Norwood–Hamilton classification

The Norwood scale, I to VII.

Seven stages of patterned hairline recession, plus the vertex and anterior variants. This is the map norwood.fyi grades against, as a set of measurements with published anchors.

See a sample report

No account needed, and nothing is stored.

The seven stages
III Vertex and the A-types
Mature hairline vs II
How it is measured
Women and diffuse patterns
FAQ

The seven stages

Hamilton described types of patterned loss in 1951. Norwood, in 1975, redrew them into the I–VII scale clinicians still use, added III Vertex, and later the anterior (“a”) variants. Incidence rises with age (Rhodes 1998). The drawings below are schematics. A real grade is a measurement.

Norwood 1 (I). No patterned recession. The hairline still forms a dome; temples are not notched.

Norwood 1 →II

Norwood 2 (II). Triangular recession at the temples. The midline is still relatively preserved; the crown is usually intact.

Norwood 2 →III

Norwood 3 (III). Deep temporal recession: the classic M. Norwood’s original threshold of clinically significant patterned loss. Crown may still be full, or thinning (III Vertex).

Norwood 3 →IV

Norwood 4 (IV). Further frontal recession plus a larger vertex. A band of hair still separates the two zones.

Norwood 4 →V

Norwood 5 (V). The band between front and crown is still there, but sparse. The horseshoe is starting to read as one shape.

Norwood 5 →VI

Norwood 6 (VI). Front and vertex have joined. The band is gone. Remaining hair is the horseshoe of sides and back.

Norwood 6 →VII

Norwood 7 (VII). The most extensive stage. A narrow, sparse horseshoe. Only the wreath remains.

Norwood 7 →

III Vertex and the A-types

Two patterns do not fit a temples-first story, and Norwood named both.

III Vertex is a vertex patch appearing while the front is still a II or a III. The band of hair between hairline and crown is still there. People with a family history of crown-first loss often live here for years, looking “fine from the front” and not fine in a mirror above the sink. See Norwood 3.

The anterior variants (IIa, IIIa, IVa, Va) recede as a front, not as an M. The temples are not the headline; the whole hairline marches back. norwood.fyi will label an anterior IV when the front is already deep and the crown is not a separate island. Forcing every head onto a temples-first cartoon is how you mis-stage these.

Mature hairline vs Norwood II

Almost every argument about “am I receding” is this boundary. A mature hairline has moved up on the order of a centimetre from the juvenile line and kept its dome: midline still the high point, temples not cut out. Norwood II is triangular recession: a share of the temple zone with no hair at all.

A common mistake, including in an earlier version of this engine, is to subtract midline height from temple height and call that recession. On a healthy head the hairline is a dome, so the temples sit lower than the midline and that difference is negative. Recession is the bald notches of the M, measured as bald columns. Full notes on that cut: Norwood 2.

How norwood.fyi measures it

Five photos. A dense facial landmark mesh builds a coordinate frame on the head, so distance, tilt and phone do not set the grade. A segmentation network traces the hairline and finds scalp showing through hair on the front, crown and back. Measurements are in H (glabella to chin) and as densities relative to the occipital reference, which is the androgen-insensitive hair at the back of the head.

Hairline height: Midline, in H. Full ≈ 0.50–0.64. Infinite if no midline hair remains.
Temporal recession: Share of temple columns with no hair. II ≈ 10–30%. III ≈ 30–60%.
Vertex density: Crown relative to the back of the head. Intact ≈ ≥ 0.85.
Loss-area ratio: Bare scalp ÷ scalp area. Anchors: 0, 0.14, 0.25, 0.34, 0.40, 0.66, 0.82.

Those loss-area anchors are per-grade values from Xi et al., Scientific Reports 2025, 761 dermatologist-graded images, Mask R-CNN segmentation. The continuous index (the “3.4” next to III) is the rule stage nudged by that area, kept inside the stage. Confidence drops when a measurement sits on a cut, when a photo is bad, or when the two temples disagree. The method page is the spec sheet: Method & sources.

The paid report also projects 5- and 10-year stage under published trial response rates (Kaufman et al. 1998, 2002 for finasteride 1 mg; Olsen et al. 2002 for minoxidil 5%), and splits genetic versus environmental drivers (Chumlea et al. 2004; Su et al. 2007). Spec sheet: Method & sources.

Women and diffuse patterns

Norwood–Hamilton is a male-pattern map. Women more often show midline widening and diffuse thinning (Ludwig / Sinclair), which this scale was not drawn to name. The measurements still run. When scalp shows through evenly without a frontal or vertex driver, the report flags diffuse rather than inventing an M that is not there. Diffuse thinning can be androgenetic, telogen effluvium (illness, stress, crash dieting, deficiency), or female-pattern loss. A roman numeral is the wrong summary of that, and the report says so.

FAQ

What is the Norwood scale?

The Norwood–Hamilton scale is the standard map of male-pattern hairline recession, published by O’Tar Norwood in 1975 on top of James Hamilton’s 1951 types. It runs from I (no patterned recession) to VII (a narrow horseshoe). Variants cover a vertex-first pattern (III Vertex) and a front-first pattern (the “a” types).

What Norwood am I?

You cannot reliably read it from a bathroom selfie. Angle, wet hair and a hatline all move you a stage. norwood.fyi takes five photos and returns a stage, a genetic vs. environmental split, ranked levers, 5- and 10-year paths, and a signed link. $7, no account, nothing stored, nothing sold after you pay. Open the sample report first if you want to see the layout.

What is Norwood 3 vertex?

III Vertex is temporal recession around stage III plus a separate thinning or bald patch at the crown, with a band of hair still between them. Classic III has the deep M and a spared crown. Once the band thins you are in V; once it is gone you are in VI. Dedicated notes: the Norwood 3 page.

Is a mature hairline Norwood 2?

Not by itself. A mature hairline has moved up about a centimetre and kept its dome. Norwood II is triangular recession at the temples, measured here as the share of temple columns with no hair at all. Full notes: the Norwood 2 page.

Does the scale work for women?

The Norwood–Hamilton types describe male-pattern recession. The measurements (hairline, density, scalp visibility) still run on anyone. Diffuse or midline-widening patterns that fit female-pattern or non-androgenetic loss are flagged as such rather than forced onto a roman numeral that does not fit.

How accurate is Norwood staging?

Adjacent grades agree about 65–78% of the time even among dermatologists. That is why the report shows the measurements behind the number. The report shows hairline height, temple bald-column share, vertex density, loss-area, a confidence score, and the photographs’ quality flags.

norwood.fyi is an informational tool, not a medical device, and not a substitute for a dermatologist. Staging between adjacent Norwood grades has about 65–78% agreement even among experts; the report shows its confidence and the measurements behind it.