Crochet Color Repeats & Stripe Planning Guide

Striping is two decisions that get muddled together: what order the colours go in, and how much of each you need to buy. They are not the same problem and they are not solved at the same time. This is about the first one — the shape of the repeat, how it joins itself, and how to decide how many rows each colour gets — and about handing the second one off cleanly once the first is settled.

How crochet stripe repeats work

A stripe pattern is a short list of colours with a row count against each, worked over and over. Four rows of cream, two of terracotta, one of sage, then back to cream: that is a 7-row repeat unit, and everything else about the pattern follows from it.

Two things are worth being precise about from the start. The unit is measured in rows, not in colours — a three-colour stripe can be a 7-row unit or a thirty-row one. And the unit repeats end to end, which means its last row and its first row become neighbours in the fabric. That second point causes more trouble than anything else on this page, and it has a section of its own further down.

The other thing that follows from the unit is where the pattern stops. A project is however many rows it is, and the repeat divides into that however it divides. A 7-row unit over 100 rows gives 14 whole repeats and 2 rows left over — so the piece ends part-way through a repeat, and the top does not match the bottom.

That is not a mistake, but it is a decision, and it is much easier to make now than at the end. Working 5 more rows finishes the unit at 105 rows; stopping 2 earlier finishes it at 98. Either is fine. Discovering the problem with two rows of yarn left is not.

On the left, one repeat unit of a stripe sequence drawn row by row: 4 rows of A cream, 2 of B terracotta and 1 row of C sage, 7 rows in the unit, bracketed and labelled as the repeat unit. On the right, a 100-row project drawn as compressed rows, showing 14 whole repeats followed by 2 leftover rows outlined in a contrasting colour. Working 5 more rows would finish the unit at 105 rows, and stopping 2 rows earlier would finish it at 98. Neither is wrong, but the top will not match the bottom unless one is chosen deliberately.
The unit on the left, the project on the right. Almost no project ends on a whole repeat by accident.

Choose colours before choosing the sequence

The sequence cannot be designed until the palette is settled, because almost every decision in it depends on how many colours there are and how they relate.

How many colours you have sets the shortest possible unit and the number of orders available. Which colour is the quiet one decides what can safely take the most rows. Whether two of them are close in value decides whether they can sit next to each other at all — two mid-tones adjacent in a sequence read as one thick stripe with a vague edge, no matter how different their hues.

Choosing a palette is its own craft and is covered properly in the colour planning guide, which is written around granny squares but whose advice about value, contrast and the quiet colour applies to stripes unchanged. Do that part first.

One thing does belong here rather than there: for stripes specifically, the order matters as much as the selection. The same four colours arranged cream, sage, terracotta, ochre and arranged cream, terracotta, sage, ochre are different fabrics, because what any stripe is read against is the stripe beside it.

Simple repeating stripes

The simplest sequence gives every colour the same number of rows: four cream, four terracotta, four sage, repeat. A three-colour, twelve-row unit, equal thirds.

It is worth knowing what it does well. Equal stripes read as a rhythm rather than a subject — no colour leads, and the eye moves along the piece rather than stopping. They are also the easiest to work, because there is one row count to remember, and the easiest to correct, since losing your place costs at most one stripe.

The width of the stripe changes the character far more than most people expect. One-row stripes in three colours read almost as a texture or a tweed at any distance. Eight-row stripes read as blocks. The colours have not changed at all; only the row count has.

Equal stripes have one structural advantage that comes up again below: with every colour the same, the join between repeats is never a problem, because the last colour and the first are always different.

Uneven stripe repeats

Giving colours different row counts is where stripe patterns get their character. A 4 / 2 / 1 unit — four cream, two terracotta, one sage — reads as a cream fabric with terracotta bands and a sage pinstripe, rather than as three equal colours.

Two practical points. The single-row stripe is the most useful tool in uneven striping and the most easily lost: a one-row stripe in a colour close in value to its neighbours simply disappears. If a colour is only getting one row, it should be the one that contrasts most.

And uneven units get long quickly. Four, two and one is 7 rows; four, three and two is nine; five, three and two is ten. A longer unit means fewer whole repeats in a given project, which means the leftover at the end matters more, and it means the pattern takes longer to become recognisable as a pattern.

Uneven sequences are also where the join trap lives, because the first and last colours are no longer guaranteed to differ.

Symmetric stripe sequences

A symmetric sequence reads the same forwards and backwards: cream, terracotta, sage, terracotta, cream. It is a natural thing to want, and it contains a trap that is obvious once seen and invisible before.

A palindrome is symmetric as a unit and not as a fabric. Repeat A B C B A end to end and the last A meets the first A of the next repeat. Every A band in the piece is one row deep except at the joins, where it is two. The sequence you designed is symmetric; the fabric it makes is not.

Repeat that five-row unit three times and you get 2 doubled bands — one at every join. It is not a rounding-level flaw; it is a visibly different stripe appearing at regular intervals.

The fix is exact and slightly annoying: drop the last row of the palindrome and repeat that. A B C B, worked over and over, gives A B C B A B C B — which reads as symmetric in the fabric and produces 0 doubled bands. The cost is that the unit on its own is no longer a palindrome, which offends people who like their pattern notes tidy.

One caveat worth stating precisely: dropping the last row removes the doubling at the join. If your sequence already repeats a colour inside itself — A B C C B A, say — those internal doubles stay, because they were deliberate.

The other approach is to not repeat it at all. A symmetric sequence worked once, centred on the middle of the piece, is a mirrored design rather than a stripe pattern, and it has no joins to go wrong. That is a different thing to make, and often the thing people actually wanted.

Random stripe sequences

“Random” stripes are popular for scrap projects, and the word hides an arithmetic fact that catches nearly everyone.

If you pick each row’s colour independently — pulling from a bag, rolling a die, letting a generator choose — then any two neighbouring rows have a one-in-k chance of matching, where k is the number of colours. Over 100 rows there are 99 neighbouring pairs. So the expected number of places where the same colour lands twice running is exactly 99 divided by k.

In three colours that is 33 doubled bands per hundred rows. In two colours it is 49.5 — half the piece.

Doubled bands expected in 100 rows when each row’s colour is picked independently. This is the exact average, not a simulation — there are 99 neighbouring pairs and each has a one-in-k chance of matching.
ColoursChance a pair matchesDoubled bands expected in 100 rows
21 in 249.5
31 in 333.0
41 in 424.8
61 in 616.5
81 in 812.4

Those doubled bands are not errors, but they are not random-looking either: they read as double-height stripes you did not choose, and they change the apparent palette by making some colours look heavier than others. If that is not what you want, you have to exclude it deliberately — pick again whenever the colour matches the previous row.

Two other things are worth doing with a random sequence. Write it down before working it, because a sequence you cannot reproduce is a sequence you cannot fix. And if a generator produces it, use one that takes a seed, so the same sequence can be recovered later.

Three columns of crochet rows. The first shows the symmetric sequence A B C B A repeated three times, with the rows where two A bands meet at each join outlined; there are 2 such doubled bands. The second shows the same sequence with its final row dropped, A B C B, repeated, which produces no doubled bands while still reading as symmetric. The third shows fifteen rows with each row's colour picked independently at random from three colours, with repeated rows outlined; 5 bands double. Every band carries its colour's letter and its own fill pattern. A palindrome is symmetric as a unit but not as a fabric, because repeating it makes its first and last rows neighbours. Picking rows at random from three colours gives about 33 doubled bands in a hundred rows on average, which is arithmetic rather than bad luck.
The palindrome doubles at every join; dropping one row fixes it exactly. Random stripes double about a third of the time in three colours.

How many rows should each colour have?

There is no correct answer to this and anyone offering a universal ratio is inventing one. What there is, is a set of consequences worth knowing before choosing.

Row count is proportion. A colour’s share of the rows is its share of the piece — and, on constant-width fabric, its share of the yarn. Give a colour half the rows and you are buying half your yarn in it. That is the link to the next section, and it is the reason this decision has a cost attached.

Row count is emphasis. The colour with the most rows becomes the background and the others become marks on it. A colour with one row per repeat is punctuation.

Row count sets the unit length, which sets how many whole repeats fit and how much is left over. Small numbers give short units that repeat often and are forgiving about where you stop.

A reasonable way to decide is backwards: settle which colour is the background, give it the most rows, then give the others whatever makes the unit a convenient length. Convenient usually means short, and short means the pattern shows up in a small piece.

Watch for one thing only: a colour appearing so rarely that it reads as a mistake. One row of a strong colour every forty rows does not look like a design decision, it looks like you ran out.

How to estimate yarn needed by colour

Once the sequence is settled, the quantity question becomes arithmetic, and it is a different job from everything above.

The principle is one line: a colour’s share of the rows is its share of the yarn, as long as every row is the same length. Count the rows each colour gets, divide by the total rows, and apply that fraction to the project’s total yarn.

For the 4 / 2 / 1 unit over 20 repeats: 80, 40 and 20 rows, so 57.1, 28.6 and 14.3 per cent of the yarn.

Two things this does not tell you, and both matter at the shop. It does not tell you the total — that comes from your pattern or from a yardage estimate. And it does not tell you how many skeins, because each colour has to round up to a whole skein on its own, which costs more than rounding the project as a whole.

The stripe yarn calculator does both of those, and this guide deliberately stops here rather than repeating its arithmetic.

Three boxes connected left to right by arrows. The first, sequence planning, lists which colours in what order, how many rows each, and how the repeat joins itself, labelled as belonging to this guide. The second, row proportions, gives 80 rows of A, 40 of B and 20 of C, labelled arithmetic. The third, quantity planning, gives 686, 343 and 171 yards, labelled as belonging to the yarn calculator. Only the middle step is pure arithmetic. It holds while every row is the same length; on a shaped piece the share of rows and the share of yarn are different numbers. The sequence is decided first and the quantities follow from it.
Sequence first, proportions second, quantities third. Only the middle step is arithmetic.

Why shaped projects need different maths

Everything in the previous section rests on one assumption: that a row of one colour uses as much yarn as a row of another. On a scarf or a blanket that is true. On anything shaped it is not.

A row across the hem of a jumper and a row across the shoulder are both one row, but one may be twice the stitches. A triangular shawl is worse again — the first row and the last can differ by a factor of ten. Counting rows there does not estimate yarn, it estimates rows, and the two have come apart.

The practical consequence is that on shaped work a colour’s share of the rows and its share of the yarn are simply different numbers, and using the first as the second can be wrong by enough to leave you short.

The fix is to weight rows by their length — count rows times average stitches rather than rows alone — which the stripe yarn calculator supports. Where one colour spans both the wide and narrow parts of a piece, the honest approach is to split the project into sections, work each out separately, and add the results.

This does not change any of the sequence planning above. The order of the colours, the join, the leftover at the end — all of that is about rows and is unaffected by shaping. It is only the conversion into yarn that breaks.

How to plan stripe changes

The join between one colour and the next is a physical thing, not just a notation, and a few decisions about it are worth making before the first change rather than the fortieth.

Change colour in the last step of the previous row. Working the final yarn-over of the last stitch in the new colour makes the change happen at the row boundary rather than one stitch into it. Otherwise the first stitch of each new stripe carries a small hook of the old colour.

Decide whether to cut or carry. Carrying a colour up the side works for stripes of a few rows and saves an enormous number of ends; over more than about four rows the carried strand gets long and loose enough to catch. Cutting is tidier and costs two ends per change. On a 7-row unit repeated twenty times, that is a real number of ends either way, and it is worth counting before choosing.

Count the ends. Every colour change is two ends to weave in unless you carry. A piece with a one-row stripe every repeat has as many changes as it has repeats, doubled. This is the hidden cost of narrow stripes and the reason a beautiful sequence sometimes never gets finished.

Keep changes on the same edge where you can, by choosing a unit with an even number of rows if you are working flat. An odd-numbered unit puts alternate changes on opposite sides, which looks fine but makes carrying impossible.

Everything in this section is practitioner convention, not a standard. No body publishes a rule about when to carry rather than cut, and the four-row figure above is a common habit rather than a measurement — it depends on the yarn, the stitch and how tightly you carry. Treat them as starting habits to test against your own work, in the same way as everything else on this page.

None of this affects the arithmetic. It affects whether the arithmetic gets built.

Worked examples

A simple repeat. Four cream, two terracotta, one sage: a 7-row unit. Over 100 rows that is 14 whole repeats and 2 rows over; stopping at 98 or working to 105 makes the ends match. Cream takes 80 of the 140 rows across 20 repeats, so 57.1 per cent of the yarn.

A symmetric one. Cream, terracotta, sage, terracotta, cream. Repeated three times it produces 2 doubled cream bands, one at each join. Dropped to cream, terracotta, sage, terracotta and repeated, it produces 0 — and still reads symmetric in the fabric.

A random one. Fifteen rows pulled at random from three colours came out with 5 doubled bands, against an expected 4.7 for that length. Over a full 100-row piece the expectation is 33. If that is unwanted, the rule is to re-draw whenever the colour matches the row below.

In all three the sequence was settled first and the quantities came after. That order is the whole argument of this page.

Use the stripe pattern generator

Once you know what shape of sequence you want, the stripe pattern generator builds and previews one for you, which is quicker than sketching rows on paper and much quicker than finding out in yarn.

What this guide adds is the judgement to bring to it: how long to make the unit, whether the join doubles, whether a random draw needs constraining, and what the row counts will cost you in yarn. The generator makes sequences; deciding which sequence you want is the part above.

Calculate yarn needed for each colour

With the sequence fixed, the stripe yarn calculator turns it into quantities: each colour’s share, its length in yards and metres, and the whole skeins to buy.

It needs your project total as an input — it divides that total rather than estimating it — and it has a weighted mode for the shaped projects described above. It will also show you something worth seeing before you commit to a palette: how many more skeins a multi-colour project costs than a single-colour one of the same size, because every colour rounds up on its own.