It is just division, and then the remainder

There is only one thing to do: divide the stitch count by the number of shaping steps, then make the interval one stitch wider for as many times as the remainder. That is what "spread them evenly" actually means.

If it divides exactly, every interval is the same. If it does not, you mix in a slightly wider interval for the leftover. That is the whole trick.

Worked example: 114 stitches down to 88

You are decreasing 114 − 88 = 26 times. The number you divide is the smaller one (88).

Here is the catch: charts and knitting books write this one number higher, because they count the row you are holding right now — the 114 stitches before you work the decreases. For decreases every interval gains +1.

As it appears in a chart

Real charts often factor out a common divisor and write it as "every 4 sts × 8 and every 5 sts × 5, repeated twice". It is the same instruction, just easier to count while your hands are busy.

That check is the most reliable test there is. If interval × count adds up to the stitch count before the row, the plan is sound.

Why tutorials disagree with each other

This is the part that makes searching for an answer confusing. For the very same "114 → 88", some tutorials say every 3 sts × 16 and every 4 sts × 10, and others say every 4 sts × 16 and every 5 sts × 10.

Both are right. Only the counting convention differs.

Counting conventionRow it countsHow 114 → 88 is written
Count the row before working it (charts and knitting books)114 stsevery 4 sts × 16 / every 5 sts × 10
Count the row after working it88 stsevery 3 sts × 16 / every 4 sts × 10

The knitting is identical; the notation is offset by one. You can tell which convention a tutorial uses by adding it up: a total of 114 means the before-row convention, a total of 88 means the after-row one.

When I built the app, my first implementation divided the larger number. I simulated every pattern and the stitch counts came out exact — but the output was one off from the reference implementation I compared against. A simulation that passes is not proof that you are right about the convention.

Increases do not take the +1

Awkwardly, increases and decreases behave differently. Treating an increase as "a decrease in reverse" will give you the wrong number.

The principle is the same — count the row you are holding — but for increases the row you are holding is already the smaller number. So no +1 is needed.

Worked example: 110 stitches up to 130

For decreases the row you hold is the larger count; for increases it is the smaller one. Both are consistent in one respect: the total always equals the stitch count before you work the row.

Do not put shaping at the very edge

Even with the arithmetic right, placing the shaping from the very first stitch hurts the finish. The outermost stitches become the seam allowance, so shaping that lands there shows once the pieces are sewn together.

In practice you leave a few plain stitches at each end and push the shaping inward. The calculation gives you "how many stitches apart, how many times" — the placement convention is not something a calculator can tell you.

From gauge to stitches and rows

The stitch count itself comes from your swatch.

With a gauge of 20 stitches and 28 rows over 10 cm, a piece 50 cm wide and 60 cm long is 100 stitches × 168 rows. Only once you have that can you work out armholes, shoulder slope and the rest.

Yarn quantity scales from the swatch too: yarn needed (g) = area to knit ÷ swatch area × swatch weight. In practice, buy a little more than the number says.

A free app that does this arithmetic

None of the steps are hard. The real burden is that a single garment needs them dozens of times. To hand that to a machine I built a free app called Amidoria (Knitting Calculator) — it is made here at GRAMSHIFT.

Summary

FAQ

What is even distribution in knitting?

It is the method for spreading increases or decreases evenly across a row. Divide the smaller stitch count by the number of shaping steps to get a quotient and a remainder, then make the interval one stitch wider for as many times as the remainder. The rest use the plain quotient as the interval.

Why do tutorials give different numbers for the same shaping?

Because they count different rows. Charts and knitting books count the row you are holding before you work it, while some tutorials count the row after it is worked. For decreases this makes every interval differ by one. Both are correct; you can tell which convention is in use by adding interval times count and seeing whether the total matches the stitch count before or after the row.

Can I put increases and decreases at the edge of the piece?

It is better not to. The outermost stitches become the seam allowance, so shaping placed there shows after the pieces are sewn together. Leave a few plain stitches at each end and move the shaping inward. The calculation gives you the interval and the number of repeats; the placement is a finishing decision.

Also in Japanese

This guide is also available in Japanese: 編み物の平均計算(割り出し)のやり方(日本語版).