The Binomial Theorem

Pascal triangle hands you the coefficients.

Squaring a bracket

Multiply (a + b) by itself, every term by every term: (a + b)² = a² + ab + ba + b² = a² + 2ab + b². The coefficients are 1, 2 and 1. The middle one is 2 because ab turns up twice, once as a times b and once as b times a.

Drawn as a square of side a + b, the area splits into four pieces: a square a², a square b², and two rectangles, each a by b. The two rectangles are the 2ab.

Cubing it

To cube the bracket, multiply the square by one more bracket: (a + b)³ = (a + b)(a² + 2ab + b²).

Multiply every term by every term. a times the square gives a³ + 2a²b + ab². b times the square gives a²b + 2ab² + b³. Together that is a³ + 2a²b + ab² + a²b + 2ab² + b³.

Now collect the like terms: 2a²b + a²b = 3a²b and ab² + 2ab² = 3ab². So (a + b)³ = a³ + 3a²b + 3ab² + b³, and the coefficients are 1, 3, 3 and 1.

(a + b)³ = a³ + 3a²b + 3ab² + b³1 cube a³ · 3 slabs a²b · 3 rods ab² · 1 cube b³n = 2n = 3pull apart

a cube of side a + b, cut at a along each edge: pull it apart to count the pieces of each kind

Choose n = 3 and pull the cube apart

A cube of side a + b, cut at a along every edge. Pull it apart and count the pieces: one cube a³, three slabs a²b, three rods ab² and one small cube b³. Those counts are the coefficients 1, 3, 3, 1. Choose n = 2 to see the square come apart into a², two rectangles ab and b².

Pascal’s triangle

Write the coefficients of each power in a row, starting with (a + b)⁰ = 1 and (a + b)¹ = a + b. Row 0 is 1. Row 1 is 1, 1. Row 2 is 1, 2, 1. Row 3 is 1, 3, 3, 1.

Each number is the sum of the two numbers above it, and each row starts and ends with 1. In row 3, 3 = 1 + 2 and 3 = 2 + 1. This arrangement is called Pascal’s triangle.

Why the two above add

Look again at the cube. The a²b term came from two places: a² times b, with coefficient 1, and 2ab times a, with coefficient 2. So its coefficient is 1 + 2 = 3. Those are the two numbers above it in row 2.

The same happens in every row. Multiplying a row by one more bracket (a + b) sends each term two ways, once times a and once times b, and each new term collects one contribution from each of the two terms above it.

Every expansion from the triangle

So the triangle gives the coefficients without multiplying any brackets out. Row 4 is 1, 1 + 3, 3 + 3, 3 + 1, 1, which is 1, 4, 6, 4, 1. So (a + b)⁴ = a⁴ + 4a³b + 6a²b² + 4ab³ + b⁴. The power of a falls by one from term to term and the power of b rises by one, and in every term the two powers add up to 4.

Row 5 is 1, 5, 10, 10, 5, 1, and row 6 is 1, 6, 15, 20, 15, 6, 1. So (1 + x)⁵ = 1 + 5x + 10x² + 10x³ + 5x⁴ + x⁵. Put x = 1 to check the row: 1 + 5 + 10 + 10 + 5 + 1 = 32, and 2⁵ = 32.

row 01row 111row 2121row 31331row 414641row 515101051row 61615201561

Rows 0 to 6 of Pascal’s triangle. Each entry is the one directly above it plus the one above and to the left: in row 6, 20 = 10 + 10 and 15 = 5 + 10.

Each entry counts choices

Multiplying out (a + b)(a + b) means choosing either a or b from each bracket and adding up every possible result. There are four ways to choose: a then a, a then b, b then a, and b then b. Two of them give ab, so ab has coefficient 2.

For (a + b)⁴ there are four brackets. A term a²b² comes from choosing b from exactly two of the four brackets and a from the other two. The number of ways to choose 2 brackets out of 4 is 6: the first and second, first and third, first and fourth, second and third, second and fourth, or third and fourth. That is why the middle of row 4 is 6.

aa
bab
babab

The choices for (a + b)²: a or b from the first bracket, then a or b from the second. Two of the four paths, the colored one and the one below it, give ab, so the coefficient of ab is 2.

The general term

The number of ways to choose r things from n is written nCr. It can be worked out without the triangle: nCr = n! ÷ (r! × (n − r)!). For example, 4C2 = 4 × 3 / 2 = 6, and 5C2 = 5 × 4 / 2 = 10.

In (a + b)ⁿ, a term with b to the power r comes from choosing which r of the n brackets give a b. The other n − r brackets give an a. So each term is nCr aⁿ⁻ʳ bʳ, and the whole expansion is the sum of these terms for r = 0, 1, 2, up to n. This is the binomial theorem.

For the x³ term of (1 + x)⁵, choose 3 of the 5 brackets to give an x: 5C3 = 5 × 4 × 3 / (3 × 2 × 1) = 60 / 6 = 10, and every other bracket gives 1. So the term is 10x³.

When the first term is not 1

The powers of each term in the bracket go into every coefficient. In (2 + x)³ the row is 1, 3, 3, 1, and the powers of 2 fall from 2³ to 2⁰: (2 + x)³ = 1 × 8 + 3 × 4 × x + 3 × 2 × x² + 1 × x³ = 8 + 12x + 6x² + x³.

A minus sign belongs to its term. In (x − 2)⁴ the second term is −2, and its odd powers are negative: (x − 2)⁴ = x⁴ + 4x³(−2) + 6x²(−2)² + 4x(−2)³ + (−2)⁴ = x⁴ − 8x³ + 24x² − 32x + 16. Check at x = 3: (3 − 2)⁴ = 1, and 81 − 216 + 216 − 96 + 16 = 1.

One term can be found without the rest. For the x³ term of (2x + 3)⁵, the 2x must be chosen from 3 brackets and the 3 from the other 2: 5C2 × (2x)³ × 3² = 10 × 8x³ × 9 = 720x³.

The usual mistakes

Using the row before. (a + b)³ needs row 3, 1, 3, 3, 1. Row 2, 1, 2, 1, belongs to the square.

Taking 4C1 for the a²b² term of (a + b)⁴. 4C1 = 4 counts the ways to choose one bracket, which gives the a³b term. The a²b² term needs two brackets chosen, 4C2 = 6.

Counting the choices in order. Choosing the first and then the third bracket is the same choice as the third and then the first, so 4 × 3 = 12 counts each choice twice; halve it to 6.

Multiplying instead of choosing. The coefficient of x³ in (1 + x)⁵ is 5C3 = 10, not 3 × 5 = 15.

Leaving out the powers of the first term. (2 + x)³ is 8 + 12x + 6x² + x³, not 1 + 3x + 3x² + x³.

Two applications

In the first application below, six coin tosses are the six brackets of (1 + x)⁶: an x for a head and a 1 for a tail. The coefficient of x² counts the sequences with exactly two heads.

In the second, every length of a block is multiplied by 1 + x, so its volume is multiplied by (1 + x)³ = 1 + 3x + 3x² + x³. When x is small, the first two terms give a close estimate.

Worked example: Tossing a Coin Six Times: The Chance of Exactly Two Heads from Pascal's Triangle

Question A fair coin is tossed 6 times. (a) Use the binomial theorem to expand (1 + x)6, and explain why the coefficient of x2 is the number of sequences of tosses with exactly two heads. Find the probability of exactly two heads. (b) Find the probability of at least five heads.

  1. 1.Each of the 6 tosses has 2 outcomes, so there are 26 = 64 sequences of heads and tails, and they are all equally likely.

    1111211331146411510105116152015616 tosses: 26= 64 sequencesall equally likely
    1111211331146411510105116152015616 tosses: 26= 64 sequencesall equally likely
    Each toss is a head or a tail, so six tosses give 26 = 64 sequences, all equally likely. Pascal's triangle is drawn down to row 6.
  2. 2.By the binomial theorem, (1 + x)6 = 1 + 6x + 15x2 + 20x3 + 15x4 + 6x5 + x6. The coefficients are row 6 of Pascal's triangle, 60, 61, …, 66.

    111121133114641151010511615201561row 6(1 + x)6= 1 + 6x + 15x2+ 20x3+ 15x4+ 6x5+ x6
    111121133114641151010511615201561row 6(1 + x)6= 1 + 6x + 15x2+ 20x3+ 15x4+ 6x5+ x6
    By the binomial theorem the coefficients of (1 + x)6 are row 6 of Pascal's triangle: 1, 6, 15, 20, 15, 6, 1.
  3. 3.Think of the six brackets as the six tosses: take x from a bracket for a head and 1 for a tail. A term in x2 takes x from exactly two brackets, so its coefficient 62 = 6 × 52 × 1 = 15 is the number of sequences with exactly two heads.

    111121133114641151010511615201561row 6x from 2 brackets (heads), 1 from the other 4(6 × 5)/(2 × 1) = 15 sequences
    111121133114641151010511615201561row 6x from 2 brackets (heads), 1 from the other 4(6 × 5)/(2 × 1) = 15 sequences
    Choosing x from a bracket is a head and 1 is a tail. The term in x2 takes x from two brackets, in 62 = 15 ways.
  4. 4.(a) The probability of exactly two heads is 1564 ≈ 0.234.

    111121133114641151010511615201561row 6P(exactly 2 heads) = 15/64about 0.234
    111121133114641151010511615201561row 6P(exactly 2 heads) = 15/64about 0.234
    (a) Exactly two heads happen in 15 of the 64 sequences, so the probability is 1564 ≈ 0.234.
  5. 5.At least five heads means five heads or six heads: 65 + 66 = 6 + 1 = 7 sequences. (b) The probability is 764 ≈ 0.109. Check: the whole row adds up to 1 + 6 + 15 + 20 + 15 + 6 + 1 = 64, so every sequence is counted once.

    111121133114641151010511615201561row 65 or 6 heads: 6 + 1 = 7 sequencesP(at least 5 heads) = 7/64check: 1 + 6 + 15 + 20 + 15 + 6 + 1 = 64
    111121133114641151010511615201561row 65 or 6 heads: 6 + 1 = 7 sequencesP(at least 5 heads) = 7/64check: 1 + 6 + 15 + 20 + 15 + 6 + 1 = 64
    (b) Five heads happen in 6 sequences and six heads in 1, so the probability of at least five heads is 764. The row adds up to 64.

Answer: (a) (1 + x)6 = 1 + 6x + 15x2 + 20x3 + 15x4 + 6x5 + x6; 15 of the 64 sequences have exactly two heads, so the probability is 1564; (b) 764

Common mistakes

  • Taking the probability of two heads as (12)2 = 14. That is the chance of two heads in two tosses. Here the other four tosses must be tails, and the two heads can fall in any of 15 places, so the probability is 15 × (12)6 = 1564.
  • Counting only five heads for "at least five", which gives 664. At least five includes six heads as well, so the count is 6 + 1 = 7.

More polynomials and the binomial theorem problems, worked step by step →

Worked example: A Block of Rubber That Swells by 2% in Every Direction: The Change in Its Volume from the First-Order Term

Question A block of rubber is a cuboid 10 cm by 8 cm by 5 cm. Left in oil, it swells so that every length grows by 2%. (a) Expand (1 + x)3, and use its first two terms to estimate the percentage increase in the volume of the block and its new volume. (b) Find the new volume exactly, to 2 decimal places, and the error in the estimate. Which term of the expansion makes up most of the error?

  1. 1.Every length is multiplied by 1 + x, where x = 0.02. The volume is multiplied by (1 + x)3 = 1 + 3x + 3x2 + x3.

    10 cm5 cm8 cmevery length × (1 + x), with x = 0.02volume × (1 + x)3= 1 + 3x + 3x2+ x3
    10 cm5 cm8 cmevery length × (1 + x), with x = 0.02volume × (1 + x)3= 1 + 3x + 3x2+ x3
    Every length of the block is multiplied by 1 + x with x = 0.02, so its volume is multiplied by (1 + x)3 = 1 + 3x + 3x2 + x3.
  2. 2.For a small x, the terms 3x2 and x3 are tiny, so (1 + x)3 ≈ 1 + 3x = 1 + 0.06 = 1.06. The volume grows by about 3 × 2% = 6%.

    10 cm5 cm8 cmx is small: (1 + x)3≈ 1 + 3x = 1.06about 3 × 2% = 6% more
    10 cm5 cm8 cmx is small: (1 + x)3≈ 1 + 3x = 1.06about 3 × 2% = 6% more
    For a small x the terms 3x2 and x3 are tiny, so (1 + x)3 ≈ 1 + 3x = 1.06: about 6% more.
  3. 3.(a) The volume was 10 × 8 × 5 = 400 cm3, so the new volume is about 400 × 1.06 = 424 cm3, an increase of about 6%.

    10 cm5 cm8 cm10 × 8 × 5 = 400 cubic cmestimate: 400 × 1.06 = 424 cubic cm
    10 cm5 cm8 cm10 × 8 × 5 = 400 cubic cmestimate: 400 × 1.06 = 424 cubic cm
    (a) The block holds 400 cm3, so after swelling it holds about 400 × 1.06 = 424 cm3.
  4. 4.Exactly, 1.023 = 1 + 0.06 + 0.0012 + 0.000008 = 1.061208, so the new volume is 400 × 1.061208 = 424.4832, which is 424.48 cm3. Check with the new lengths: 10.2 × 8.16 × 5.1 = 424.4832.

    10.2 cm5.1 cm8.16 cm1.023= 1 + 0.06 + 0.0012 + 0.000008400 × 1.061208 = 424.4832 cubic cm
    10.2 cm5.1 cm8.16 cm1.023= 1 + 0.06 + 0.0012 + 0.000008400 × 1.061208 = 424.4832 cubic cm
    The swollen block, drawn to scale over the first, is 10.2 cm by 8.16 cm by 5.1 cm, and 400 × 1.061208 = 424.4832 cm3.
  5. 5.(b) The estimate is short by 424.48 − 424 = 0.48 cm3. The term 3x2 is worth 400 × 0.0012 = 0.48 cm3 of it, and the term x3 only 400 × 0.000008 = 0.0032 cm3.

    10.2 cm5.1 cm8.16 cmshort by 424.48 − 424 = 0.48 cubic cmthe 3x2term: 400 × 0.0012 = 0.48
    10.2 cm5.1 cm8.16 cmshort by 424.48 − 424 = 0.48 cubic cmthe 3x2term: 400 × 0.0012 = 0.48
    (b) The estimate is 0.48 cm3 short. The term 3x2 is worth 0.48 cm3 and the term x3 only 0.0032 cm3.

Answer: (a) (1 + x)3 = 1 + 3x + 3x2 + x3; an increase of about 6%, to about 424 cm3; (b) 424.48 cm3, so the estimate is 0.48 cm3 short, almost all of it the term 3x2

Common mistakes

  • Saying that the volume also grows by 2%, because every length does. The volume is a product of three lengths, so each of the three growths adds about 2%, and the volume grows by about 6%.
  • Adding 2% three times to the lengths and then multiplying, 10.6 × 8.48 × 5.3. Each length grows by 2% once, to 10.2, 8.16 and 5.1; it is the volume, not a length, that grows by about 6%.

More polynomials and the binomial theorem problems, worked step by step →

Practice The Binomial Theorem in the app