Programme 01 · Research

How often did the literature report on skin like yours?

One question, one answer, and no account. This reports on a published literature, not on you. It asks which band you are closest to only so it can tell you what was counted for that band.

Which of these is closest to your skin?

I Always burns, never tans II Usually burns, tans minimally III Sometimes burns, tans evenly IV Rarely burns, tans easily V Very rarely burns, tans darkly VI Never burns, deeply pigmented Not sure Show the figure for all six bands

We ask for nothing else. No email, no name, no photograph, no account, and nothing at all before the result.

Representation Check · Your band: III

01 · Your plate

At your skin type, 7 of every 12 photons make the round trip.

This is why a camera struggles with skin like yours. It is physics, not a flaw in you. The plate below is yours to keep.

Fitzpatrick is a phototype scale describing how skin responds to ultraviolet light. It is not an ethnic or racial classification, and nothing here describes anything about you beyond the band you picked.

Schematic A teaching schematic, not a measurement of you. The photon counts illustrate a double pass absorption model with illustrative melanin volume fractions. The anatomy and the mechanism are real and well established; the numbers are there to make it legible.

Put your name on it

Your name stays in your browser. We never receive it.

02 · What that costs

A model can lose information at your band and still report a good average.

Illustrative The shape of this problem is well documented. These specific numbers are placeholders that describe that shape, not a measurement of any model, ours or anyone else's.

Conventional model Skin tone stratified method

See the same data as a table

Balanced accuracy by Fitzpatrick skin type, illustrative
Fitzpatrick type	Conventional model	Skin tone stratified	Gap	n
I	94%	93%	-1	not stated
II	94%	93%	-1	not stated
III	93%	92%	-1	not stated
IV	88%	92%	+4	not stated
V	76%	91%	+15	not stated
VI	71%	90%	+19	not stated

03 · Who this literature serves

04 · What we are counting

The protocol is fixed before the counting starts.

We are running a scoping review, reported against PRISMA-ScR, of how often studies of image based skin diagnosis report their performance by skin type at all. The method goes up in public, dated, before a single paper is screened. A method fixed in advance cannot be adjusted to fit the answer, which is the whole reason to do it in that order.

What proportion of studies report performance broken down by skin tone, rather than pooled.
Which scale they use: Fitzpatrick, Monk, ITA, or none at all.
What proportion report any result for the darkest bands.
The median number of people behind each band, because reported but tiny is its own finding.

05 · If you want the rest

There are six plates. You have one.

The full series is the six bands drawn as one sheet, laid out to print as a single poster: the same cross section at every phototype, side by side, so the change in what light does is visible in one look rather than one band at a time.

We are writing the next one. Read it as it is written.

The set comes with the Reading Room, a finite sequence of seven pieces on how these models are built, who they are tested on, and what a sensor measures that a camera cannot. One every ten to fourteen days, for about three months, and then it ends.

Take your plate first. The set is offered once you have it.

See what is in the sequence ›

This is not a medical device and not a screening tool. It makes no statement about your skin, your health or your risk. The drawing is a schematic of how light behaves in skin generally, not a picture of you. Nothing you selected was stored or sent anywhere.