Oblong vs Rectangle Face Shape: Which End of the Published Corner Band

The Widths Both Words Share, in Millimetres and Inches

Two of this row's five checks are widths divided by a width, so their units cancel and they come back as plain numbers: your forehead width divided by your cheekbone width, and your jaw width divided by the same cheekbone measurement, taken across the widest part of the face just under the eyes. The published band for the forehead on this row is 0.87 to 1.03; for the jaw it is 0.80 to 1.04. The table below turns those printed edges into the units a tape measure comes in, at three cheekbone widths.

The three widths in the first column are display points inside a range this site already publishes for adult cheekbone width — 128.2 to 148.8 millimetres, named on its round-face glasses page, with 136.0 as the working line. Choosing three of them is this page's presentation decision, not a claim that faces come in three sizes: anything between the narrowest and widest is one of the rows above, and you can redo any cell with your own measurement in place of the printed one.

Each cell is one multiplication and nothing else: a published edge of a band, times the width in the first column. No new measurement, model or re-fit appears anywhere in the table, and the last column is there because of what it does not do. Both words you are choosing between own every cell on this page — the row does not have a version for each of them.

What that means in plain terms: at an average cheekbone width, a forehead between roughly 118 and 140 millimetres and a jaw between roughly 109 and 141 are both inside this row, whether you call yourself oblong or rectangle. Nothing in the tape-measurable part of the row changes with the word.

Four things worth knowing about these cells. (1) The edges are the band this site publishes — the centre plus one tolerance step, rounded to two decimal places — so a reading just outside one is inside the row's tolerance, not outside the shape: at two tolerance steps the middle row above would read 107.4 to 151.0 mm (4.23 to 5.94 in) for the forehead and 92.5 to 157.8 mm (3.64 to 6.21 in) for the jaw. (2) The rounded edges and the unrounded ones differ by at most half a millimetre once multiplied out; the table uses the rounded ones because those are the numbers printed elsewhere on this site. (3) The face-length band for this row, and the length bands of the rows either side of it, are multiplied out on the page comparing square with rectangle — this page does not print that column a second time. (4) The corner band is printed in degrees, not millimetres, and it is identical in all three rows: it does not depend on how wide your face is.

• A face whose forehead and jaw both land inside those two corridors — the two width ranges in the table above — is inside the row on both width checks — which is the measurable version of what every answer in this set means by three widths sitting close to each other.
• Both corridors include a ratio of 1.00. That is worth noticing, because a taper is not required anywhere on this row: a forehead exactly as wide as the cheekbones is inside, and so is a jaw exactly as wide.
• The corridors are wide: 20.5 to 23.8 mm (0.81 to 0.94 in) of forehead room, and 30.8 to 35.7 mm (1.21 to 1.41 in) of jaw room, depending on face size. A single tape reading near the middle of one tells you very little on its own.

Every Dated Answer Found for This Question

This table is the reason the rest of the page can be brief. Each row is one dated, published position on this exact pair of words, oldest first, with whatever number that position offered. Six of the nine rows are this site's own; the three outside the top ten of the search results are marked, and nothing on this page leans on them.

Nine rows, four years, nine months and three days from the first to the most recent. The column on the right is the one to read down.

Four of the answers found for this question could not be read at all: one image-sharing site and two question-and-answer pages sit behind logins, and one hairstyling page now redirects to its own installation script, so that site's article is not reachable in any state. None of them is a row here, and none of them is quoted. Two more pages carry no date at all in their markup, which also keeps them out. A relative age on a comment is not a publication date.

What Moved, and What Did Not

Read down the third column. Across the five rows that are not this site's — a span of four years, six months and nineteen days — the position changes at every step, and every step goes the same way. In 2022 a page lists oblong as one of seven shapes and does not discuss the other word. By 2025 a page merges the two by declaring one an alias of the other. By the summer of 2026 a page merges them and then files the difference as a habit of some sources rather than a fact about faces. The last of those closes with the honest version: the line is not standardised. What never appears in any of those rows is a number attached to the difference.

The last column tells the opposite story: it stays empty on every external row, for the whole four years and nine months, and it is filled only on the four rows belonging to this site. That is the one thing this page can add without arguing with anyone. The pages that answer this question do not have a corner measurement; the page that does have one says, in its own table note, that the corner is the check it trusts least.

There is a second thing hiding in that column. The corner band has been on this site since 2026-09-12, and in every row since then it is the same 110 to 126 degrees — not because nobody edited it, but because all four pages read one set of constants when they are built, so there is no second copy free to drift. Elsewhere on the web the numbers attached to face shapes do move. Here, the number that could have been argued about is the one nobody argues about, and the adjectives are what change.

What that means in plain terms: four years and nine months of published answers to this question, oldest row to newest row on this page, have produced no measurement of the difference, and the one measurement that exists on this site is a range wide enough to hold both readings of the corner.

The Number on the Front Page Is a Door, Not a Line

The most detailed answer in the top ten does contain a number, and it is worth putting next to this site's, because the two do different jobs. Its instruction, quoted exactly, is that if the result of length divided by width is 1.5 or above and the three widths sit close to each other, you are in the rectangle and oblong family. That number opens a door: it takes you into a group whose two names it then declines to separate. It is not a line inside the group, and the page says so twice — once by putting both words on the same side of it, once by adding that its own distinction worth caring about is between rectangle and square.

This site's row is entered at 1.47 rather than 1.50, which is the only place the two statements disagree. The gap is 0.03 on that scale, which on an average cheekbone width is about four millimetres of face length — a strip of ground where one page would let you in and the other would not, and where neither page has anything to say about which of the two words you should use. Above the door, both pages have the same answer: you are in this family, and the word is yours to pick.

So the two statements are not competing accounts of the same line. One is an entry rule, one is a row of five bands, and neither of them divides oblong from rectangle. Reading down the top ten, this is the only quantity anyone attaches to this pair at all, and it points at the family rather than at the boundary.

What that means in plain terms: the number you may have arrived holding is a gate you are either through or not, and it does not care which of the two words you answer with.

Why the Corner Is the One Check a Tape Cannot Give You

Here is the part that changes what you can do with any of this at home. Four of the five checks on this row are lengths: face length, forehead width, cheekbone width, jaw width. You can take all four with a soft tape, divide two of them by the cheekbone number, and compare the results to the table at the top of this page — that is the whole of the corridor arithmetic, and it is reproducible in your bathroom.

The fifth check is the angle at the jaw corner: how sharply the underside of the jaw turns up into the side of the face. It is not a length, so a tape has no way to produce it. On this site it comes from a photograph read by a detector — the tool on this site's home page, which finds the points along the outline and works out the turn between them; the same detector publishes its own limit on that reading, and the limit is severe. Across the fifty photographs this site calibrated on, meaning the fifty it measured in order to set its numbers, the angle did not separate a round face from a square one — the two groups landed within a degree of each other, where the written rule set puts them twenty-four apart. Its own table note therefore calls the angle a weak check and the ratios strong ones.

Put those two facts in the same sentence and you get the thing no page in the top ten says. The one reading the two words are ever said to differ on is the one reading nobody at home can take a tape to, and on this site it is also the reading that has been measured least reliably. That is why the answers you are up against are adjectives. An adjective is what you write when the quantity is real, load-bearing, and not available to the person reading.

What that means in plain terms: the tape settles which row you are in. It cannot settle which word you are, and this site does not claim it can.

• What a tape can settle at home: face length divided by cheekbone width, forehead divided by the same, jaw divided by the same, and chin width divided by jaw width.
• What it cannot settle: the corner's angle. Nothing in the four tape measurements converts into it, which is why the last column of the table above has no millimetres in it.
• How this site handles that anyway: the five readings are weighed together and a borderline face is reported with the margin against each candidate, rather than as a verdict.

How Much Room the Two Adjectives Have

If you insisted on drawing a line inside that band, what would you be drawing with? The band is 110 to 126 degrees, and it is one band, not two: cut that printed range at its own middle and each half is eight degrees wide. That is the whole of this section's arithmetic — one published range, halved once. Two faces whose corners differ by two degrees are therefore inside the same half, and so are two faces that differ by six.

Be clear about what that arithmetic is and is not. The halves are this page's way of showing how much room there is, not a rule this site applies; nothing here cuts the band, and nothing on this page says which end belongs to which word. The site's own published sentence is softer than that: within that row's jaw-angle band the tighter angles are what people usually mean when they say rectangle. Usually, in what people mean — that is a report about usage, not a threshold.

The reason to print the width anyway is that it puts a number on the vagueness rather than leaving it vague. Every external answer in the table above says the distinction is slight. This row's band says how slight: the whole distance between a corner sharp enough to be called rectangular and a curve soft enough to be called oblong fits inside one half of one weak reading's published band.

What that means in plain terms: however finely two people mean those two words, the range they fit inside is wider than the difference between them.

Where This Site Files the Two Words

This site is not an outside observer of the renaming it is describing, and pretending otherwise would be the least useful thing on this page. Its photo detector's result page for this family is titled with one word. Its long guide to the shape uses that same word, and the word you searched with appears there only inside a bracket. Both of its pages that use the other word in their address show up when you filter the blog by the first word — that filing is deliberate, and it is written into those pages themselves.

The page comparing square with rectangle keeps a full section for this problem, including the sentence that if you keep arriving at pages about oblong faces when you searched with the word rectangle, nothing has gone wrong. That section is the place to read it; this page will not say it a third time. What belongs here is the narrower consequence for a number: because the two words sit inside one row, every band on this page is one row's band, and there is no second copy of the table to compare them with.

There is one surface where you will notice this most. The result pages for confused-shape pairs are generated from a registered list of which shapes get compared, and rectangle is not among the seven names on this site, so there is no oblong-versus-rectangle result card to wait for. Your two words arrive as one row on one page.

What that means in plain terms: if a search you ran produced pages about a different word, you were not sent to the wrong place, and the numbers on those pages are the numbers on this one.

What To Actually Do With Four Measurements

Measure first, decide about words afterwards, and the measuring steps are on the page that holds them; this page does not repeat them. The short version: hair back, camera level, then four lengths with a soft tape — face length from hairline to chin tip, and the three widths across the forehead, the cheekbones and the jaw — and each of the three width readings divided by the cheekbone reading. Face length is measured from the hairline, which is why hair covering the forehead changes that one number without changing your face.

Then read your four results against the table at the top of this page, not against a single target number. If your forehead and jaw land inside those two corridors at your own cheekbone width, you are in this row on both width checks, and the word is a preference. If your length ratio sits near the start of the row's length band, the useful question is not oblong-versus-rectangle at all: it is the one the square-versus-rectangle page answers, because that is where a wide unused strip between two rows is laid out.

And if you want the corner quantified rather than described, that is the one reading the tape cannot give you, so the only place on this site where it is produced is the photograph you take. Expect that answer to arrive with a margin rather than a verdict, and expect it to be reported as the weakest of the five checks — which is exactly how this site words it.

What that means in plain terms: check the two corridors with a tape, use the length bands to decide whether you are actually near a real boundary, and let the corner be the one thing you do not insist on measuring yourself.

What This Page Cannot Prove

Every millimetre figure here is a published edge multiplied by a published width, so the table inherits whatever is wrong with those bands. They come from fifty photographs, with the shapes' labels drafted by one person rather than agreed between several, and three of the seven rows are interpolated between measured neighbours rather than measured. The corridor cells are also built from the rounded edges: recomputed from the unrounded ones, individual cells shift by up to half a millimetre.

The claim that the corner cannot be produced by a tape is a claim about this site's method, not a claim that no method exists. A clinic with a protractor and a printed tracing could take the angle in some other way; this page says only that nothing in the four tape lengths converts into it, and that the site's own calibration did not resolve it between two rows that the written rule sets twenty-four degrees apart.

The external readings are what could be read, when it could be read. Four of the ten answers reached for this question had no accessible text at all — two behind logins, one behind another login, one replaced by its own installer — and two more carry no date anywhere in their markup, so they are not in the table. One answer was quoted from a page where people upload photographs and ask, and its date there was stated as an age rather than a day, so it is not a row either. Nothing in this page's argument rests on a page whose text could not be read.

The half-band division in the section about the band's width is a display choice of this page. It is not a rule, not a threshold, and not the site's opinion about which word belongs to which end. And the absence of numbers in the other nine rows of the dated table is not evidence that their authors think the difference is nothing: it is evidence that they published no measurement, which is a smaller and better-supported claim.

What that means in plain terms: this page can show you which published numbers are checkable with a tape and which one is not. It cannot tell you what your jaw corner measures, and it cannot tell you which word your stylist will use.

Questions this page answers