context do not silently change the population parameters. Extremely remote normal tails can be physically implausible; the normal model is an approximation.

What “bigger” means here

The default asks about a partner who is both longer and thicker. Let q be the joint probability that one draw exceeds your length and your circumference. For n independent hypothetical partners:

P(at least one longer AND thicker) = 1 − (1 − q)ⁿ

At ρ = 0, q is the product of the two marginal upper-tail probabilities. At other correlations, we integrate the conditional normal distribution. This is a joint exceedance probability, not an overall size percentile. Length-only and girth-only modes use their respective marginal distributions. Those probabilities do not depend on the assumed correlation.

With 6 inches of length, 5 inches of girth, and 10 hypothetical partners, the marginal length percentile is approximately 89.92% and the girth percentile is 82.78%. With independent dimensions, the chance of at least one partner who is longer and thicker is about 16.06%. The length-only chance is 65.43%; the girth-only chance is 84.89%. These answer different questions.

Zero partners means zero chance of a bigger partner in the scenario. Numerical calculations use stable logarithmic formulas and conditional-normal quadrature, checked independently against a higher-resolution implementation. Accurate arithmetic does not remove uncertainty about the reference population or your measurements.

The result section also counts simulated partners who are longer and thicker, longer, and thicker. These counts overlap: one person may appear in all three. Each model average is n multiplied by the corresponding single-partner exceedance probability. A draw’s integer count can differ from that expected average.

Where newer research fits

A 2023 meta-analysis and a 2025 review report different averages. Their pooled-mean confidence intervals and standard errors cannot substitute for the spread of individual measurements. We retain the explicitly dated 2015 mean/SD pairs. This is a comparison to a particular reference, not a current worldwide ranking.

Inclusive language, accurate labels

Choose she/her, he/him, they/them, ze/zir, or a name. The wording changes throughout the experience. The preference layer continues to show women’s study responses. We have not created a male or nonbinary preference surface from that data. You can hide the preference layer and explore only your position and a simulated lineup.

Pronouns do not establish someone’s anatomy, tastes, or history. Set the count for the prior partners you want to include in this anatomical comparison. No graph can determine an individual partner’s preferences.

Measurement and privacy

Both inputs are erect measurements; girth means circumference, not diameter. A top-side bone-pressed length and gentle mid-shaft circumference provide a consistent self-measurement approach. Published measurement techniques vary, and self-measurements can differ from clinician measurements. Input bounds of 2–35 cm are interface limits, not clinical limits. One inch is exactly 2.54 cm.

Calculations, names, orientation, and simulated dots stay in your browser. We do not upload or save these inputs. Default sharing contains only the study map and a generic headline. If you opt into your map position, the image axes reveal approximate length and girth. Shared links never include measurements, partner names, orientation, counts, or results. See privacy details.