RateMyDick.ai
HomeThe Studio NEWArcadeSamplesFAQs
Analyze →
RMD Research

Seven proprietary measurements.

Every scan is reduced to a traced outline — hundreds of boundary points, a reconstructed centerline, and a width profile sampled along it. These seven numbers are computed from that trace, which is why a human rater, a chatbot, or a tape measure cannot produce them. Each ships in every report, ranked against 7414 confidence-gated scans. Below is what each one means, in brief.

01 / 07PROPRIETARY
Taper index
How evenly width falls from base to tip — the single number behind a "clean taper".
How it's measured
The ratio of fitted widths at the 25% and 75% stations of the traced width profile, normalised by mean width.
Why a tape can't
A tape gives you one girth at one spot. This needs the whole width envelope.
02 / 07PROPRIETARY
Flare ratio
How much the head widens against the shaft directly below it — the number behind what people call the classic look.
How it's measured
Maximum head width divided by shaft width just below the corona, both read off the traced profile.
Why a tape can't
It's a ratio of two widths at anatomically detected positions — you'd need to find the corona first.
03 / 07PROPRIETARY
Peak girth locus
Where you are widest, as a share of length from the base. Below the midpoint reads as a taper; above it as top-heavy.
How it's measured
The argmax of the fitted width envelope along the centerline.
Why a tape can't
Knowing your widest girth doesn't tell you where it lives. Position is the whole point.
04 / 07PROPRIETARY
Midline wobble
Fine left-right wiggle along the centerline — smaller than a curve, invisible to the eye, and completely invisible to a tape measure.
How it's measured
The RMS lateral deviation of the reconstructed centerline from its own smooth fit.
Why a tape can't
There is no tape-measure equivalent. This only exists once you have a traced centerline.
05 / 07PROPRIETARY
Shape rarity
How far your whole silhouette sits from the corpus average — distinctive, not better or worse.
How it's measured
Distance from the corpus centroid across eleven geometric axes at once, mapped to a percentile. Vision axes are excluded so photo quality can't inflate it.
Why a tape can't
Rarity is a property of the whole shape at once, not of any single measurement.
06 / 07PROPRIETARY
Range position (PC1)
Where you sit on the one continuous shape axis that replaced the old "type" buckets — clustering the corpus produced no stable types.
How it's measured
The projection of your normalised outline onto the corpus's first principal component of shape variation.
Why a tape can't
It's a coordinate in shape space. You need every outline in the corpus to even define the axis.
07 / 07PROPRIETARY
Outline smoothness
How clean your traced boundary is once capture noise is separated out — a measure of the shape, not the photo.
How it's measured
One minus the normalised high-frequency energy of the outline contour.
Why a tape can't
Smoothness lives in the fine structure of the boundary — hundreds of points, not two.
Next from the lab
Published findings — coming soon.
Full write-ups for each measurement: corpus distributions, what covaries with what, the null results we're publishing alongside the positive ones, and replication on a held-out half.

See all seven measured on you.

Every report computes the full set from your traced outline and ranks each against the corpus — alongside the 22 standard metrics.

ANALYZE MY DICK →VIEW SAMPLE REPORT