Multisize me
Grading a knitting pattern for multiple sizes

You've designed a new sweater, and you'd like to publish the pattern. The only trouble is that you've only designed it to fit one size: yours. Now what?

Presumably by this point, what you have is a set of notes or pattern instructions for one size. You worked them out, perhaps, by trial and error, or by computation and tweaking as you go. In fact, even if you do characterize your design technique as trial and error, your design was probably still underscored by certain fundamental calculations: how many stitches to cast on here, the circumference of the sweater there, and the length between the two points. (If you don't have any notes or instructions jotted down for your prototype garment, sorry, can't help you there: you'll need to work backwards to figure out the key measurements, stitch and row counts for your original garment. And then, if you want to make a habit of replicating or multi-sizing your patterns for others, change your note-taking habits!)

Given this starting point, grading your knitting pattern for other sizes amounts to three steps that should sound somewhat similar to what you just did:

1. Figure out the key finished measurements of the original garment, and the ease involved for each of these measurements.

2. Determine what these key finished measurements will be for each size of the pattern.

3. Compute the stitch and row counts for each of these measurements.

4. Insert these numbers into your written instructions, tweaking as necessary.

These four steps might sound a little simplistic, but in reality, multisizing a pattern is just more arithmetic and logic piled on top of the arithmetic and logic you used the first time around. If you were able to figure out instructions and numbers for one size, you can do it for another.

Of course, the process isn't always that simple. Sometimes, the truly hard part is figuring out what those other sizes should be. Other times, the construction method or design details of the garment throw a monkey wrench into the computations.

How do I know what sizes other people are, anyway?

This is the million-dollar question, isn't it? How do you find a set of sizes that will fit all knitters, from extra-extra-extra-small to double-venti?

It's all very well to fit yourself and tailor the number of stitches and rows to your own body as you knit, but the general idea of a knitting pattern is that complete instructions are provided for the knitter to follow to arrive at a finished product. This doesn't mean that the knitter doesn't think while following instructions, but "increase until it's big enough to fit around your chest" isn't a pattern instruction -- it's just the intended result. Even if you do empower the knitter to recognize where adjustments can be made, and how to make them, the knitter still (quite fairly) has some expectations that you'll provide numbers for each pattern size. The knitter also hopes that you'll also cover the size that he or she wants to knit.

However, there is no set of standards that will apply satisfactorily across a broad range of sizes. (If there was one, clothes shopping experiences would probably be more pleasurable.) There are standard sizing guides out there, and they may or may not be the right fit for you and the knitter of your patterns. For example, there are the Craft Yarn Council of America's sizing standards, which are used by some knitting publications; other publishers may have different sizing guides. You can also obtain sizing information in some knitting or sewing reference books, or from various standards organizations (but be prepared to pay for it), or from digging into the data used by knitting software programs (more on that later). You can even do your own detective work and survey ready-made clothing, sewing patterns, or neighbourhood children.

Standards and guides, though, are based on statistical averages and assumptions: for example, standards often presume that a woman has a B cup, and that on average, there are height differences between sizes (a woman with a 32 inch bust is presumed to be shorter than a woman with a 40 inch bust). In the aggregate, these presumptions are fairly made, but it means that no sizing methodology will ever truly account for the ways of all flesh and where it's located. A full bust measurement of 36 inches could belong to a small-boned woman with a C or D cup, or to a woman who perfectly matches the measurements in a published standard -- yet, these women would probably have different cross-shoulder measurements, sleeve lengths, back waist lengths, and hip and waist measurements. If you conduct your own research to develop your own size guidelines, your data will suffer from the same flaws of your source. If the clothing lines you inspect tend to be designed for skinnier bodies, or if the pattern line is reputed to fit too tightly in the shoulder area, you may wind up importing those problems into your patterns. Even guides published in reference books will become outdated with time, as average adult height and weight increases over decades. On top of that, some of these standards or guides may lack some dimensions that you need for your grading work. Some guides are complete; for others, you may have to fill in some details yourself.

Despite all of these issues with standards and size guides, you will still need to have a starting point to assist you in grading your pattern up and down, and that starting point will probably be somebody else's sizing charts. At a minimum, for grading a woman's pullover, you'll want to find a table that correlates at least these measurements:

• full bust circumference (assuming that the body shape reflected in these measurements has a B cup)

• cross-shoulder (shoulder point to shoulder point)

• wingspan (from wrist to wrist, arms slightly bent, across the shoulders)

• back waist length (from back neck to waistline)

• waist circumference

• hip circumference

• hip depth (from waistline to hipline)

Some tables will express wingspan in a different way: they may give a half-wingspan value (from center back, over the shoulder, then down the slightly bent arm to the wrist), or they may give a sleeve length (from shoulder point to wrist) plus a cross-shoulder measurement. In theory, twice the sleeve length plus the cross-back measurement should equal the total wingspan. If you find a data set that doesn't quite add up this way, then either the sleeve length was made longer to ensure wrist coverage when the arm is bent sharply, or the shoulder point used to determine the sleeve length is not the same shoulder point used for the cross-shoulder measurement. This can lead to confusion when drafting set-in sleeves.

These are the basic measurements, but if you've ever tried to grade a pattern, you can probably think of additional data that you'd love to have:

• armhole depth (vertical distance between shoulder and armpit)

• raglan depth (diagonal distance between crew neckline and armpit)

• bicep circumference (measured around the fullest part of the arm)

• wrist circumference

• head circumference (although this is obviously useful for hats, it's also useful to know whether a neckline will stretch enough to fit over the wearer's head)

• neck width (not neck circumference, but the horizontal distance)

This data is out there (particularly the last three items), but information about depths and biceps can be harder to come by; one reason is because these dimensions, when they exist, are often expressed in terms of finished garment measurements, which are heavily influenced by the design and wearing ease of the garment, and not in terms of actual body measurements. You should make a note, by the way, whether the dimensions in your sizing standard are finished measurements, or actual body measurements.

Range and depth: selecting the sizes

There are two aspects to selecting a size range for a pattern. The first, which is the most obvious, is choosing the range of sizes: what is the smallest size? What is the largest size?

We won't go into the politics of offering a garment in a range of sizes. It is probably an objective truth to say that not every clothing design will flatter every figure, but what a designer does with that truth is a personal (and business) choice. Some designers may consider that their design integrity is compromised if a pattern is adjusted to fit a body shape for which it was not intended. Other designers, however, figure that they cannot be the judges of what will look good on a body that they have never seen, and realize that the customer knows what he or she wants.

The second aspect is the number of sizes within the range. If you are offering sizes ranging from a 32 inch to a 62 inch full bust measurement, do you need to offer sizing in increments of two inches? Four? Six? The choice of increment could have you crunching numbers for six or sixteen sizes.

When a pattern is published by a third party, the designer (here, when we use the term "designer", we're not only referring to the person who greated the design and perhaps first prototype, but also to the pattern writer who creates the instructions for different sizes -- these roles may be filled by the same or different people) may or may not have any say about the range or number of sizes. A publisher may expect instructions for a specific range and number of sizes, which relieves the designer of the burden of decision-making. But in a self-publishing situation, it's up to the designer to decide how many sizes to offer, and over what range. These decisions may be constrained by space and publishing considerations (there is a cost to having instructions reviewed by technical editors, or having multiple sizes test knit). However, other design considerations may impact the choice of size increment:

Ease: is the garment intended to fit loosely, or snugly? A loose-fitting garment is usually a little more forgiving; with an intended ten inches of ease in the body, it will probably not matter if the wearer has a 40 or a 43 inch chest as much as it would if the garment was intended to be tight-fitting.

Scope for adjustments: is it easy for the knitter to make his or her own adjustments without significantly altering the pattern, and if so, is it possible to provide fewer sizes along with instructions on fine-tuning the fit?

Whether you can "cheat" this way will depend on the garment style and patterning. Provided some guidance was included in the pattern, it would probably be easy for a knitter with a small amount of experience to alter the size of a raglan pullover by an inch or two. But it wouldn't be so easy if the pullover involved an allover stranded colourwork pattern that would be cut off at the wrong point if the stitch counts were changed. This kind of decision requires some consideration -- the knitter may or may not appreciate being given this much latitude.

These two factors are related. Let's consider an example of a pullover to be offered in sizes to fit bodies measuring 30 inches to 60 inches, full bust measurement.

If the garment was intended to fit rather loosely, like a drop-shoulder sweater, experience might suggest that the pullover would look just as good with six inches of ease as with ten inches of ease. Is it necessary to provide pattern instructions at 2-inch increments? Likely not, given the loose fit, and here is why:

With such a range of tolerable ease for a given size, there will be some overlap between sizes. For a 36 inch full bust measurement and between 6 and 10 inches of ease, the finished size of the garment could be 42 to 46 inches (we could also express this as 44 inches, plus or minus 2 inches). For a 38 inch full bust measurement, the finished size of the garment could be 44 to 48 inches (or 46 inches, plus or minus 2 inches), and for a 40 inch full bust measurement, the finished size of the garment could be 46 to 50 inches (or 48 inches, plus or minus 2 inches).

For a sweater that can tolerate between 6 to 10 inches of ease, you can see that there will be overlap in the range of acceptable finished garment sizes for these three "adjacent" body sizes.

Note that the appropriate finished sizes for the 38 inch bust size are covered by the potential ranges of the 36 inch and the 40 inch sizes, and in fact the smallest finished size for the 40 inch bust measurement is also the largest size for the 36 inch full bust measurement.

If we provided instructions for every 2 inch increment for this design, we would have to provide instructions for a full bust measurement of 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50, 52, 54, 56, 58, and 60 inches: sixteen sizes. But given the overlap between sizes 36 and 38, and 38 and 40, it seems that we could perhaps drop every other size, and still provide the knitter with decent coverage. We would then be left writing up instructions for full bust sizes of 30, 34, 38, 42, 46, 50, 54, 58, and... well, the original intention was to end at 60 inches, so you could either provide the 60 inch size, and respace the sizes between 30 and 60 (or not). Alternatively, you may choose to either end with 58 or 62 instead. At least, if we choose a final size for a 62 inch full bust measurement, we now have only nine sizes to write up instead of sixteen.

Even if the size 38 pattern instructions are omitted, the size 36 or the size 40 instructions will cover the wearer between these two sizes.

The rule that we just followed in this example is to increment the body sizes by an amount equal to the range of acceptable ease. Here, we had 4 inches of leeway in the ease (6 to 10 inches of ease); we could therefore increment the sizes by 4 inches (e.g., for a body measurement of 36 inches, to 40, to 44, to 48, etc.). With this 4-inch increment, an appropriate amount of ease will be available for any wearer with a full bust measurement betwen 30 and 62 inches -- even if the wearer falls between two of the pattern sizes.

When we actually choose the finished dimensions for each size, we would logically choose to add 8 inches of ease -- the midpoint between the 6 and 10 inches of acceptable ease.

But if the garment was intended to have a slinky, tight fit -- say, 0 to 1 inch of wearing ease (every fitting style can admit some flexibility in ease, whether it ranges from -0.5 to +0.5 inches, or 0 to 1 inches) -- then incrementing the sizes by four inches would leave some gaps.

In a garment style with little tolerance for ease variation, providing pattern instructions in large increments leaves a number of sizes uncovered.

However, if we follow the same rule as above -- that we increment the sizes by an amount equal to the range of acceptable ease, which here is 1 inch -- we would have to generate instructions for thirty-one sizes. Even if this would make every knitter happy, that's a crazy number of sizes to work into a single pattern.

If you are determined to provide all available sizes in one pattern, you'll need to make a judgment call: provide all possible sizes, or drop some sizes and leave them as an exercise to the reader? This is where the second consideration, scope for adjustment, comes in; it could be that if the garment design is simple enough to allow knitters to easily alter the garment size by an inch or two, you can omit some sizes and leave it to the knitter to tweak the sizing. If at all possible, though, make this alteration option abundantly clear in the pattern instructions.

Number-crunching: applying theory to practice

Now, you've got your target body measurements for each size. You've figured out the increments for your pattern sizes, and you've also managed to determine the finished garment size (the full bust measurement, at least) for each garment size.

Now for more arithmetic! After the anguish of locating an appropriate sizing standard and determining the range of sizes to provide, this part should seem easy.

Figure out the key finished measurements of the original garment, and the ease involved for each of these measurements: You can likely characterize your prototype pattern as a set of key finished measurements -- the circumference at the hem, waist, and bust; the width at the shoulders; the depth of the armhole and the front and back neckline, and so forth. Based on those finished measurements, you will have also worked out the amount of wearing and design ease. Write these things down:

Some possible key finished measurements and ease amounts for the original pattern.

One important point to consider is whether your prototype fits within the sizing guide that you intend to follow. If the prototype was made for you, for example, and you are unusually long-waisted or are very small-boned, it may be that you will need to tweak your prototype measurements to fit your sizing standard. Your original measurements, though, will provide a guide as to how much ease to add to all the different measurements: ease at the bust, around the wrist, across the shoulders, and so on. These amounts of ease will likely remain fairly constant across all sizes; you might, with your own personal fitting philosophy, choose to vary the amount of ease for some size ranges, but the original amount of ease for each measurement of your prototype will be your guide.

Finished lengths may be treated differently. The neckline depth at front and back will likely be fairly constant across sizes, although as sizes increase you may wish to increase the front depth of a crew neck to ensure there is ample room for the neck, and of a V-neck or other decorative neckline to maintain design proportions. The shoulder height (the vertical distance created during shoulder shaping) will likely also be fairly constant; while some people have shoulders that slope more or less than the average (whatever average is), you can't really anticipate this in a knitting pattern.

Some length measurements require a little more planning. Sizing standards will tell you that the back waist length will increase with size; the overall length of a garment, therefore, will theoretically increase by the amount that the back waist length increases. However, the design of the garment (if there is a specific vertical pattern repeat that is used) may prevent you from providing instructions in such increments; you might need to fix a single length for every size, so that your Fair Isle pattern or your cable panels begin and end at the right place. Also, looking at your schematic, you will realize that there are two places where length can be added to the garment: below and above the underarm. As you increase in size, you may find that you will be increasing the armscye depth, so you may not need to make any adjustments below the underarm. Similar logic may apply to the sleeve cap depth and the sleeve length between wrist and bicep.

Determine what these key finished measurements will be for each size of the pattern: Perhaps the most obvious way to work this out is to construct a table for yourself -- either by hand or using a spreadsheet program, if you are so inclined -- so that you can fill in the key finished measurements for each of your intended sizes. If you are more of a visual person, you might consider sketching a pattern schematic for each size, and writing numbers on the schematic.

Below is an example of a table for a relatively simple set-in sleeve pullover, knit flat with a separate front and back, with no waist shaping and four inches of ease at the chest (remember, that's two inches for each of the front and back). Each column corresponds to a size of the pattern, and each row corresponds to a finished measurement of the garment. The one column of numbers that is filled in corresponds to the one set of numbers you know at this time: your prototype. For ease of reference, the table also includes the amount of planned ease:

Some of the "ease" values, particularly length measurements (which aren't really "ease") might be left blank, depending on how things need to be adjusted based on finished lengths; also, for some values you might not actually know the amount of ease, but only the finished dimension -- for example, if your armscye depth might be drawn from a sizing guide that provides finished measurements, and not the underlying body dimensions. For each size column, the empty cells will need to be filled in with a number, which will be derived in part from the ease that you intend to insert for each measurement, and from your sizing guide.

It may be that your prototype measurements and your pattern sizes fall between the sizes in the standard you're following; for example, your standard may give dimensions for full bust measurements of 36, 40, and 44 inches, whereas you are intending to provide pattern sizes for 38, 42, and 46 inches. In that case, you will need to interpolate or extrapolate from the numbers you have in order to fill in this table: for example, if your standard said that a wrist measurement for a size 36 is 7 inches, and for a size 40 is 7.25 inches, it would be fair to conclude that the wrist measurement for a size 38 is 7.125 inches.

You will therefore be able to gradually fill in these numbers (in this example, the numbers are fictional):

Compute the stitch and row counts for each of these measurements: This is a matter of converting your numbers to stitches or rows, based on your gauge. In some instances, you may need to round up or down to the nearest even or odd number in order to maintain symmetry in your design; if you need to have a central stitch on each body piece, for example, then you will need to make sure that all of your stitch counts for the body are odd. Here, our fictional design has a gauge of 5 stitchs and 7 rows per inch, and it doesn't matter whether the sleeves have an even or odd number of stitches, but it does matter that we have an even number of stitches on each of the front and back:

You'll note that because of the restriction that the front and backs each have an even number of stitches, we've had to make a change: 5 stitches per inch, times 19 inches, equals an odd number (95). Here, we've increased the stitch count by one. Given that the pieces of this garment are knit flat, a seam allowance will be needed anyway. Even after taking a conventional one-stitch allowance for seaming, the final width will be very close to the intended finished width.

You'll also note that other values, like the cross shoulder width, no longer increase with each size -- an effect of rounding off to a convenient stitch count. Here, values were rounded down to the nearest even number to try to ensure that the seamline between the body and sleeve stays as close to the shoulder point as possible; however, you might choose to round up, particularly if the garment was meant to have an easy fit.

Also, lengths such as the armscye depth or neckline depth may be rounded off to a convenient even number, even if the accurate value is an odd number of rows. Although it's not necessary to do this, often knitting patterns provide lengths in terms of even numbers of rows, so that the next stage of shaping always begins on a right-side row. (This presumes, of course, that the work began on a right-side row, too.) This is not mandatory -- it's only a convention, of sorts, and is not as conventional as other "rules", such as the direction in which a chart is read.

Finally, note that even though it might not matter whether stitch counts for the sleeve are odd or even, they need to be consistently odd or even for each size(assuming your sleeve is symmetric, and not tailored closely to actual body shape -- which isn't symmetric). This is an adjustment you might make at this stage, or during the "tweaking" stage when you insert the numbers into your written instructions. (Here, we've already made that adjustment.)

Insert these numbers into your written instructions, tweaking as necessary: Your pattern instructions are the instructions that get you from one of these finished measurements to another. For example, if your prototype notes told you (for the 42 inch size) to increase from 45 stitches at the wrist to 91 stitches at the bicep over 140 rows, then you might have written these instructions as:>

Increase 1 stitch each side every 6 rows 23 times; work 2 rows even.

You would then figure out something similar for all of the other sizes.

For the 34 inch (body measurement) size, you need to increase from 43 stitches to 79 stitches over 134 rows:

Increase 1 stitch each side every 7 rows 18 times; work 8 rows even.

Another convention is to space increases so that they always fall on the right side of the work -- in other words, to specify that increases are worked every 6 or 8 rows, but not every seven. You may choose to follow this convention if you wish. One example:

Increase 1 stitch each side every 6 rows 9 times, then every 8 rows 9 times; work 8 rows even.

For the 38 inch size:

Increase 1 stitch each side every 6 rows 20 times; work 16 rows even.

For the 46 inch size:

Increase 1 stitch each side every 5 rows 25 times; work 19 rows even.

Alternatively, because it may not be desirable to have 19 rows worked even at the end, but to have them spaced out a little more:

Increase 1 stitch each side every 5 rows 15 times, then every 6 rows 10 times; work 9 rows even.

Having figured this out, you'll then need to merge all of these instructions -- how confusing you wish to make your pattern instructions is up to you. One example, using the more complex instructions for the 34 and 46 inch sizes:

Increase 1 stitch each side every (6, 6, 6, 5) rows (9, 20, 23, 15) times, then every (8, 0, 0, 6) rows every (9, 0, 0, 10) times; work (8, 16, 2, 9) rows even.

The exact format will depend on the formatting standards dictated by the publisher, if you are not publishing the pattern yourself.

This is a simple example; you can imagine that when you work out the shaping instructions for more complex shapes, like sleeve caps, armscye shaping, and necklines, that compressing the instructions into concise statements may be more of a challenge. Clear pattern writing takes some skill!

As you work out these instructions, the need to tweak the numbers you had computed will arise. One of the tweaks you will need to consider is a seam allowance: do you need to add extra stitches for seams? If you have a neck finish or a ribbing that requires a specific multiple of stitches, will the numbers you've computed work, or do you need to change the numbers or sneak in some increases or decreases? Changes like these make pattern writing a somewhat iterative process; although you will have worked out some numbers in the previous step, those numbers might have to change, and you may need to recompute some other measurements to confirm that your instructions will still yield pieces of the correct dimension.

It may also be that the garment design imposes restrictions on the possible sizes: think of an Aran sweater composed of specific cable panels that are all three inches wide, and that need to be centered on the front and back -- while the obvious answer to multisizing is to increase each size by 12 inches (two panels each on the front and back -- that's rather a lot!), other creative solutions are available to cover intermediate sizes.

Next time: Dealing with restrictive design considerations; estimating yardage amounts; and using knitting pattern software.

References

Find Knitty's sizing guides here, based on the CYCA guidelines, and a link to Ysolda Teague's expanded guide that provides what the CYCA information is missing.

You can also consider references like Righetti's Sweater Design in Plain English (St Martin's Griffin, 1990), which includes reference tables for men, women and children.

Other measurement sources include standards bodies, such as ASTM International. Reference D5585, for example, is a standard table of body measurements for adult female misses, sizes 2-20 (32 to 44.5 inch full bust measurement); D6960 covers sizes women's plus sizes 14W to 32W (39.5 to 57.5 inch full bust measurement). ASTM data includes values such as the armscye girth (around the armscye, front and back) as well as armscye ("scye") depth), and upper arm girth.

However, as cautioned above, all data has its flaws, and despite the fact that the size guide you use is called a "standard", you may find that the data is not representative of the knitters of your patterns.