Measuring a garment flat: the method question that decides two centimetres

The flat table we keep, and how it gets filled in

Every garment comparison on this site comes back to 5 numbers taken on a flat surface. The table below is the one we use, and it is a method rather than a dataset: the same 5 rows work on a shirt, a jacket or a pair of trousers.

Filling it in for 1 garment takes about 5 minutes, and 4 of those minutes go on finding a flat surface and a tape that does not stretch. A soft tailor tape held in the air reads 1 or 2 cm longer than the same tape laid on a table.

A tape and a wall are enough for the whole method, and 2 tools make it easier: a hard ruler for short spans and a table edge to hold the garment still. Neither is essential, and both cut the reading error that comes from measuring on a bed or across a knee.

The table is deliberately short. Charts with 15 rows encourage people to copy numbers that decide nothing, while 5 rows cover the differences that make a garment unwearable rather than merely different.

How each number is taken, and why the doubling rule exists

Doubling is where the first argument usually starts, and it rests on 1 rule: a chest measurement taken seam to seam on a flat garment is half a circumference, so comparing it against a body circumference without doubling it makes a normal shirt look half its size.

Fixed points matter as much as doubling. Chest is taken 2 cm below the armhole rather than at the widest point, because the widest point moves with the cut while the armhole seam stays where the pattern put it.

Fabric behaviour sits underneath all of it. The same shirt measured straight out of a packet and again after 2 hours on a hanger can differ by 1 cm, which is why the method says smooth rather than stretch, and why a single measurement is a snapshot rather than a verdict.

• Lay the garment flat and button or zip it, because an open front measures wider at every row.
• Smooth the fabric with a hand instead of pulling it, since a stretched knit holds the stretch while the tape is on it.
• Measure to the nearest 0.5 cm; finer precision on a soft garment is noise rather than information.

Doubling is also where a chart and a garment appear to disagree when they do not. A listing publishing a flat 52 and a listing publishing a circumference of 104 describe the same shirt, and reading 1 as the other turns an ordinary garment into an impossible one. The [sizing topic](/table/sizing/) treats that reading as the first thing to settle.

Notes on the rows that get read wrongly

2 rows cause most misreadings, and both of them are about where a tape starts rather than about what it measures.

1. Length: from the collar seam at the back of the neck, not from the top of the collar and not from the front.
2. Shoulder: seam to seam across the back; on a shirt with a yoke, the yoke seam is not the shoulder seam.
3. Sleeve: from the shoulder seam to the cuff edge, along the top of the sleeve rather than underneath it.
4. Waist: at the narrowest point unless the listing marks a different point, and listings often do.

The length row is the one people get wrong by 2 or 3 cm most often, because a collar adds height at the top while the hem stays where it is. A listing measuring from the top of the collar reports a figure that looks 3 cm longer than the same shirt measured from the collar seam.

Sleeve length hides a similar trap on raglan garments, where 1 of the 4 usual seams does not exist. There the measurement runs from the centre back of the neck to the cuff, which spans both sleeves and is not comparable with a set-in sleeve figure.

The hem row is the 1 most charts leave out and most garments need. A 55 cm chest with a 60 cm hem hangs differently from the same chest with a 48 cm hem, and the number that separates the 2 is not printed on any size label.

Outliers: where a flat number stops describing the fit

Some garments produce numbers that are correct and useless at the same time, and 3 constructions do it more often than the rest.

• Drop shoulder: the seam sits down the arm, so the shoulder row reads 4 or 5 cm more than a set-in sleeve of the same fit.
• Raglan sleeve: the shoulder figure disappears and the sleeve figure becomes a span across the body.
• Stretch knit: a ribbed fabric can relax 5 cm across the chest after a day of wear, so the flat figure describes its smallest state.

Read an outlier against the same construction rather than against the same number. A 55 cm shoulder on a drop shoulder shirt is unremarkable, while the same 55 cm on a set-in sleeve describes a far larger garment.

Where a garment is cut from stretch knit, the useful extra figure is the circumference of the fabric stretched to a comfortable degree, which a seller can supply in 1 message and which no published chart contains.

Boots and heavy outerwear behave like a 5th category, because internal volume matters more than outside length. Where a seller can give you the internal length and the internal width, those 2 figures answer a fit question that no flat measurement of the outside can reach.

Coat sizing is the 4th case worth naming, because a coat is worn over something else. A flat chest figure for a coat only means something next to the layer you intend to wear under it, so the honest comparison is coat against coat rather than coat against shirt.

Applying the method to a listing before you buy

The method earns its keep at the point where you still have a choice, which is before the item ships. Four steps, and the whole sequence takes under 10 minutes.

1. Measure the garment you already own and write the 5 figures into the order notes.
2. Compare each figure with the chart row by row, rather than comparing the size label with the size you usually wear.
3. Where a row differs by more than 2 cm, ask the seller to confirm that specific measurement.
4. Where 3 or more rows differ by more than 2 cm, the chart describes a different garment and the label is no guide at all.

A row by row comparison also tells you what kind of difference you are looking at. A garment 2 cm narrower at the chest and identical everywhere else is a slim cut, while one that is 4 cm longer with the same chest is a different style entirely.

If the item has already reached the warehouse, the same 5 numbers turn into a QC question rather than a buying question, and the [QC photo stage](/steps/qc-photos/) is where a seller-confirmed figure can still be checked against the physical item.

Write the answers into the order notes as they arrive. A measurement confirmed by a seller on day 2 is still the same number on day 20, while a figure remembered from a chat thread has a way of turning into a different 1 by the time the parcel lands.

Keep the questions in 1 message. Sellers answer a list of 5 rows more readily than 5 separate messages, and the reply is easier to use later when it sits in one place instead of inside a long thread.

Recording a baseline you will still be using in 2 years

A baseline is worth building once. Two garments that fit well, measured with the same 5 rows, cover most of what a wardrobe needs, because most purchases are variations on something you already own.

Store those figures somewhere searchable rather than in a photo roll, and note the construction beside them. A 52 cm chest on a knit and a 52 cm chest on a woven shirt are not the same anchor, and 1 line of description keeps the 2 apart.

Review the baseline when your own measurements change rather than when a chart looks wrong. A 3 cm change around the waist is normal across a couple of years, and updating 5 numbers is easier than re-measuring a wardrobe, which is the kind of routine the [QC coverage tool](/rig/qc-coverage/) is built around.