Methodology

Every number on this site comes from one of two places: a published industry standard we cite directly, or StitchHow's own calibration, which we show you the working for below. We never invent a number silently. Where we could confirm a figure from two or more independent sources, we say so and name them; where we could only confirm one, or none, we say that too, plainly, rather than presenting a thin or missing source with false confidence. See the crochet calibration table below for a real example of both.

Where the standards come from

Two categories of data on this site are third-party published standards, not our own invention:

• Yarn-weight gauge ranges (lace through jumbo) follow the Craft Yarn Council's Standard Yarn Weight System, the same category numbers (0–7) printed on most commercial yarn labels.
• Needle and hook size conversions (US to metric to UK/Canada old sizing) follow the widely published industry-standard equivalence chart used across pattern publishers.

Those numbers are reproduced, not calibrated — see the Yarn Weights and Needle & Hook Size pages for the tables themselves.

The yarn-quantity model: what StitchHow calibrated ourselves

The Yarn Quantity Estimator's gauge-based mode uses a model we built and calibrated in-house, because no single public source gives a stitch-by-stitch yarn-consumption constant for every stitch pattern. Here is the whole model, in the open.

The formula

For a piece of fabric with a target area, at a known gauge, in a known stitch pattern:

• Stitch width w = 1 / stitches-per-inch
• Stitch height h = 1 / rows-per-inch
• Yarn used per stitch L = K_pattern × (w + h) inches, where K_pattern is a constant specific to the stitch structure
• Total yards = area (sq in) × stitches-per-inch × rows-per-inch × L / 36

K_pattern is not a universal constant — it has to account for how much a stitch structure "takes up" yarn beyond the flat linear width and height of the stitch. A rib pulls in and eats more yarn per stitch than plain stockinette at the identical gauge; single crochet is denser than either. That's what the constant captures.

How we calibrated K_pattern

We anchored the model on stockinette stitch, using a controlled reference point: published gauge-swatch guidance reports that a 6-inch square stockinette swatch (36 square inches), worked at a typical worsted-weight gauge of roughly 4.5–4.75 stitches per inch, consumes about 29–32 yards of yarn. Solving the formula above for K_pattern at the midpoint of that range gives a value of roughly 2.8.

We cross-checked that figure at a larger scale using published yardage guidance for a typical DK-weight adult hat (a ribbed brim plus a stockinette-dominant body and crown): commonly reported as 200–300 yards. At typical DK gauge and an estimated ~200 square inches of knitted fabric, the implied constant comes out closer to 3.5.

Averaging the two calibration points and rounding to a clean, defensible number, we set the stockinette anchor at K = 3.0. Every other stitch pattern's constant is stated as an explicit multiple of that anchor, not derived independently, so the relationships between patterns stay consistent even though the absolute number is our own calibration.

The full derivation table

Stitch pattern	Method	Multiple of stockinette	Resulting K	Why
Stockinette	Calibration anchor	1.00x	3.00	See above.
Garter	Direct multiple	1.18x	3.55	Published guidance for textured stitch patterns commonly recommends roughly 15–20% more yarn than stockinette. Re-verified against a real worsted-weight garter baby-blanket estimate landing in its published band.
1x1 Rib	Direct multiple	1.20x	3.60	Top of the textured-allowance range — 1x1 rib draws in the most of any pattern here.
2x2 Rib	Direct multiple	1.15x	3.45	Bottom of the textured-allowance range. Re-verified against a real worsted 2x2 rib scarf estimate (~582 yards) landing in its published band.
Seed stitch	Direct multiple	1.17x	3.50	Mid-range textured allowance.
Single crochet	Direct project calibration	1.46x	4.37	Solved directly from a real published pattern — see below.
Half double crochet	Interpolation	1.55x	4.66	No trustworthy direct anchor was found (see below) — interpolated as the midpoint between single and double crochet's directly-calibrated K values.
Double crochet	Direct project calibration	1.65x	4.94	Solved directly from three real published patterns, three different designers — see below.
Granny square motif	Structural fallback	1.65x	4.94	No sourced figure exists for a granny motif specifically (see below) — uses double crochet's K as a conservative ceiling.

The crochet constants: two rounds of correction, and what's still unresolved

The crochet rows above went through two corrections, and we're describing both rather than quietly presenting only the final number, because the process is exactly what tells you how much to trust it.

Round 1 — a units bug. An earlier version of this table took a published per-area comparison ("single crochet uses 1.60x as much yarn per square inch as stockinette") and multiplied it directly onto K_pattern, which is a per-stitch constant. Those are different units. The correct conversion factor between a per-area ratio and a per-stitch constant is (stitches-per-inch + rows-per-inch), and that factor is roughly half as large for a tall, open stitch like double crochet as it is for stockinette. Applied directly, the bug didn't just shrink the correction — for double crochet it flipped the sign: the site briefly shipped double crochet at K = 3.80, less than stockinette's 3.00, the opposite of what the source it cited actually said.

Round 2 — double-counting. Fixing the units conversion (deriving K from the ratio through a reference gauge) reproduced the cited ratios faithfully, but that fix had its own blind spot: this model already takes your gauge as an input, and (stitches-per-inch + rows-per-inch) already reflects the fact that crochet is worked at a looser density than knitting. Taking a ratio measured by comparing knit and crochet swatches at their own different natural densities, and then applying it on top of a model that already receives a gauge which differs by craft, double-counts that difference.

What we did instead. Single and double crochet's K values here are now solved directly from real, finished, published projects — the exact same method used for the stockinette anchor itself: K = (published yards ÷ 1.1) × 36 ÷ (area in sq in × (stitches-per-inch + rows-per-inch)), using each project's own stated gauge and finished size.

• Single crochet is solved from one project: Elise's Single Crochet Blanket (easycrochet.com), stroller size — 104 sc × 131 rows at 14 sts/15 rows per 4in, 1,006 yards. (Its own stitch and row counts imply 29.7 × 34.9in, matching the pattern's own stated "35 × 30in" almost exactly — a consistency check most other candidate sources we checked could not pass, usually for missing a row gauge, a finished size, or mixing stitches.)
• Double crochet is solved from three independently-designed projects and averaged: a 36×48in blanket (1,300yd, daisycottagedesigns.net), a 5.5×54in scarf (240yd, sweetbeecrochet.com), and a ~30in-square baby blanket (~710yd, daisycottagedesigns.net) — individually implying K of 4.69, 5.29, and 4.92. A fourth candidate (a 40×50in blanket) was excluded: its ~2,100-yard figure mixes a primary color with three unquantified accent-color skeins used for stripes, and the split between "covering fabric" and "sitting partly used in a stripe skein" isn't recoverable — omitted rather than guessed at.
• Half double crochet has no trustworthy direct anchor. One well-documented multi-size pattern was found (the Haylee Afghan, eyeloveknots.com, 15 sts/12 rows per 4in), but solving K independently per size gave 5.00 for its smallest (Newborn) size and ~3.1 for its two largest (Throw, Queen) — consistent with each other, but not with the small size (a small piece's border is a much bigger fraction of its total yardage, which biases K solved from it upward), and a K of ~3.1 would rank half double crochet's yarn use below single crochet's, contradicting the stitch's own structure (hdc sits between single and double crochet in height and wraps, by every source's description). Rather than use an anchor that produces a structurally implausible result, half double crochet's K is an interpolated midpoint of single and double crochet's directly-calibrated values instead.
• Granny square has no sourced figure at all, direct or ratio-based — every source we checked agrees qualitatively that the open chain spaces between clusters use less yarn than solid double crochet, but none gives a number. Rather than invent an open-space discount, the calculator uses double crochet's K as a conservative ceiling and says so on the Granny Square page: a real motif likely uses somewhat less than this estimate.

A real, unresolved disagreement. Two controlled lab studies — Interweave's "Yarn Usage for Knitting vs. Crochet" and an independent 2014 test by Flagel Knitting Files — each wove/crocheted matched swatches in one yarn and weighed them, finding single crochet uses about 1.41–1.60x as much yarn per square inch as stockinette, and double crochet about 1.19–1.27x. Checked at a typical worsted gauge for each stitch, the direct-project K values above imply something different: single crochet at roughly 1.04x stockinette (still more, roughly agreeing on direction), but half double and double crochet at roughly 0.85x and 0.87x — about the same as stockinette, or slightly less, not clearly more. As one concrete illustration: at a fixed 144 sq in, this site's own worked comparison computes stockinette at 4 sts/5.5 rows per inch as 125.4 yards, single crochet at 3.5/4 as 144.21 yards (more), and double crochet at 2.75/1.75 as 97.81 yards (less) — see the Yarn Quantity Estimator for the same math on your own numbers.

We have not resolved this, and we're not going to paper over it with a single averaged number. Our working theory: the lab studies measure yarn density with no incentive to shade the number either way, but at a single matched yarn/hook combination that may not reflect the (much looser) gauge most patterns are actually written at. The direct-project anchors are real finished patterns at real crocheters' actual gauges, but they're independent hobbyist-blog patterns rather than lab-tested ones, and free blog patterns are known to sometimes understate required yardage — a generous-looking number is more shareable than a discouraging one. That bias, if real, points in exactly the direction we're seeing. But it's a theory, not a citation. If you have a better-sourced answer, use the contact page.

Until this is resolved, treat the double and half double crochet estimates from the gauge-based model as the more uncertain figures on this site, and prefer the swatch-weighing mode below — which uses no K_pattern constant at all — for any double or half double crochet project where the total genuinely matters.

The reference projects we checked this against

• 6x6in stockinette gauge swatch, worsted weight — 29–32 yards reported, primary calibration point.
• Adult hat, DK weight — 200–300 yards reported, used as a second-scale cross-check.
• Adult medium sweater, worsted weight — 1,000–1,500 yards reported. Garment area varies too much by construction and ease to solve K precisely from this alone, but it's consistent with the swatch- and hat-derived values, which is what we checked it for.
• Elise's Single Crochet Blanket (easycrochet.com) — 1,006 yards for a 30×35in stroller-size blanket at a stated 14 st/15 row per 4in gauge. Where the single crochet constant comes from directly.
• Three independently-designed double crochet patterns — a 36×48in blanket, a 5.5×54in scarf, and a ~30in-square baby blanket, from two different designers. Where the double crochet constant comes from directly.
• Two controlled knit-vs-crochet density comparisons (Interweave; Flagel Knitting Files) — published side-by-side square-inches-per-gram figures for stockinette, single crochet, half double crochet (Interweave only), and double crochet, at matched yarn and hook/needle size. Kept as a documented cross-check against the direct-project figures above, not the derivation path — see the disagreement noted above.

The honest limits of this model

This is StitchHow's own calibration, built from a small number of real reference points, not a laboratory-verified physical constant. Yarn thickness, fibre content, your own tension, and the exact stitch pattern all shift the real answer. Treat the gauge-model estimate as a planning number, and always buy a little more than the raw total — the calculator's ease/seam/edging allowance is shown as its own explicit percentage on top of the base estimate, never folded in silently.

The accurate path: weigh your swatch

If you want a number that doesn't depend on any constant at all, use the calculator's "I weighed my swatch" mode. Knit or crochet a swatch in your actual yarn, at your actual gauge, in your actual stitch pattern. Weigh it, measure its area, and weigh (or read the label of) your full ball or skein. The tool then scales by simple density — grams per square inch, times your target area, times yards per gram from the ball's own label. No K_pattern, no calibration, no assumptions about your yarn. This is the accurate path, and the tool says so on the page.

Needle and hook conversions

The needle and hook conversion tables are reproduced from the standard published US/metric/UK(old) equivalence chart used across the pattern industry. One detail that trips people up: UK (old) sizing runs inverse to US and metric — the UK number gets smaller as the needle gets bigger. A US 0 needle (2.0mm, one of the finest listed) is UK 14; a US 11 needle (8.0mm, one of the largest) is UK 0. Below UK 0, the convention switches to 00 and 000 for the largest traditional sizes, which are bigger needles despite having more zeros. Our conversion tool encodes this relationship directly rather than treating it as a simple lookup, and our own test suite specifically asserts the inverse direction never breaks.

Cast-on rounding

The Cast-On Calculator doesn't just round your raw stitch count to the nearest whole number — it rounds to the nearest multiple that actually matches your stitch pattern's repeat, so a 1x1 rib project always gets an even number and a k3,p3 rib project always gets a multiple of 6. We round to whichever valid multiple is numerically closer to your raw gauge math, not always up and not always down, and we show you both the raw number and the adjustment so you can see exactly what changed and why.

Questions or corrections

If you find a figure on this site that looks wrong, or you have a better-sourced number for something we've marked as our own calibration, use the contact page — we'd genuinely like to fix it.