Necklace Length Chart

How Deep a Chain Hangs: One Length, Four Necks

Every row below starts from a tape reading this site does not have and you do not have yet: the measure round your neck. Take it where a chain would sit, snug rather than tight, and read the column that matches it. The depths are measured down from the level where the chain leaves the two sides of the neck — not the throat line, not the collarbone line. A chart that reports a position on the body is describing one of those two; this one describes the third thing.

The identity underneath the cells is short enough to argue with. Half the neck's measure is spent going round the back, so a chain of length L on a neck of measure N has L − N/2 of free chain in front, running as two equal strands. The two points it leaves the neck at are at the sides, so they sit one diameter apart, and a neck's cross-section of circumference N is one diameter of N/π wide. Two strands, a known span between them, one right triangle each: depth = √(f² − (N/(2π))²), where f is (L − N/2) ÷ 2.

Read the last column before reading any depth cell, because it is the reason the table is not one column wide. On an eighteen-inch chain, the four necks in this table sit between 146.0 and 104.0 millimetres apart, 42.0 mm of landing point from the neck alone. The step from the fourteen-inch row to the sixteen-inch row at the same neck is 27.9 mm. The input nobody asks you for is worth more than one nominal size; at the tight end it is worth two of them.

Every depth cell is one subtraction, one division and one square root: f = (L − N ÷ 2) ÷ 2, then √(f² − (N ÷ 2π)²), with L and N in millimetres, rounded to one decimal. Each inch cell is the same unrounded depth divided by 25.4 and rounded separately, so an inch can land 0.01 away from dividing the millimetres beside it. The spread column is subtracted from the unrounded cells, so it can sit 0.1 mm away from taking the difference of the two printed cells in the same row — the sixteen-inch row reads 47.0 here and 47.1 by eye. The eight lengths are the sizes chains are actually cut to; the four neck measures are a plain 40 mm ladder from 300 to 420 mm, chosen to straddle a range of adult necks, and they are not typical values: nothing on this page depends on which column is yours, because the working goes from whatever the tape says. Three idealisations are inside the square root and are stated here rather than assumed away. The chain is taken to lie against the neck over exactly half its cross-section, which is the same assumption as saying the two free runs leave the neck one diameter apart — those are one choice, not two. The free runs are taken as straight strands meeting at one lowest point, so a stiff cable or a heavy pendant that pulls the V narrower sits deeper than the cell says, and a chain that stands off the neck sits shallower. The chest is not modelled: below roughly the collarbone the body intercepts the strand, which is why the thirty-and thirty-six-inch rows are depths of a hanging line rather than predictions of where metal ends up on a person. One parameter is genuinely this page's own — the half-wrap. Moving it to 0.6 and 0.7 of the cross-section moves the fourteen-inch cell at a 300 mm neck from 91.0 to 75.1 and 61.7 mm, the eighteen-inch cell at 340 mm from 133.0 to 115.7 and 100.5 mm, and the twenty-four-inch cell at 380 mm from 200.9 to 181.9 and 164.7 mm. The three moves are 29.3, 32.5 and 36.2 mm, or 0.58 to 0.71 of the 50.8 mm a two-inch step adds to the chain. Two further readings of the same table, both arithmetic and both checkable: a chain begins to hang free at all only when L passes (1 ÷ 2 + 1 ÷ π) × N = 0.81831 × N, which is 245.5, 278.2, 311.0 and 343.7 mm at the four columns — so a fourteen-inch chain at the 420 mm column has 11.9 mm of slack over that boundary and falls to 28.8 mm; and the back-wrap alone is 47.8% of a fourteen-inch chain at 340 mm but 18.6% of a thirty-six-inch one, which is the whole reason the spread column shrinks as the chain grows.

What Already Sits at the Front of This Question

Start with the part that is not a gap. Eleven write-ups were opened for this page in one pass this month, the ten at the front of the question plus one further down the list read as a comparator, and their visible text runs from 874 words to 4,503. Five of the ten return a hit for centimetres and four of those five return more than one — 13 on a page, 10 on another, 30 on a third, 2 on a fourth, a single hit on the fifth — and exactly one page in the whole set prints millimetres, six of them. Seven of the eleven carry at least one table, nine tables in all. One page, published on 24 September 2026 and the newest in the set, runs a three-column table of inches against centimetres against where the chain lands on the body, line by line from fourteen inches at 35 to 36 cm against the base of the neck to thirty-six inches at 90 to 91 cm. Two pages return no hit for centimetre, millimetre or inch at all, and one of those two is titled as a sizing chart.

The position words are there too, and heavily: the word for the shoulder line appears 44 times across those eleven pages, in phrases about sitting at it, just below it, or at the base of the throat. So a chart that converts a nominal length into centimetres and a landing place is already made, more than once, and one of them is a week older than this page.

Two of the eleven also recommend by face shape rather than leaving the shape out. One has a section on body type, neck size and face shape, and for a round face it offers a name from the same nominal ladder and the instruction to fall below the chin. The other makes face shape its third step and offers twenty to twenty-four inches for a round face, sixteen inches and a short loop for a heart, and an exemption for an oval. Both give sizes in inches and both give names. Neither gives a millimetre, and neither gives a centimetre that was worked out rather than converted.

Here is the part that is a gap, and it is one number wide. Across all eleven pages opened, the count of hits for a face length is zero. Six of them tell you to measure your neck or your body, and none of them puts a facial measurement anywhere in the arithmetic. Every landing place they do give names a part of the body: the base of the neck, the shoulder line, below the chin. None of them names a length of the reader's own face. The contribution below is not centimetres, which arrived here from a conversion of inches long before this page, and not the observation that a face has a shape, which two of the eleven already make. It is the second input and the second unit: your neck as the thing that spends the chain, and your own face length as the thing the landing point is measured against.

The Vertical Your Face Already Has

This site publishes five readings, and one of them is a vertical: face length over cheekbone width, taken with a soft tape from the middle of the hairline to the chin point. Multiply that published ratio by the cheekbone width you measured and you have a face length in millimetres, which is the only vertical this site has any number for. The seven centres below are read off the same calibration that generates every ratio band printed elsewhere on this site, and the three widths are the adult cheekbone band and the working line published where that band lives — this page reproduces none of that dataset, it only uses the two numbers.

The span is the finding. At the 136 mm working line, an oblong outline carries 224.4 mm of face length and a round outline 142.8: the same multiplication on the same cheekbone width puts 81.6 mm between them, 3.29 of this site's own standard deviations. The same arithmetic gives the opposite on the horizontal. On the width axis, the gap between a round jaw and a square jaw at that line is 13.8 mm, or 0.85 of one spread inside a single shape — smaller than the difference between two people who are both called round. On the vertical, the labels separate by more than three spreads.

Each millimetre cell is the ratio in the second column multiplied by the width in that column's heading, rounded to one decimal; each inch cell is the same unrounded product divided by 25.4 and rounded separately. The ratios are on the soft-tape scale — the scale a tape measures on, which is not the scale the photo detector's raw quotient lives on — and they are the centres that generate the seven published bands, so the centre plus or minus the published deviation reproduces that band cell for cell. The published deviation on this feature is 0.18 on the tape scale, which at 136 mm is the 24.8 mm of the last column. Two limits are this site's and are not smoothed over. The round and square centres are identical in this table, to the last printed digit, so a face length separates neither of them from the other; and the heart, diamond and triangle centres are interpolated along a published ordering from one, three and zero measured faces, so read those three rows as positions on a scale. The 128.2 and 148.8 mm columns are the ends of the adult cheekbone band this site works from and 136 mm is its working line; the dataset behind that band is named where this site publishes it and is not reprinted here, and anyone outside the band should use their own width, since the arithmetic is one multiplication.

Where Does a 24 Inch Necklace Fall

Twenty-four inches is 609.6 millimetres of chain, and 60.96 of centimetres, exactly: the conversion is the easy half and several of the pages read for this page have already done it. The interesting number is the depth. At the 340 mm neck the free front is 609.6 − 170 = 439.6 mm, two strands of 219.8, spanning 108.2 mm between the sides, which gives 213.0 mm, or 21.3 cm and 8.39 inches, below the line where the chain leaves the neck. Move the neck and the same chain moves to 224.8 mm at 300 and 188.3 mm at 420, a spread of 36.5 mm. At this end of the ladder one nominal size is worth 26.3 mm, so the four necks in the table are worth 1.39 sizes of landing point, and the only one of those inputs a shopper is ever asked for is the size.

In the other unit, the same cell reads differently depending on whose face is doing the measuring. 213.0 mm of drop is 1.49 of the 142.8 mm face length this site publishes for a round outline at the working line, and 0.95 of the 224.4 mm it publishes for an oblong one. On a 300 mm neck the same cell is 1.57 of the shorter of those two face lengths, and on a 420 mm neck it is 1.32. That is an arithmetic comparison, not a claim about appearance, and nothing below says a longer face wants a longer chain. One cell looks like a match: the 300 mm neck at 24 inches lands at 224.8 mm, while an oblong face at the working line is 224.4 mm long, 0.4 mm apart. The near-equal pair is a coincidence of two round numbers and is treated as nothing more here.

A Necklace Drop Centimeters Chart, Built From Two Readings

A landing point is only useful if you can state it as something you intend. The rows below take the drop you would want and ask how much chain it costs, expressed in a unit off your own face: your face length from the section above, used as a ruler rather than as a diagnosis. Half a face, three quarters, one whole face, one and a quarter — those four fractions are this page's own ladder, chosen to straddle the range in which a chain actually lands, and they are not a recommendation about where anything should sit on anyone.

The inversion is the same identity read from the other side: chain = N/2 + 2√(d² + (N/(2π))²), so a wanted depth d is one square root from a length. Two faces, one neck of 340 mm, and the row at one whole face length runs 475.4 mm of chain for the shorter of the two published face lengths and 631.7 mm for the longer — 18.72 inches against 24.87, a gap of 156.2 mm opened by the face alone, with the neck held constant. The last column is that gap, and it is the number this page exists to print.

Now compare what it costs to be wrong about the other things. At that row, each millimetre of depth is worth 1.87 mm of chain on the shorter face length and 1.94 on the longer. One full standard deviation of this site's own published spread in face length, 24.8 mm at the working line, is worth 46.4 and 48.2 mm of chain, or 0.91 and 0.95 of a two-inch step. A cheekbone width misread by 5 mm moves the face length by 5.3 and 8.3 mm and the required chain by 9.8 and 16.0 mm. So the shape difference of 6.15 inches survives everything the sample can take back: the spread costs about one step, the labels cost six inches, and that ratio is the honest version of an instruction by face shape.

The two face lengths are the published centres for the shortest and longest outlines in the table above, both at the 136 mm working line — they are positions on a scale, not two people. The four multiples are this page's own ladder and carry no claim about where a chain should sit. Chain cells use the same 340 mm neck throughout so the last column isolates one variable; change the neck and the whole row moves: the one-face row at the shorter length runs 451.1 mm at a 300 mm neck, 475.4 at 340, 500.2 at 380 and 525.3 at 420, which is 2.92 inches across the four columns, and 608.8 to 678.3 mm at the longer one. Margins at the one-face row are 1.870 and 1.944 mm of chain per millimetre of depth, from the derivative of the same square root. Sensitivities, all stated wide rather than tightened: one standard deviation of the published face-length spread (0.18 of the ratio, 24.8 mm at 136 mm of cheekbone width) costs 46.4 mm and 48.2 mm of chain, which is 0.91 and 0.95 of a two-inch step; a cheekbone width read wrong by 2, 5 and 10 mm costs 3.9, 9.8 and 19.6 mm of chain at the shorter length and 6.4, 16.0 and 32.1 mm at the longer. The half-wrap idealisation from the first table applies to every cell here as well, and it moves the wanted depth by roughly a third of a step over the range its two ends were given. No cell in this table is a measurement of a reader; all of them are one multiplication, one division and one square root away from becoming one.

Choker vs Long Necklace Face Shape: What the Choice Actually Moves

A short loop and a long one differ in more than depth, and the arithmetic says where the difference is worst. At a 340 mm neck the back wrap alone is 170 mm of chain, or 47.8% of a fourteen-inch loop and 18.6% of a thirty-six-inch one. The same fixed subtraction eats a different share of the price depending on how much chain you bought, and that is why the top of the first table is the least stable part of it. Across the four necks a fourteen-inch chain moves 62.2 mm of landing point, or 2.09 nominal steps, while a thirty-six-inch chain moves 33.4 mm, only 0.43 of the six-inch step that row belongs to. Tight loops are neck-dominated. Long ones are length-dominated.

The half-wrap parameter is also unkind where the chain is short. Moving it from a half to seven tenths of the cross-section costs 29.3 mm at fourteen inches on a 300 mm neck, 32.5 at eighteen inches, 36.2 at twenty-four — so in proportional terms the same assumption eats about a third of the fourteen-inch cell, a quarter of the eighteen-inch one and just under a fifth of the twenty-four-inch one. Anyone reading the fourteen-inch row should read it as a band from roughly 62 to 91 mm and not as a point, and should know that the band has no far end: at fourteen inches on a 420 mm neck a seven-tenths wrap leaves the chain nothing to hang with.

What face shape can honestly do here is narrow, and it is the same thing this site said about a garment's opening: it changes the unit, not the arithmetic. A choker's depth is decided by your neck and by nothing on your face; a long chain's depth is decided by the length you paid for, less a fixed amount your neck takes. Where the outline enters is in the ratio between the landing point and the face you are looking at in the mirror, which is the last column of the third table and the only place on this page where a shape label carries millimetres. No row above says a round face should buy short and an oblong face long, and no row above claims a chain makes a face read longer or narrower than it is — those are the sentences the rest of this category is made of, and they are not arithmetic, so they are not here.

Pendant Length for Face Shape: Where the Names Stop

Two of the ten write-ups read for this page do recommend a length by face shape, and they are worth describing precisely rather than dismissed. Both work on the round, heart and oval outlines; both give real sizes; the sizes they give are inches, taken from the same nominal ladder every seller uses, and the ones they avoid are given by name — the three or four words in the language that stand for a length without stating it. One of them also tells a round face to choose a drop that falls below the chin, which is a position instruction with a landmark and no number, and it is the closest any of the eleven came to asking for a measurement of the reader's face. None of them asks for a face length, and none of them prints a millimetre that was derived rather than converted.

So the gap here is not that the advice by outline is missing. It is that the advice by outline stops at the label while the underlying quantity is a millimetre, and that a reader is left to translate a name into a size without the one input the size depends on. This page supplies the two translations: nominal length to depth, in the first table, from your neck measure; and wanted depth to nominal length, in the third table, stated in your own face length. Put the two together with a chin landmark and the instruction that a round face wants a drop below the chin becomes arithmetic you can check once per person rather than a name to remember — the drop you want is your chin-to-wherever distance in millimetres, and the third table turns that into a chain.

What this page will not do is put a number in a shape's column and call it that shape's size. On the vertical the labels separate — 3.29 spreads between the shortest and longest face length this site publishes — but the separation is in a reader's own proportions, not in a chain length, and the step from one to the other needs a neck measure and a body landmark. Both of those are on the body. This site's five published readings are all taken on a face, and no ratio printed above converts into a depth without one measurement of the body it belongs to, which is why the third table asks for your neck and the section above it asks you to state the drop you want rather than inherit one.

What Changed Between July 2022 and September 2026

Seven dated entries, sorted by the day each one was published, six of them from the front of the question and one from further down the same list, opened as a comparator. Four pages at the front of it carry no date this month's pass could locate — not evidence that they have none, and they are therefore absent from this time series rather than entered on a guessed day, which is a limit on this page's reading. Two of the eleven pages blocked a plain fetch and were read in a real browser instead, so their word counts include navigation and folded sections and are not strictly the same measure as the other nine.

What moves across the four years and two months is the vocabulary and the units. The 2022 row already has a face-shape step and gives it inches; the 2024 row prints millimetres and leaves the shape out entirely; the newest row, a week before this page, has the most complete length table in the set — inches, centimetres and a position on the body, line by line — and no shape content at all. What does not move in any row is the count of inputs: every one of these tables is a chain of arithmetic from a single number, and none of them asks the reader for a second one, let alone for a measurement of the face.

Read the first row as late as its revision date allows and the span from the oldest publication day to the newest is still 1,528 days — four years and two months, not four and a half; the two newest rows are 29 days apart, and three of the seven carry a revision date as well as a publication date, 1,452, 706 and 107 days after publication. Dates were taken from each page's own statement — a printed line, its metadata or its structured data. The columns describe what was read in one pass this month and are not quotations. Eleven write-ups were opened for this comparison and seven are dated here; the four that are not sit in the earlier section with all their counts intact, and their absence from a time series is this page's limitation rather than a finding about them. The 2022 row is not one of the ten at the front of the question: it was opened because it is the only other page in the reading that makes face shape a step of its own.

How Far to Trust Each Row

The first thing to distrust is the geometry, and it is distrusted by one division. Every depth in the first table hangs on the choice that the chain lies against half the neck's cross-section. Move that to seven tenths and the fourteen-inch row at a 300 mm neck falls from 91.0 to 61.7 mm; move it to six tenths and it sits at 75.1. Those three costs are 29.3 to 36.2 mm across the three cells named in the note under the first table, between 0.58 and 0.71 of the step the ladder is built from. Each cell is therefore a band of roughly two-thirds of a nominal size, printed as a line because a band of forty numbers would be unreadable. Know which of the two you are reading.

The second is the round trip from a face. Face length here is a published centre multiplied by a published width, which means it is a position in a sample, not your face, until you take the tape. The centre plus one spread is 24.8 mm at the working line, and the third table carries that through: one spread of face length costs about 46 to 48 mm of chain — a step. Anyone inside a printed row who wants their own number should multiply the ratio they measured by the cheekbone width they measured, and the multiplication is the whole method.

The third is what none of this says. A depth is a place where metal sits, and every row above is about a place. This page makes no claim that any of them changes how a face reads to anyone, and that is not a hedge — it is that the quantity is not in the arithmetic, so it cannot come out of it.

Four things to do with the tables, in order: measure the neck and read one column, not the table; measure cheekbone width and look up your ratio once, so the face-length row is yours; decide the drop you want in millimetres before choosing a length, and use the third table's inversion rather than the first table's rows. The fourth is the boundary check — multiply your neck measure by 0.81831 and if the chain you are looking at is shorter than that, it will not hang at all.

• Neck measure first. Half of it is spent before any of the chain is free, and it is the input every chart on this question omits.
• One row, one column. Any two cells in the first table are the same square root with different numbers in them; redo one of them at your own measure and you have checked the page rather than trusted it.
• Face length as the unit, not the diagnosis: 142.8 to 224.4 mm at the working line is why one wanted drop costs 156.2 mm of chain more for one outline than another.
• A body landmark is still missing, and this site does not have it. Chin to the line where a chain would sit is a measurement of a neck, and no ratio above supplies one.
• Read the sensitivity before the number: two-thirds of a step for the wrap, one step for one spread of face length, and 4.3 mm for a 10 mm error on the neck tape in the fourteen-inch row on a 340 mm neck — the same 10 mm moves the thirty-six-inch row by 2.8 mm.

Questions this page answers