Facial thirds test — are your thirds balanced?
The Vitruvian ideal is 1:1:1 — upper, middle, and lower thirds equal. Measure all three free by marking four lines on your photo below. Or upload a photo to the scan: the free scan returns your Face Score and strongest metric, and whether your middle or lower third runs long, and by how much, is in the full report.
Start my free scanYour facial thirds are one of 17 metrics that determine perceived attractiveness.
Thirds ratio alone won't tell you why your face reads the way it does. The audit reads every photo you send on the same measurements (canthal tilt, FWHR, eye shape and more), then a personalized 5-page PDF tells you which to fix first.
Sample thirds result
Your middle : lower split
56 : 44 → long midface
Middle third 6pt over an even split · Upper third not scored (no hairline landmark)
Upper
Not scored
Middle
56%
Lower
44%
Sample numbers. Run the test below for your actual thirds distribution.
Facial thirds calculator: measure all three yourself
Free, and the photo never leaves your device: it is drawn in your browser and nothing is uploaded. Mark four lines (hairline, brow line, nose base, chin) and you get each third as a share of your face height against the 1:1:1 ideal.
Each arrow press moves a line 1% of the photo's height; hold Shift for a quarter of that.
Your results expire after this session — save or screenshot
Your Facial Thirds Score
Upload a front-facing photo to get your score
Upper third the one reading long? Take the big forehead test — it measures the forehead share specifically.
Measure your facial thirds in 60 seconds
This is how to measure facial thirds, and it doubles as a facial thirds test online: equal facial thirds are the commonly cited ideal, so an equal facial thirds test is really just a check of how close one photo sits to that range. People arriving from looksmax threads sometimes call a low reading a sub 3 face — that is their slang, not a verdict this page issues.
Every word of our 60-second facial thirds explainer, on the page:
What are facial thirds? And does balanced actually mean better?
Divide the face at the brows and the base of the nose. Forehead, midface, lower face. Three bands.
Roughly equal thirds is a commonly cited range. It's not a law, and balanced thirds don't make anyone a better person.
Tip your chin up and your lower third stretches. That's the camera, not your face.
Can you change them? Honestly, no. Thirds are set by bone. You didn't do anything wrong, and uneven thirds are not a flaw.
A fringe or a different camera height changes how the thirds read in a photo. Not your face.
See your own thirds measured. One of seventeen measurements, free, at realsmile.online.
The Vitruvian 1:1:1 ideal — and why it matters
Facial thirds is the oldest measurable rule in facial aesthetics. Leonardo da Vinci codified it from earlier Greek and Roman work in his late-15th century Vitruvian studies: the face divides into three vertical sections — hairline to brow line, brow line to nose base, nose base to chin — and the harmonious face has all three sections roughly equal.
Modern orthodontics, maxillofacial surgery, and computational rating systems all use this 1:1:1 baseline as the reference for "balanced."
It survives as a reference line rather than a description of real faces. When Farkas and colleagues tested the classical canons on 153 young adults, the vertical profile proportions were the ones faces matched worst, and they concluded the canons do not represent average facial proportions (Farkas et al., 1985).
The flip side: small deviations mostly go unnoticed; large ones do not. In a rating study that varied chin height on idealized faces, chin heights from about 30% up to about 40% (men) or 50% (women) of lower face height went largely unnoticed, and raters wanted surgery only well outside that range (Naini et al., 2012).
Knowing which of your thirds is off — and by how much — is the highest-leverage diagnostic in looksmaxxing because it tells you whether to fix the third (chin work, hairline cuts) or compensate around it (styling, beard, glasses).
The 3 thirds, region by region
Ideal is 33% ± 2% per third — the classical reference band, not a pass mark; most real faces sit outside it somewhere.

| Third | Reads long as | Reads short as | Compensatory levers |
|---|---|---|---|
Upper Thirdhairline → glabella (between brows)33% ± 2% of total face height | Reads long asLong upper third reads as receding hairline or "forehead heavy." Common after Norwood 2+ or naturally high hairline. | Reads short asShort upper third reads as low hairline or "no forehead." Common with low brow ridges or coarse-textured hair growing forward. | Compensatory leversCut to face shape (curtain bangs, side-fringe), brow positioning via shaping, hair density treatment if recession is the cause.Face shape test → |
Middle Thirdglabella → subnasale (base of nose)33% ± 2% of total face height | Reads long asLong midface. Reads as elongated, "pulled-down," or vertically asymmetric. | Reads short asShort midface reads as compact and youthful but can flatten the cheekbones. Less common than long midface but appears in some heart and round face shapes. | Compensatory leversLong midface is largely structural — bone-anchored. Compensatory plays: facial hair to break vertical line, glasses with strong horizontal frame, hair volume up top to balance, makeup contour on cheekbone (women).All face shapes guide → |
Lower Thirdsubnasale → menton (chin tip)33% ± 2% of total face height | Reads long asLong lower third reads as "horse-faced" or chin-heavy. Often paired with mandibular prognathism (jaw forward of upper jaw). High among long-face/dolichocephalic skull shapes. | Reads short asShort lower third reads as a recessed chin or weak jaw. Strongly correlated with mouth-breathing development and low attractiveness ratings, particularly in men. The single highest-leverage third for surgical or restorative work. | Compensatory leversLower third is the most fixable third. Chin filler, jaw surgery (mandibular advancement), orthodontics for occlusion, body-fat reduction to unmask gonial angle, mewing posture for long-term.Jawline angle test → |
8-metric suite
Your complete looksmax scorecard
One AI scan measures all 8 metrics from a single selfie. You've explored one — get the full picture.
6 thirds archetypes — find yours
The pattern your photo lands in tells you whether to fix, compensate, or leave alone.
| Pattern | What it reads as | The play |
|---|---|---|
1:1:1 (perfect thirds) | What it reads asVitruvian baseline. Uncommon in real faces: most measured faces deviate from equal thirds in at least one section. | The playDon't fix what isn't broken. Direct effort at single-feature halos (eyes, smile, skin) instead. |
Long midface (>36%) | What it reads asSeen in dolichocephalic / long-face skulls. Reads as elongated through the middle of the face. | The playCompensatory styling — break vertical with facial hair, horizontal-frame glasses, hair-volume top. Surgical fix is rare; LeFort I osteotomy is invasive and only indicated with malocclusion. |
Short lower third (<30%) | What it reads asRecessed chin pattern. Often paired with class II malocclusion and mouth-breathing development. High looksmax leverage — chin work has outsized return. | The playChin filler ($600–1500, instant), genioplasty surgery ($4–8k, permanent), or mandibular advancement orthodontics if young. See /procedure-roi for ranked options. |
Long lower third (>36%) | What it reads asChin-heavy / horse-face read. Often class III malocclusion (lower jaw forward of upper). | The playReduction genioplasty if severe. Beard trimmed shorter under chin (avoid the visual extension downward). Hair volume top to redistribute the read upward. |
Long upper third (>36%) | What it reads asHigh hairline or receding hairline. Reads as forehead-heavy. Can be natural (always had high hairline) or progressive (Norwood pattern). | The playCut frames forehead — curtain bangs, side fringe, textured top. If progressive: minoxidil + finasteride or transplant. Never over-comb back; that exposes more upper third. |
Short upper third (<30%) | What it reads asLow hairline or hair-grown-forward look. Less penalized than the long-third patterns but can read as "primitive" or unrefined frame. | The playPluck/wax forehead-side hair growth, work with a barber on hairline cleanup. Brow positioning matters — over-low brows compound the look. |
Why the mirror lies and the test doesn't
You can't see your own thirds in a mirror — the brain self-anchors on the eyes and back-fills the rest. Most people with long midface or short lower third have no idea until they see a measurement.
The test is brutally honest because the landmark detector has no ego: it places the glabella, subnasale, and menton from raw pixel data and computes the ratio (the upper third is not scored — see how the thirds are measured below).
The test compares your middle third with your lower third, with an even 50/50 split as the reference, and it takes no ethnicity or demographic input. Proportions do differ between groups — in Le and Farkas's 2002 measurements of Chinese, Vietnamese and Thai adults, the lower face ran shorter than the forehead — so read a small deviation as normal variation, not a verdict.
How the thirds are measured — including the honest part
The scan runs a 68-point facial-landmark detection (iBUG 300-W scheme) — in your browser on desktop, and on mobile the photo is processed in memory by our scan server and deleted immediately after the landmarks are measured — either way, nothing is kept unless you choose to save it.
Two of the three thirds are measured directly from those landmarks: the middle third from the brow-line midpoint down to the base of the nose, and the lower third from the nose base to the chin point.
The upper third is different, and we would rather say so than pretend: the 68-landmark scheme has no hairline point, and a hairline read from a front photo is unreliable (bangs, hats, recession), so the upper third is not scored at all. The score uses only the middle and lower thirds, which carry full landmark precision.
Your balance score compares those two: each is taken as a share of brow-to-chin height, an even 50/50 split is the reference, and the further the split moves from even, the lower the score. That reading is then placed against other RealSmile scans, so 50 is a typical result.
Lower-third dominance is also one of the gender-differentiated markers scored on the masculinity test, so if your lower third runs long, that page shows the other side of the same measurement.
If you want to check the math yourself — with a ruler in the mirror or on a printed photo — the classical method uses four landmarks in a vertical chain: trichion (the hairline midpoint), glabella (the smooth point between the brows), subnasale (the base of the nose), and menton (the bottom of the chin).
- Locate the trichion — the midpoint of your natural hairline. This is the landmark our scan cannot see (the 68-point scheme has no hairline point), which is exactly why a manual check is how you get your upper third.
- Measure trichion to glabella — hairline down to the point between your brows. That distance is your upper third.
- Measure glabella to subnasale — brows down to the base of the nose. That distance is your middle third.
- Measure subnasale to menton — nose base down to the bottom of the chin. That distance is your lower third. Divide each segment by the trichion-to-menton total to get the three percentages.
What the "ideal" ratio actually is
The classical canon — the one used in cephalometric and orthodontic literature — is three roughly equal thirds, about 33% each. That is the reference line, not a pass mark: as the data section below notes, real faces normally deviate from it, and a deviated third is the common case, not a defect.
Thirds balance is one proportion among many, so matching the equal-thirds canon does not make a face attractive on its own, and missing it does not make a face unattractive.
What our data shows
In that literature, an elongated lower third — above roughly 36% of face height — is the most common imbalance, and faces with all three sections within 2% of equal are uncommon (Farkas & Munro, 1987; Naini, 2011). In other words, a deviated third is the normal case, which is why the result tells you which third and by how much instead of a pass/fail.
The research
| Study | Source | Finding |
|---|---|---|
Vitruvian proportions in facial aesthetics | Farkas et al., 1985 | Nine neoclassical proportion canons were tested on 153 young adult North American Caucasians. The vertical profile proportions matched worst, with equality of no more than two sections, and the authors conclude the canons do not represent average facial proportions. |
Lower facial third in male and female attractiveness | Naini et al., 2012 | Chin height was varied on idealized male and female front-view faces and rated by orthognathic patients, clinicians and laypeople. Chin heights from about 30% up to about 40% (men) or 50% (women) of lower face height went largely unnoticed; raters wanted surgery only beyond about 50% (men) or 58% (women), or below about 20%. |
Ethnic variation in facial thirds baseline | Le, Farkas et al., 2002 | Chinese, Vietnamese, Thai and North American Caucasian young adults (60 per group) were measured against six neoclassical canons. In none of the groups were the three sections of the facial profile equal, and in the Asian groups the lower face was shorter than the forehead. |
Mouth-breathing development and lower third length | Zhao et al., 2021 | A meta-analysis of 10 studies in children: mouth breathers had steeper mandibular plane angles than nasal breathers (SN-GoGn about 4° higher), and the authors conclude that the mandible and maxilla rotate backward and downward. |
Facial thirds — FAQ
What are facial thirds?
The face divides vertically into three sections — upper (hairline to glabella between the brows), middle (glabella to base of nose), and lower (base of nose to chin tip). The Vitruvian / Renaissance ideal is 1:1:1 — all three roughly equal, each about 33% of total face height.
It is a reference line, not a description of a typical face: when Farkas and colleagues tested the classical canons on 153 young adults, the vertical profile proportions, equal thirds among them, were the ones real faces matched worst (Farkas et al., 1985).
How accurate is a photo-based facial thirds test?
The scan measures the middle and lower thirds from 68 detected landmarks on a front-facing photo (brow line, nose base and chin), and its Facial Thirds Balance reading moves from photo to photo: two photos of the same face on the same day differ by a median of 4 points on its 0–100 scale, and one pair in 20 by more than 32 (our scans up to August 24, 2026). In the full report, a change under 15 points between two scans is that same photo noise. The four-line calculator is as accurate as the lines you place: a line 1% of face height off moves two thirds by about a point each.
One honest caveat: the hairline is not a detected landmark, so the scan does not score the upper third; its result covers the middle and lower thirds only. The four-line calculator on this page covers all three, because you place the hairline yourself. Photos with the chin tilted up or down degrade the measurement most.
Which third deviation matters most for attractiveness?
The lower third has the clearest published data. In a rating study that varied chin height on idealized male and female faces, small changes went largely unnoticed, and patients and clinicians wanted surgical correction only for large departures from the classical proportion (Naini et al., 2012). Upper third is the most forgiving; styling can hide most of it.
Can I fix my facial thirds without surgery?
Partially. Lower third is the most fixable: chin filler, jaw posture (mewing), and body-fat reduction can shift the visible proportion noticeably without surgery. Midface is largely bone-anchored — compensatory styling (facial hair, glasses, hair volume) is usually the realistic play. Upper third responds to hair work (cuts, density treatments, hairline shaping). Surgery (LeFort, genioplasty) is reserved for severe deviations with functional issues.
How does facial thirds testing differ from a face shape test?
Face shape (oval, round, square, heart, oblong, diamond) describes the outer contour and width-to-height. Facial thirds measures internal vertical proportions. The two are independent — you can have an oval face shape with a long midface, or a square face shape with perfect 1:1:1 thirds. Together they describe the face in 3D. Run both for a complete read: this test plus the /face-shape-test.
Is the 1:1:1 ratio universal across ethnicities?
The 1:1:1 ideal was derived from European Renaissance proportions and skews toward Caucasian skulls. Measured faces do not follow it in any group: in a study of Chinese, Vietnamese, Thai and North American Caucasian young adults, the three sections of the profile were equal in none of the groups, and in the Asian groups the lower face was shorter than the forehead (Le, Farkas et al., 2002).
RealSmile takes no ethnicity or demographic input: the scan compares your middle third with your lower third, where an even split is the reference. Read a small deviation as normal variation, not a verdict.
Is a long lower third bad?
No single third is "bad" on its own. An elongated lower third — above roughly 36% of face height — is the most common imbalance reported in cephalometric literature (Farkas & Munro, 1987; Naini, 2011), so it is closer to the norm than an outlier; faces with all three sections within 2% of equal are uncommon.
A longer lower third also reads as more mature, and in men it overlaps with masculine markers like jaw dominance. It becomes a styling question mainly when it pairs with a recessed chin — and a beard trimmed shorter under the chin visually rebalances it without changing anything structural.
Why is my lower third so long?
Usually because that is simply how your skeleton grew — an elongated lower third is the most common thirds imbalance in cephalometric literature (Farkas & Munro, 1987; Naini, 2011). Developmental research also associates longer lower-face growth with chronic mouth breathing and malocclusion during childhood; that is a published association, not a diagnosis.
In adults the proportion is skeletal and stable, so what you can realistically change is the read: beard shape, hair volume up top, and photo angle all shift how long the lower third appears.
Can you measure facial thirds without seeing the hairline?
Two of them, yes — the scan never uses the hairline at all. The 68-point landmark scheme has no hairline point, so the middle third (brow line to nose base) and lower third (nose base to chin) are measured directly and scored against each other, and the upper third is left out.
That means bangs, a hat line, or a receding hairline won't break the scan. For your upper third, mark your hairline in the free four-line calculator on this page, or measure hairline to brow by hand; the steps are on this page.
What's the relationship between facial thirds and the golden ratio?
The golden ratio (1.618:1) describes width-to-height and feature-spacing relationships. Facial thirds describes vertical thirds within total face height. They are independent measures. A face can hit 1:1:1 thirds without hitting golden ratio, and vice versa. RealSmile scores both — see the /golden-ratio page for the geometric companion to this test.
Honest caveats
- The upper third is not scored — the 68-point landmark scheme has no hairline point, and a hairline read from a photo is unreliable. If you want it, mark it in the four-line calculator or measure hairline to brow by hand (the steps are above).
- Chin tilt up or down skews lower-third measurement. Use a neutral, eye-level photo for the most accurate read.
- Heavy beard coverage hides the chin tip and inflates the lower third. Trim or use a clean-shaven photo for an accurate baseline.
- Thirds is a single dimension of facial harmony. A balanced thirds ratio doesn't guarantee high attractiveness — it removes one penalty. The other 16 metrics still matter.
Background reading: the rule of facial thirds and the midface ratio in our glossary, plus how the 17 metrics average out across real scans in the State of Looksmaxxing 2026 report.
Terms measured on this page
- Rule of Facial ThirdsThe rule of facial thirds divides the face vertically into three equal regions — forehead, midface, and lower…
- Neoclassical Facial CanonsThe neoclassical canons are a set of nine ideal facial proportions formalized in 18th-century European art — u…
- Midface RatioThe midface ratio compares the vertical height of the midface (brow to upper lip) to other facial thirds — a m…
Test Your Other Metrics
Each test measures a different facial metric. Run them all to get the full picture.
Get the written photo audit — $19.99, 5 pages, emailed in about 2 minutes.
Thirds is one metric. The audit reads your full face, ranks what to address first, and writes a personalized 5-page report on the highest-leverage moves (procedure ROI included).
Get the $19.99 photo audit →Premium Dating Photo Audit · Delivered in 1–2 minutes
Get your dating photos audited.
Rank 2 to 6 photos free and see which one should lead. Then unlock the full written audit on up to 10 photos — each photo ranked with its strongest and weakest metrics, what to delete, three reshoot fixes, a 20-dater AI voter panel + downloadable PDF — for $19.99.
✓ 1–2 min delivery · ✓ 17 metrics measured · ✓ Lead-photo verdict · ✓ Stripe secured