Pascal's Triangle for Kids: Find the Hidden Patterns

Pascal’s triangle gets handed to children as a curiosity, and that is why it bores them. Built properly it does two things no curiosity does. Shade in every even number and a fractal appears out of nowhere, drawn by nothing but addition. Then flip four coins thirty-two times, tally how many heads you got, and the counts come out as 1:4:6:4:1, which is row four of the triangle you just drew. Pascal’s triangle for kids is worth an afternoon because it stops being a pattern and becomes a prediction.

Key Takeaways

• You’ll build a 12-row triangle with real numbers, a 20-row triangle marked only odd or even, and then test row four with coins.
• It teaches that a single rule, add the two numbers above, generates powers of two, triangular numbers, Fibonacci and a fractal.
• Best from 8. The addition is easy, but staying accurate for 12 rows needs the patience of a solid eight-year-old.
• One arithmetic slip poisons every row below it. Check row sums as you go; they double every time.
• Skip the full numbers for the fractal. Marking only odd or even gets you to 20 rows in ten minutes instead of an hour.

What You’ll Need

Item	How much	Where to get it	If you don’t have it
Large paper	2 sheets of A3, or 4 sheets of A4 taped together	Printer, or the back of wrapping paper	Squared exercise paper, one number per two squares
Pencil with an eraser	1	School supplies	A pen makes the inevitable correction messy
Ruler	1	School supplies	Fold lines into the paper instead
Two crayons or highlighters	2 colours	Anywhere	Shade with pencil cross-hatching
Coins	4 identical	Pocket	Four bottle caps, marked on one side
Calculator	1, optional	Phone	Only for checking row sums

Time: numbered triangle 25 min, shading 15 min, coin test 20 min Cost: nothing Mess level: low

Do the Experiment: Pascal’s Triangle for Kids, Row by Row

Step 1 – Set out the grid before writing any numbers

Rule 12 rows of circles in a triangle, staggered so each circle sits between two above it. Row one has one circle, row twelve has twelve. Leave 3 cm per circle because the middle numbers get to three digits. Doing the grid first stops the triangle drifting sideways, which is the usual reason these end up lopsided.

Step 2 – Fill in the rule, out loud

Put 1 at the top. Put 1 and 1 in row two. From then on, every number is the sum of the two directly above it, and the ends are always 1 because there is only one number above them. Say the addition aloud for the first three rows and then let them take over.

Step 3 – Check each row by adding it up

As each row finishes, total it. You should get 1, 2, 4, 8, 16, 32, 64 and so on, doubling every time. This is a running error check and it is the difference between a triangle that works and twenty minutes of silent wrongness. Row 12 should total 2,048.

Step 4 – Hunt the diagonals

Read down the second diagonal: 1, 2, 3, 4, 5. Now the third: 1, 3, 6, 10, 15, 21. Those are the triangular numbers, the ones you get stacking bowling pins, and they are sitting in a triangle built by addition. Have your kid draw dots to confirm that 10 really does make a triangle.

Step 5 – Build the parity triangle and shade it

On the second sheet, draw 20 rows of circles and do not calculate anything. Write only E for even or O for odd, using three rules: even plus even is even, odd plus odd is even, odd plus even is odd. Twenty rows takes about ten minutes this way. Then colour in every even circle.

Step 6 – Look at what you shaded

A pattern of triangles within triangles appears, each one made of three smaller copies of itself with a hole in the middle. That is the Sierpinski triangle, a fractal, produced by nothing more than keeping track of odd and even. It does not become convincing until about row 16, which is why the parity shortcut matters.

Step 7 – Test row four with coins

Flip four coins at once and record how many came up heads: 0, 1, 2, 3 or 4. Do this 32 times. Row four of your triangle is 1, 4, 6, 4, 1, which adds to 16. So in 32 trials, expect roughly 2, 8, 12, 8, 2. Compare your tally with the row.

What you should see: The numbered triangle takes 20 to 25 minutes and almost always contains one error, caught by the row-sum check. The parity triangle is fast and the Sierpinski pattern is unmistakable by row 16: three big triangles with a hole, each subdivided again. The coin test is noisier than children expect. Expect two heads to be the most common outcome and zero or four heads to be rare, but do not expect exactly 2, 8, 12, 8, 2. With 32 trials, counts within three or four of the prediction are normal, and saying that before you start saves the result from feeling like a failure.

What’s Actually Going On

Every number in Pascal’s triangle counts something specific: the number of different routes from the top to that position, moving only down-left or down-right. There is one way to reach either end of any row, which is why the edges are 1. There are six ways to reach the middle of row four, which is why that entry is 6. The addition rule works because any route to a position must have passed through one of the two positions above it, so you add the routes to each.

That counting interpretation is why the coins work. Flipping four coins is the same as walking four steps down the triangle, with heads as left and tails as right. There are 2⁴ = 16 equally likely sequences. Exactly one gives four heads, and exactly six give two heads, because there are six ways to choose which two of the four coins land heads. Row four is therefore the exact probability distribution of the coin experiment, and that is why the row sum is 16: 1 + 4 + 6 + 4 + 1. Row sums double each time because each extra coin doubles the number of possible sequences.

The diagonals follow from the same idea. The second diagonal counts the ways to choose one item from n, which is just n. The third counts the ways to choose two from n, which gives the triangular numbers. Add along the shallow diagonals and you get the Fibonacci sequence, which is harder to see and worth a ruler.

The fractal is the strangest result and has a clean explanation. Odd and even behave consistently under addition, so you can run the whole triangle using only two symbols. The resulting pattern of odd entries is self-similar: the top half of any big block repeats, scaled down, in each of two blocks below, leaving an even-numbered hole between them. Keep going and you approach the Sierpinski triangle. A fractal has emerged from integer addition with no geometry anywhere in the process, which is a genuinely remarkable thing to have on your kitchen table.

Finally, the name. Yang Hui set the triangle out in his Detailed Analysis of the Mathematical Rules in the Nine Chapters, written in 1261, and said he had learnt it from Jia Xian, whose own treatise is lost. It appears earlier still in Indian and Persian work on combinations. Pascal wrote his treatise on it in the 1650s; it was published in 1665, three years after his death, and his name stuck in English and French. It is Yang Hui’s triangle in China and Tartaglia’s triangle in Italy. Telling a child that the same mathematics was found independently in several places, and that the name depends on where you live, is a better history lesson than a date.

How to Explain It to Your Kid

Ages 5–8: add the two above

No names, no patterns. Draw six rows of big circles and let them do the additions with you. Every number is the two above it added up. That is the whole activity, and the arithmetic practice is real: forty small additions with an immediate self-check, which beats a worksheet. Then show them the 1s running down each side and ask why those never change.

Ages 9–12: find the row sums and the fractal

Introduce the row-sum check and let them discover the doubling themselves by writing the totals in a column. Then do the parity triangle. The question that matters comes after the shading: “we only wrote odd and even, so where did the picture come from?” You do not have to answer it well. Sitting with the strangeness is the right response.

Ages 13+: combinations and the binomial expansion

Give them the notation: the entry in row n, position k, is n choose k, the number of ways to pick k things from n. Then expand (a + b)⁴ by hand and watch Pascal’s row four fall out as the coefficients. Then set the real investigation: run the coin experiment with six coins, 64 trials, and compare against row six. Compute the expected count for each outcome and the standard deviation, and check how many of their counts sit within one.

The question to ask: “If I flip six coins instead of four, which number of heads will be most common, and how much more common will it be than getting all six heads? Work it out from the triangle before we flip anything.”

Why Teaching This Early Pays Off

Pascal’s triangle is the only object in school mathematics that connects arithmetic, probability, algebra and fractal geometry in one picture a child can hold. A kid who has built one has a place to put the binomial theorem when it arrives, and a reason to believe that combinations are counting rather than a formula.

The specific payoff is the link between counting and chance. Probability goes wrong for most people because they never see that a probability is a count of possibilities divided by a count of possibilities. Here the count is written in front of them in their own handwriting, and then they measure it with coins. The Framework for K-12 Science Education calls this using mathematics and computational thinking, and the coin tally is the same analysing-data practice as any science experiment.

Keep the claims honest. Nothing about one afternoon changes test scores, and the research does not say otherwise. What it does say is more useful: Chi and Wylie’s ICAP framework ranks constructive work, where the learner produces something beyond what they were given, above merely active work. Building the triangle is active. Predicting the coin distribution from it and then being partly wrong is constructive, and that is where the learning sits.

Before You Start: What Parents Should Know

Best age	8 and up. The arithmetic is simple but there is a lot of it, and 12 rows demands sustained accuracy. A six-year-old can do six rows happily with you doing the checking.
Supervision	You are the error-checker. Verify each row sum as it completes rather than at the end, because a mistake in row 5 ruins everything below it.
Real hazards	None. This is paper and pencil. Coins are a choking hazard for under-3s, so put them away afterwards.
Setup	A table and a large flat sheet. This does not work in an exercise book: row 12 is 12 numbers wide and needs about 35 cm.
Frustration point	Around row 9, when the numbers reach three digits and the additions stop being automatic. Offer to take alternate rows. Sharing the work does not spoil anything here.
Don’t	Don’t compute full numbers for the fractal version. Twenty rows of real arithmetic is an hour of work and the error rate makes the picture fail.

When It Doesn’t Work

• The row sums stop doubling. There is an addition error in that row or the one above. Find the first row whose total is wrong and fix only that row, then recompute downward. Do not start again.
• The triangle drifts off the side of the paper. The circles were not staggered consistently. This is why step 1 exists; redraw the grid rather than trying to squeeze row 12 into the margin.
• The Sierpinski pattern looks like noise. Too few rows. It is genuinely unconvincing before row 14 and obvious by row 18. Extend the parity triangle; it is quick.
• The coin tally does not match 2, 8, 12, 8, 2. It usually will not, exactly. Thirty-two trials is a small sample. If two heads is the most common and the extremes are rarest, the experiment worked. Double to 64 trials and it tightens visibly.
• A younger child loses interest halfway. Stop at row eight and go straight to the coins. The connection between the triangle and the coins is the memorable part, and it works fine with the smaller triangle.

Take It Further

Shade multiples of 3 instead of 2. A different self-similar pattern appears, built on triangles of a different size. Then try multiples of 5. Comparing the three shadings is a serious investigation into what the prime number does to the structure, and a teenager can get real traction on it.

Run the coin experiment at two sample sizes and compare. Do 32 trials with four coins, then 128. Compute how far each count is from the prediction as a percentage. The percentages shrink as the sample grows, which is the law of large numbers showing up in a completely different activity from the usual dice one.

What to Watch For Over the Next 3 Months

• Week 4: They draw a Sierpinski triangle for its own sake, usually on the back of something else, and sometimes get the recursion right.
• Month 2: The phrase “how many ways” appears when they are reasoning about something unrelated, which is the sign the counting idea transferred.
• Month 3: The question worth encouraging is why the coin flips and the triangle agree. If they can explain that in their own words, they have understood more of probability than most adults.

Frequently Asked Questions

How many rows do we actually need?

Twelve for the numbers, diagonals and row sums, which is about 25 minutes of arithmetic. For the Sierpinski fractal you need at least 16 and ideally 20, which is only practical if you mark odd and even instead of calculating the values.

Why does shading the even numbers make a fractal?

Because odd and even behave consistently under addition, so the triangle can be run with just two symbols, and that two-symbol version is self-similar. Each large block of odd entries reappears, scaled down, in two blocks below it, with an even-numbered gap between them.

Did Pascal discover the triangle?

No. Yang Hui wrote it out in 1261 and credited Jia Xian, whose work is lost, and it appears earlier in Indian and Persian mathematics. Pascal’s treatise was published in 1665, after his death. It is called Yang Hui’s triangle in China and Tartaglia’s triangle in Italy.

Our coin results looked nothing like the prediction. Did we do it wrong?

Probably not. With 32 trials, random variation of three or four counts in each column is completely normal. What should hold is the shape: two heads most common, zero and four heads rarest. Increase to 128 trials and the match tightens noticeably.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. O’Connor, J. J., & Robertson, E. F. “Yang Hui.” MacTutor History of Mathematics Archive, University of St Andrews. https://mathshistory.st-andrews.ac.uk/Biographies/Yang_Hui/
2. O’Connor, J. J., & Robertson, E. F. “Blaise Pascal.” MacTutor History of Mathematics Archive, University of St Andrews. https://mathshistory.st-andrews.ac.uk/Biographies/Pascal/
3. NRICH, University of Cambridge. “Exploring Wild and Wonderful Number Patterns.” https://nrich.maths.org/problems/exploring-wild-wonderful-number-patterns
4. Chi, M. T. H., & Wylie, R. (2014). “The ICAP Framework: Linking Cognitive Engagement to Active Learning Outcomes.” Educational Psychologist, 49(4), 219–243. https://doi.org/10.1080/00461520.2014.965823
5. Alfieri, L., Brooks, P. J., Aldrich, N. J., & Tenenbaum, H. R. (2011). “Does discovery-based instruction enhance learning?” Journal of Educational Psychology, 103(1), 1–18. https://doi.org/10.1037/a0021017
6. National Research Council. (2012). A Framework for K-12 Science Education. National Academies Press. https://doi.org/10.17226/13165

Related reading: rolling 1,000 dice to see probability, counting Fibonacci spirals in real pinecones, and pattern recognition across maths and music.