Hats for Face Shape: The Two Divisions That Set Crown Height and Brim Width

Neither division appears in the positions that rank for this subject. The two words for crown and brim occur 190 times across the ten pages read on 6 October 2026, and a millimetre figure occurs zero times in all ten — the one clause that puts a brim and a number together belongs to a single item's description.

What a Crown Adds to the Length Ratio

The length ratio is a division, so anything added to the length has to be divided by the same width to be understood. Take your lift `a` and your cheekbone width `C`, and the change to the ratio is `a ÷ C`. That is the whole of the first table. Rows are the lift in millimetres with the inches beside them; columns are the adult cheekbone band this site publishes — 128.2 mm and 148.8 mm at its ends, with the 136 mm working line in the middle — because the same 25 mm of crown is a different change on a narrow face and a wide one.

Hat styles by face shape is what the published lists actually sort, and the arithmetic above is in none of them. Of the ten positions read on 6 October 2026, the two words for crown and brim appear 190 times between them and a millimetre figure appears zero times in all ten. One of those pages does print a face as a ratio, a length-to-width figure used to decide which outline you are, and it stops at the naming: the crown and brim sentences that follow carry no number at all.

Read the table in whichever direction your question runs. If you know the effect you want, find the change in ratio and read the lift across that row. If you have a hat in front of you, measure its lift, find your row, and read what it does at your own width. Anyone whose cheekbone measure falls outside the published band should use their own number: every cell is one division, and it can be redone on the back of an envelope in the time it takes to try a hat on.

Every cell is the row's millimetres divided by the column's cheekbone width, to three decimals, and the inch column is the millimetre row divided by 25.4 and rounded separately — so an inch can sit one hundredth away from dividing the millimetres printed beside it. The three widths are the adult cheekbone band this site publishes: 128.2 and 148.8 mm are one standard deviation either side of the measured means, and 136 mm is the working line its eyewear pages use; the dataset behind that band is named where the band is published and is not reprinted here. The published length band is 1.05 to 1.65 with a standard deviation of 0.18, which is what makes one full deviation of that ratio equal to 23.4 mm of lift at the narrow width, 24.8 mm at the working line and 27.1 mm at the wide one — 0.92, 0.98 and 1.07 in. Read the other way: 0.18 of ratio is worth 23.1, 24.5 and 26.8 mm of crown. The whole published span of 0.60 is 3.29 deviations, which is 76.9 to 89.3 mm of lift depending on width. Two things are this page's own and are labelled as such. The first is the landmark — a crown's height measured as the excess above the top of your own head rather than as the height of the object — because no page, this site included, has ever published a measurement of a hat standing on a person. The second is the ladder: 5 to 45 mm in 5 mm steps is the edge of a worksheet, not a finding, and there is no published band for a lift to sit inside. Sensitivity, so you know what a bad reading costs: 5 mm of error in the lift moves a row by 0.037 of ratio at the working line and reorders nothing; the same lift read at 128.2 mm gives a cell 16.1% larger than at 148.8 mm; and 3 mm of error in the width itself costs 0.004 of ratio, which is under a quarter of a row step.

How Tall Is Your Decision, Not the Ratio's

A search for hat crown height face shape asks the ratio for something it cannot hand over. The division returns the size of a change, not a target: nothing in 1.05 to 1.65 says whether you should wear 20 mm of lift or 40 mm, and no page in print or on the web — this one included — publishes a crown height that stands for a population of faces.

What the arithmetic does settle is how much room you are spending. This site publishes the oval outline at a centre of 1.45 with the same spread of 0.18, so the oval band runs 1.27 to 1.63 and the slack above the centre is exactly one deviation. At the working line that slack is 24.5 mm of crown lift, which is a little under an inch: a hat standing that far above your head spends the entire upper half of the published oval band in one fitting. The same lift is worth 23.1 mm of crown at the narrow end of the width band and 26.8 mm at the wide end, so how tall is too tall is a width reading before it is a style reading.

That is a statement about a silhouette and nothing else. A hat does not change what your face measures — the ratios in this site's calculator are taken on a bare head, and a crown sitting 25 mm above it enters no classification anywhere on this site. What it changes is where the top of your head reads as being in a photograph, and the use of the table is that it lets you choose that change instead of meeting it in a mirror you did not expect.

The Three Numbers That Follow the Head, Not the Face

Three quantities in every hat are set by the skull and the sweep, and none of them contains anything measured on a face. The reading underneath all three is taken around the head, over the hair. Work out the size a crown has to fit ↓

Every hat carries a second column of quantities, and the ratios do no work in it. The ring that has to close around your skull, the depth your head takes up inside the crown, and the sweep a brim is cut to all follow the head. Two printed millinery course manuals from 1915 are the clearest statement of that column written out. One measures a support wire 12 in (304.8 mm) from one side of a crown over the top to the other, with 2 in (50.8 mm) turned down at each end so the piece comes to 16 in (406.4 mm). The head itself is drawn by wire as an oval 8½ in front to back and 7 in side to side, which is 215.9 mm by 177.8 mm, and that oval appears in both parts. The other strikes a flat circular brim piece 18 in (457.2 mm) across by measuring 9 in (228.6 mm) twice, and cuts a bias strip 6 in (152.4 mm) wide for a rolled edge.

Sort that ladder against the two divisions above and three kinds fall out, which is the step none of the current ten takes. The band that closes around the skull, the depth the head occupies inside a crown and the circle a brim has to clear are the head's figures. A chooser's figures sit next to them: the same manuals allow a box crown anywhere from 1 to 5 in — 25.4 to 127.0 mm — and take 3 in, 76.2 mm, for the worked example because an example needs one number rather than a range. Only the two divisions in this page's tables belong to the face, and they are the only ones that decide whether a given hat reads as tall or as wide on a given head.

The head column also sets a floor the ratio cannot go under. In the manuals' worked example the head's own oval is 177.8 mm across at the level a brim is attached. The brim widths the forehead arithmetic produces at the working line run from 128.8 mm to 162.8 mm, so the ratio's answer is 49.0 mm smaller than that head at the narrow end of the published band and 15.0 mm smaller at the wide end. The head wins. What the ratio adds is the knowledge of where the edge will actually fall against your temples once the head has had its way — which is the next table.

Millimetres are the printed inches multiplied by 25.4 and rounded to one decimal, and a range cell is its two ends done separately. Three of these rows are why the column matters. The support wire and the head oval are fixed by the skull, and no ratio appears in either. The box crown's 1 to 5 in is left to judgement by the manual itself — and that whole ladder, 101.6 mm wide, is broader than the entire published length-ratio band converted into crown lift at the working line, which comes to 81.6 mm. A catalogue's word for tall and this page's number for a ratio are two different ladders, which is why a crown height quoted without a head under it tells you nothing about a face. The 27 in band is not a head measurement either: it is a head's measure plus a 3 in lap, and the overlap is stated because a cloth ring has to be joined somewhere. No head-size range is printed on this page. This site publishes one on the calculator named above, and a standard figure from a 1915 course manual is not the same object as this site's working band, so the two are kept apart rather than averaged.

How Wide Should a Hat Brim Be

Hat brim width is the second quantity that can be run against a face, and the measurement it belongs to is the forehead rather than the cheekbones. The visible edge of a brim either falls short of your temple or stands past it, and the distance is read per side, which is why everything below is halved: a total width and an overhang are different numbers, and much of the advice in circulation uses the word for one when it means the other.

A forehead's width is a ratio times your cheekbone width. The forehead ratios this site's rule set spans on a soft tape reach from 0.80 at the narrow end of its seven outlines to 1.05 at the wide, and those two ends are published centres rather than a range invented for a table; the 1.00 row is the case where a forehead is exactly as wide as the cheekbones, which is why its cells repeat the cheekbone band's own ends. The spread published with that ratio is 0.08, which at the working line is 5.5 mm per side (0.22 in), and the whole 0.25 span of the band is 34.0 mm across the face, 17.0 mm or 0.67 in per side. The table multiplies each ratio by the three published widths; the last column is one self-defined question, how wide a brim has to be to clear your forehead by 10 mm (0.39 in) on each side.

Read that last column against the head floor before treating it as an order. The 10 mm allowance is this page's own and not a finding: set it to 5 mm per side at the working line and the answer moves from 156.0 mm to 146.0 mm (5.75 in); set it to 15 mm and it moves to 166.0 mm (6.54 in). What is not arbitrary is what the table shows next to it. A single fixed brim width does not give a single look. Hold the 156.0 mm row against both ends of the published band and the same hat stands 6.6 mm past one forehead and 23.6 mm past another — 3.58 times the overhang — with nobody's styling decision made anywhere in between.

Every cell in the three middle columns is the row's ratio multiplied by the column's cheekbone width, rounded to one decimal, with the inches taken from the unrounded product and rounded separately. The last column is the 136 mm cell plus 20 mm, which is this page's own 10 mm allowance counted twice. Nothing published fixes that allowance, and it is stated as this page's rather than as a typical figure because the published record does not contain the measurement it would be typical of: the only clause in the ten positions that puts a brim and a number in one sentence belongs to one item's description, gives a width of 10 cm — 100 mm, 3.94 in — and does not say whether that is the sweep or the whole distance across, nor name a face it applies to. A column of head sizes printed beside a navigation label that mentions wide brims is not a brim figure, and this page does not count it as one. Sensitivity on the two self-defined ends: the same 156.0 mm brim is 21.8 mm narrower than the 1915 example head across, so for a head that size the width is not available at all and the ratio is describing where an edge you cannot buy would fall; and a forehead measure 6 mm wider than your row gives an overhang 3 mm smaller on each side, which is over half of the whole allowance.

What a Century of Written Advice Left Unmeasured

The table below orders what could be dated by the date each source carries, and lets the undated ones sit at the bottom with their word counts so the gap in the record stays visible instead of being smoothed over. Two printed manuals and four of the ten positions carry a date; six positions carry none, and one of those shows only a copyright year in its footer. Two of the ten are a size-picker page and a product filter rather than text, and they get no row.

The last column is the one to read, because across the whole span it says the same thing. From the first of the 1915 imprints to the most recent dated position is 40,198 to 40,562 days depending on which day of that year the manual left the press, which is 110.06 to 111.05 years, and in all of that distance the crown word and the brim word never once get a millimetre attached to them. Inside the recent rows alone the finding repeats: 1,208 days, 3.31 years, between the 2022 guide and the January 2026 one; 852 days, 2.33 years, between the 2022 guide and the piece that carries a machine-readable timestamp of 29 January 2025; 247 days, 0.68 years, between the May 2025 article and the newest row. The 2026 position answers with two tables of names. The 2022 position answers with a list of outlines and no figure beside any of them. The timestamped 2025 piece uses the words for ratio and proportion eight times and never once with a number after them. The 1915 manuals answer in wire and cloth. Ask any of the four how far a crown should stand above a particular head and each reaches for a different kind of object.

Two positions carry no row and are accounted for here rather than dropped: one is a shopping page whose hat advice is a filter over outlines, and one is a size picker whose only figures are head sizes in a navigation column. The 1915 rows are counted as one row because both parts were issued the same year by the same course; their figures appear in the second table with the conversions. The word counts are counts of text after markup was stripped, so a page that carries a catalogue and a story list reports a large number for reasons that have nothing to do with depth.

What These Tables Do Not Settle

Do hats flatter face shape is a question these two divisions cannot answer, and it is worth being exact about why not. A division tells you how far a thing moves; it does not tell you which direction you wanted. Nothing in 0.037 or in 23.6 mm says whether you would prefer a taller crown, and this site publishes no measurement of preference of any kind — the five readings in its rule set are widths, an angle and a ratio, all taken on a bare face.

Three further things sit outside the tables. They are silent about how a hat sits: a crown pushed back on the head and one pulled down to the brow carry the same measurement and change the same lift, and the 1915 manuals note that a covering strip goes on with slack so an edge can be shaped afterwards, which is a fitting variable rather than a ratio. They say nothing about asymmetry, because both divisions are taken straight across. And they suppose a brim's edge level with the temple at the sides — a dipped or swept brim moves the reading at the front while leaving the side number exactly as measured.

What survives all of that is the shape of the answer, which is a division and not a list. If a page on this subject cannot tell you which two numbers to hold against a mirror, it is answering a different question, and the chains above are everything in it that can be settled with a ruler while your face stays precisely as it is.

• Two divisions, one per ratio: lift ÷ cheekbone width, and brim width read against a forehead that is itself a ratio times that width. Everything else a catalogue offers you is a name.
• The head's column wins any argument about a floor. A brim cannot be narrower than the skull it sits on — 177.8 mm across in the 1915 example, against 128.8 to 162.8 mm from the forehead arithmetic at the working line.
• A hat never changes your measured ratio. It changes where the top of your head reads as being, which is the only reason the number is worth taking at all.
• If your lift or your width falls outside a table, do the division instead of the lookup. The grids are worksheet ends, and this site publishes no band for a crown standing above a head because it has never measured one.

Questions this page answers