Why does a tiny rule, repeated over and over, make a giant detailed pattern?
After you watchWhy does a tiny rule, repeated over and over, make a giant detailed pattern?
The short answer
A giant detailed pattern doesn't need giant detailed instructions. One tiny rule — 'replace every straight line with a smaller bent copy of itself' — applied to a triangle again and again grows bumps on bumps on bumps, building a snowflake-edged shape. That's a fractal: endless complexity hiding inside one repeated rule.
Try this next
- What if the rule tented the middle HALF of each line instead of the middle third? Picture a fatter, taller tent on every line. Predict first: will the edge get bumpy faster or slower, and will the bumps poke out further past the box? Then watch how a bigger bite changes the shape each round.
- What if you only ever did the rule once, and never repeated it? Set the repeat count to 1 and predict the shape. Then drag it up one round at a time and watch where the new detail appears — it's only ever added to the lines made in the round before.
- What if you started from a square instead of a triangle? Predict whether 'tent the middle third, then repeat' still makes a self-similar bumpy edge. Then look for the same rule on a fern frond or Romanesco and check if a different starting shape still ends up looking 'the same when you zoom in.'
Now you — bend it
- What if Change the rule so every straight line becomes a tent on the middle HALF instead of the middle third. Predict whether the edge length grows faster than the original 4/3-per-round.Count the new edges and their lengths per round. A bigger bitten-out middle means each line still splits into four, but the tent sides are longer, so the multiplier is bigger than 4/3 — the length runs away even faster.
- What if Invent a rule that splits each line into MORE than four pieces per round. Predict how that changes how quickly the edge length climbs.Length multiplier each round is (number of new pieces) x (length of each piece as a fraction of the old line). Pick numbers that make that product bigger than 1 and the edge diverges; smaller than 1 and it shrinks away.
- What if Predict whether ANY rule that only adds tiny bumps just outside the old edge can still fit inside a small box.Edge length and enclosed area are different. As long as each new bump pokes out only a little and gets tinier every round, the area you add settles to a finite total even while the edge length never stops growing.
Can you prove it?The snowflake's edge length grows by exactly 4/3 every round, so it never stops getting longer. — Take one round. Each straight edge is replaced by 4 edges, each 1/3 as long. So new total length = 4 x (1/3) x old length = 4/3 x old length. Drag the repeat count up one at a time and check the length meter multiplies by about 1.33 each step — it keeps climbing with no top.
Design your own test:Before you drag it, write down: will a smaller tent fraction make the edge length grow slower or faster, and will the final snowflake look pointier or smoother?
Explain it to a 6-year-old: Draw a triangle, then put a tiny bump in the middle of every line, then do it again and again — one little rule makes a whole crinkly snowflake.
The whole story
How it works
Start with a plain triangle. The rule says: take every straight edge, and replace its middle third with a little triangular tent, turning one straight line into four shorter bent lines. Now apply the exact same rule to all of those new lines, and then again to all of those, and again. Each round of the rule makes four times as many edges, each one a third as long, so the total edge length grows by 4/3 every round and climbs without limit — yet the whole shape still fits inside a small box. Because every round does the same thing at a smaller scale, every bump looks like the whole snowflake shrunk down: the final intricate pattern is fully predictable from the one tiny rule plus 'do it again.'
What people get wrong
People think a complicated, detailed pattern must come from a complicated, detailed instruction. But the opposite can be true: one short rule repeated forever can produce endless detail. The Koch snowflake's entire design fits in a single sentence — 'tent the middle third of every line, then repeat' — and yet its edge is so bumpy it has infinite length.
The catch
The tiny rule is amazing to store and describe — one sentence draws infinite detail that looks the same at every zoom — but the edge length runs away to infinity, so the snowflake has no finite outline length you could ever measure with a ruler; the more carefully you measure, the longer it gets. And 'endlessly detailed' is not the same as 'smooth': the snowflake edge is jagged everywhere, with a sharp corner at every single point and no straight stretch anywhere.
Questions kids ask
How can the edge be infinitely long but still fit in a small box?
Length and area are different things. Every round of the rule multiplies the edge length by 4/3, so it grows forever with no limit. But each new bump is tiny and tucked just outside the old edge, so the whole shape only ever pokes a little past the starting triangle and stays inside a small box. Infinite wiggly edge, finite space.
Is this how real things in nature get so detailed?
Yes, in spirit. Coastlines, fern leaves, snowflakes, lungs, and branching rivers all repeat a similar small pattern at smaller and smaller scales, which is why they look detailed and 'the same' whether you're up close or far away. A simple repeated rule packs in a lot of detail cheaply, which is exactly what a growing plant or a coastline does.
What does 'fractal' actually mean?
A fractal is a shape made by repeating a rule at smaller and smaller scales, so it looks similar no matter how far you zoom in. Because the detail never stops, a fractal can have a wiggly edge with no fixed length, and it behaves like it's 'between' a line and a flat area.
Does the snowflake ever finish?
No. Each round adds a finer layer of bumps, and you can always do the rule one more time, forever. A drawing has to stop after a few rounds because the bumps get too tiny to see, but the true fractal keeps going with detail at every zoom level.
Talk about it
- Guess first: if I gave you one short sentence of instructions, could you draw something with endless detail? How?
- Where have you seen the same shape repeat when you looked closer — a fern, a coastline, a snowflake? Why do you think nature does that?
- Which is bigger: a snowflake's edge or the line around our table? What if I told you the snowflake's edge never stops getting longer?
For grown-ups
This is the Koch snowflake, a classic fractal. Each round replaces every line segment with four segments each one-third as long, so the total perimeter multiplies by 4/3 every round and diverges to infinity, while the enclosed area converges to a finite value (8/5 of the starting triangle). The curve is self-similar and continuous but nowhere differentiable, with fractal (Hausdorff) dimension log4/log3 ≈ 1.262 — between a line and a filled area. Self-similar rules like this (and their cousins, L-systems and iterated function systems) are how just a few lines of code generate coastlines, ferns, lungs, and branching river networks.
Keep going
What else makes you wonder?
- If you zoom into the snowflake's edge forever, would you ever reach a 'last' bump, or does it really go on with no end?
- What other shapes in your house might be hiding a tiny repeated rule inside them?
- Could a single short rule, repeated, ever build something that looks alive, like a tree or a lung?