A negative tilt
The outer corner sits lower than the inner corner. One angle, measurable from a straight-on photo.
Negative canthal tilt: what it means, and how to check yours
Negative canthal tilt means the outer corner of your eye sits lower than the inner corner. This check measures each eye from a photo on your device, and shows how much a head lean alone can move the reading.
Check each eye for a negative canthal tilt
Front-facing photo, even light, hair off the forehead
- Right eye would read −1.0°Negative tilt
- Left eye would read +5.0°Neutral to mild
Nothing about the eyes changes when you move the lean slider. Only the photo does. The check on this page undoes the lean before it measures.
The short answer
- It is one angle. The line from the inner to the outer eye corner, measured against level. Below zero is negative.
- Each eye has its own. The two eyes of one face often differ a little, so check them one at a time.
- Photos move it. A head leaning 2° toward one shoulder shifts one eye 2° down and the other 2° up.
- No home method has a measured effect. We found no exercise or massage study that measured the angle, and the best age study found the corner height did not change.
What negative canthal tilt means
Your eye has two corners. The inner one, beside the nose, is the medial canthus. The outer one, at the side, is the lateral canthus. Draw a line between them and compare it with level. If the outer end is higher, the tilt is positive. If it is lower, the tilt is negative. The name says nothing about the size of the eye, how open it is, or the lids above and below it.
The number is small in real life. If an eye is 30 mm wide, a whole degree of tilt is roughly half a millimetre of height between the corners. That is why a mirror is a poor judge and why the way a photo was taken matters so much. The bands this page uses come from one 2020 clinical paper, not from a population norm.
Negative, neutral and positive tilt side by side
The angle ranges are the class boundaries from Lee, Choung and Choung (2020), who measured 76 Korean patients aged 13 to 38 from standard front photographs while assessing them for surgery. They are that paper's convention, not a worldwide norm. The last two columns are what gets mistaken for each band in real photos.
| Tilt | Angle | Outer corner | Often mistaken for it | Not ideal for judging it |
|---|---|---|---|---|
| Negative | below 0° | Sits lower than the inner corner | A head leaning toward that eye's shoulder; upper lid skin folding over the outer third | A squinting photo, which pulls the outer corner |
| Neutral | 0° to 5° | Level with the inner corner, or a little above | A slightly negative or positive reading that is inside the photo's own precision | A single selfie: one small lean can cross the zero line |
| Positive | above 5° | Clearly above the inner corner | Eyeliner drawn in an upward wing; a head leaning toward the other shoulder | Photos with makeup along the lash line |
In that study the patients averaged 8.45° on the right eye and 8.42° on the left, with standard deviations of about 3.5 and 3.8 degrees. So in that clinical group the average eye sat about 8 degrees above level. That tells you where one group sat, not where yours should.
How to measure your eye angle at home
You can run a negative canthal tilt test with a phone and a ruler. The steps below are the same ones the check at the top follows, done by hand.
- 1
Take a fair photo
Camera at eye level, face square to it, relaxed, eyes open normally. No eyeliner, no squinting into light, no smile.
- 2
Level your head
Turn on the phone's grid and keep the line through both pupils on a grid line. Leaning is an easy error to make.
- 3
Mark the four corners
For each eye, the point where the upper and lower lids meet next to the nose, and where they meet at the side.
- 4
Measure each eye alone
Width between the two corners, and how far the outer corner sits above or below the inner one.
- 5
Turn it into degrees
Angle = arctan(height ÷ width). A 1 mm drop over 30 mm of width is about −1.9°; 0.5 mm is about −1.0°.
- 6
Retake before you trust it
Take two more photos and measure again. A result that flips sign between photos is level, not negative.
The check above does steps 2 to 5 with 478 landmark points. It undoes head lean from the pose of your head, measures the photo five times to find how precisely it can place the corners, and calls an eye level when the angle is closer to zero than that. Our full canthal tilt test explains the landmark indices and the published normal ranges in more detail.
Sample faces: how far each is from reading negative
Every card is an AI generated face, not a real person, measured by the same check as the tool above. The line in bold is the smallest uncorrected head lean that would make one of its eyes read negative. Tap a card to run that face in the tool.
Faces that measured negative
0 of 16
Including 0 of the 3 faces whose prompt asked for downturned eyes.
Negative after a 3° lean
4 of 16
In a photo nobody corrected. None would flip with 2° or less.
Median lean needed
4.9°
The uncorrected lean at which half of these faces would show one negative eye.
These faces tell you about the image generator, not about people. What they do show is the margin a photo has to work with: the closest face needed a 2.1° lean to show a negative eye, and half needed 4.9° or less.
Why a photo can show a downturn your eyes do not have
Head lean is the one error we can put an exact number on. Tilting your head toward your right shoulder turns the line through both eye corners by the same angle. That lowers the outer corner of your right eye against its inner corner and raises the left one. So the right eye reads the lean lower and the left eye reads the lean higher, and a pair of level eyes becomes one below zero and one above. Move the lean slider in the simulator above to see it.
| In the photo | What it does to the reading | Corrected here? |
|---|---|---|
| Head leaning toward one shoulder | Exactly the lean, down on that side and up on the other | Yes, from the head's pose |
| Head turned to one side | Compresses the far eye, which changes its measured angle | No, retake square on |
| Looking above or below the lens | Moves the lid edges, and the corner points with them | No, look into the lens |
| Squinting into light | Pulls the outer corner | No, use a relaxed face |
| Eyeliner along the lash line | Moves the point read as the outer corner | No, measure without it |
Is a downturned eye angle a problem?
On its own, no. It is a shape, and shapes vary. The spread between healthy people is wide: in a 3D study of healthy young adults (Gao 2025), the standard deviation among Caucasian men aged 18 to 30 was 14.7 degrees. That study measured a different angle, taken at the inner corner, so its numbers cannot be compared directly with the ones on this page, but a spread that wide leaves no room for a single ideal. The bands on this page come from one clinical paper, not from a population.
Things around the eye are worth a doctor's look, and they are about function rather than angle: a lower lid that has gone lax or pulled away from the eye, more white showing below the iris than you used to have, exposure of the eye, or a sudden change on one side. An oculoplastic surgeon checks those. A photo angle does not.
Changing the angle: what has actually been measured
Searches for how to fix negative canthal tilt turn up exercises, taping and surgery with confident numbers attached. This table lists what we could find measured, in degrees where anyone reported degrees. The results are thin.
| Approach | Measured change in tilt | Evidence behind it | Not for |
|---|---|---|---|
| Exercises, massage, mewing, taping, gua sha | None measured | Eleven papers on canthal tilt with these words, zero interventional studies | Anyone hoping to change the angle itself |
| Filler around the eye | None measured | No filler study we found reports canthal tilt | Expecting a number of degrees from it |
| Barbed threads | +2.2° at 6 months, +1.5° at 12 | 22 Asian women, uncontrolled, bruising in 36%, swelling in 23% | Anyone wanting a permanent or at-home change |
| Canthoplasty or canthopexy | Not reported in degrees | Outcomes are published as millimetres of lid position or success rates | Anyone expecting a number of degrees |
| Eyeliner and lash placement | Look only, the corner stays put | Changes where the eye appears to end; it also moves the point a photo tool reads | Anyone measuring: take it off first |
How to fix negative canthal tilt without surgery
There is no measured way. The broadest review of conservative facial techniques, twelve studies and 321 women, found the evidence insufficient to establish that they work, and it did not measure canthal tilt. What you can change without surgery is how the eye reads: eyeliner and lash placement shift where the eye appears to end, and they do it without touching the corner. If you want to see your real angle, remove them before you measure.
Canthoplasty and canthopexy are different operations
Both work on the lateral canthal tendon, which holds the outer corner to the bone. The distinction is whether the tendon is divided, and pages that call one of them the only real fix skip it. We found no published before-and-after tilt in degrees for either, so a promised number of degrees has nothing behind it. Both are conversations for an oculoplastic surgeon, usually about lid function first.
Downturned, droopy or hooded: three different things
Droopy or downturned-looking eyes
How the outer third of the eye reads. Can come from lid skin, even with level corners.
Hooded lids
Upper lid skin folding over the lid crease. Medically dermatochalasis. It changes the look, not the corners.
Does the outer corner drop with age?
It does not appear to. A study of 320 people aged 10 to 89, twenty men and twenty women in each decade, found the height of the outer corner was not affected by age, and no difference between the sexes. The eye around it does change: the opening lengthens by about 3 mm between 12 and 25, then shortens by about 2.5 mm between 35 and 85, the lower lid comes down around a millimetre, and the outer corner moves inward rather than downward. An older eye can look more downturned while its tilt has stayed put.
Why celebrity eye-angle lists disagree
Lists of negative canthal tilt celebrities are built from press photos, and press photos are often taken with heads turned or leaning. A 3° lean moves one eye 3° down and the other 3° up, which is enough to put the same person on a negative list and a positive list from two different red carpets. The simulator above shows it with any starting angle. We have not made a list of our own, because one photo of a person cannot settle their angle, and measuring someone from a picture they did not take for that purpose would put a number on a stranger. If you want an example to compare against, the sample faces above are measured the same way as yours.
Questions about a downturned canthal tilt
Is negative canthal tilt bad?
+
It is a description of where one corner of the eye sits, not a health finding and not a score. There is no agreed cutoff for a normal angle: the positive, neutral and negative bands this page uses come from one 2020 paper on 76 Korean patients being assessed for surgery. The spread between healthy people is also wide. In a 3D study of young Caucasian men the standard deviation alone was 14.7 degrees, on a different angle convention. What does matter clinically is function: a lower lid that has gone lax or pulled away from the eye, more white showing under the iris than before, or a sudden change on one side. Those are questions for an oculoplastic surgeon, and a photo angle cannot answer them.
How do I know if I have negative canthal tilt?
+
Take a photo straight on, at eye level, with a relaxed face and no eyeliner, then look at each eye on its own. Find where the upper and lower lids meet next to your nose and where they meet at the outer edge. If the outer meeting point sits lower than the inner one, that eye is negative. By hand, measure the height difference and the width between the two corners: one millimetre of drop over an eye 30 mm wide is about minus 2 degrees. The check at the top of this page does the same with 478 landmark points, undoes any head lean first, and tells you when the angle is too close to level for the photo to decide.
Can you fix negative canthal tilt?
+
The only measured change we found is from barbed threads, and surgery is not reported in degrees. A search of the medical literature for canthal tilt alongside exercise, massage, mewing, taping or gua sha finds eleven papers and no interventional study. No filler study measures the angle. The one measured change we found was barbed threads in 22 Asian women: from 2.1 to 4.3 degrees at six months and 3.6 at twelve, in an uncontrolled series with bruising in 36% and swelling in 23%. Canthoplasty and canthopexy exist, but their results are published as millimetres of lid position or as success rates, so a promised number of degrees has no study we could find behind it. Makeup changes how the eye looks, not where the corner is.
Does negative canthal tilt get worse with age?
+
The corner itself does not appear to drop. The best age study we could find measured 320 people aged 10 to 89, twenty men and twenty women in every decade, and found the height of the outer corner was not affected by age, with no difference between the sexes. The same authors note that the idea of the outer corner sinking with age had never been backed by measurement. What does change is the eye around it: the opening shortens by about 2.5 mm between 35 and 85, the lower lid descends around a millimetre, and the outer corner moves inward rather than downward. So an eye can look more downturned with age while the corner has stayed where it was.
Why is only one of my eyes negative?
+
The two eyes of one face almost always differ a little, and a photo adds its own difference. A head leaning toward one shoulder lowers the outer corner on that side and raises the other one by exactly the lean, so a 2 degree lean can turn a level pair into one eye below zero and one above. This check undoes that lean from the pose of your head before it measures. What it cannot undo is a head turned to the side, which foreshortens one eye. Across the AI generated faces we measured, the gap between a face's two eyes was never larger than the photo's own precision, so a one-sided result of a degree or two is more likely the photo than the eye. Retake it square to the camera before trusting it.
Is negative canthal tilt the same as downturned eyes?
+
They overlap but are not the same. Negative canthal tilt is one measured angle: the outer corner of the eye sits lower than the inner corner. Downturned eyes is used more loosely, for how the whole eye reads, and an eye can read that way with level corners, for example when upper lid skin folds down over the outer third. That fold is hooding, which the medical literature calls dermatochalasis, and it is about lid skin rather than corner position. So a person can have downturned-looking eyes with a neutral tilt, or a slightly negative tilt that nobody would call downturned. The check here measures only the corners.
Sample statistics measured 2026-09-16 and on page load from the same AI generated faces; median canthal tilt across the shape set +4.9°.
Negative vs positive canthal tilt, converted to millimetres
In millimetres, the difference between negative and positive canthal tilt is small. The height of the outer corner above or below the inner one is the eye's width times the tangent of the angle. The table does that sum for three common eye widths, so you can check a ruler measurement against it directly.
| Tilt | Band | 28 mm eye | 30 mm eye | 32 mm eye |
|---|---|---|---|---|
| −4.0° | Negative | −2.0 mm | −2.1 mm | −2.2 mm |
| −2.0° | Negative | −1.0 mm | −1.0 mm | −1.1 mm |
| −1.0° | Negative | −0.5 mm | −0.5 mm | −0.6 mm |
| +0.0° | Neutral | 0.0 mm | 0.0 mm | 0.0 mm |
| +2.0° | Neutral | +1.0 mm | +1.0 mm | +1.1 mm |
| +5.0° | Neutral | +2.4 mm | +2.6 mm | +2.8 mm |
| +8.0° | Positive | +3.9 mm | +4.2 mm | +4.5 mm |
On a 30 mm eye, the whole neutral band from 0° to 5° is about two and a half millimetres of corner height, and a mildly negative eye at minus 2° sits about one millimetre low. Differences that small are easy for an eyeliner line or a slight head lean to hide, which is why makeup has to come off before you measure, and why the same person can read negative in one photo and neutral in the next.
Is a negative tilt the same as prey eyes?
Not by itself. A negative reading here means one outer corner sits below its inner corner. In the hunter eyes test that lowers one of four measurements, but the brow gap, the lid over the iris and the opening are scored separately, so an eye with a slightly negative tilt can still lean hunter if those three sit at the hunter end.
The same photo cautions apply. A head lean of two or three degrees moves the two eyes in opposite directions, which is why that test scores each eye on its own and shows a range around its number before it names a band.