← Back to FractionRush

Binomial Distribution S1 Statistics

Grade 11 · Statistics 1 · Cambridge A-Level 9709 · Age 16–17

Welcome to the Binomial Distribution!

The binomial distribution is one of the most important discrete probability distributions in Cambridge A-Level Statistics 1 (9709). It models situations where you repeat a fixed number of independent trials, each with only two possible outcomes and a constant probability of success. Real-world applications range from quality control testing to genetics, medical trials, and opinion polls.

X ~ B(n, p)  |  P(X=r) = nCr × pr × (1−p)n−r  |  E(X) = np  |  Var(X) = np(1−p)

Learning Objectives

  • State the four conditions required for a binomial distribution to apply
  • Recognise when a binomial model is appropriate (and when it is not)
  • Use the formula P(X=r) = nCr pr(1−p)n−r to calculate exact probabilities
  • Calculate the mean E(X) = np and variance Var(X) = np(1−p) = npq
  • Use cumulative probabilities: P(X≤r), P(X≥r), P(a≤X≤b)
  • Use a calculator (binompdf / binomcdf) and binomial tables efficiently
  • Find n or p given information about probabilities or moments
  • Identify critical regions for hypothesis testing (introductory)

Four Conditions

Fixed n, independent trials, two outcomes, constant p

P(X=r) Formula

nCr pr qn-r — choose × success × failure

Mean E(X)

np — expected number of successes

Variance Var(X)

npq where q = 1−p

Cumulative P(X≤r)

Sum P(X=0)+…+P(X=r)

Complement

P(X≥r) = 1−P(X≤r−1)

Calculator

binompdf(n,p,r) and binomcdf(n,p,r)

Critical Region

Tail where P < significance level

Learn 1 — Binomial Conditions

The Four Conditions for B(n, p)

We say X follows a binomial distribution, written X ~ B(n, p), if and only if ALL four of the following conditions hold:

1. Fixed number of trials n    2. Trials are independent    3. Only two outcomes per trial (success / failure)    4. Constant probability p of success at each trial

Condition 1 — Fixed Number of Trials

The experiment is repeated exactly n times, and n is known in advance before the experiment starts.

Example: "A fair coin is tossed 10 times" → n = 10 is fixed. ✓
Counter-example: "Toss until you get a head" → n is not fixed. This is a geometric distribution, not binomial.

Condition 2 — Independent Trials

The outcome of each trial does not affect any other trial. Knowing one outcome gives no information about another.

Example: Rolling a die repeatedly — each roll is independent. ✓
Counter-example: Drawing cards one by one without replacement from a pack — removing a card changes the composition, so trials are not independent. ✗

Condition 3 — Two Outcomes Only

Each trial results in exactly one of two mutually exclusive outcomes, called success and failure. We choose which to label "success" based on what we are counting.

Example: Testing whether a component is defective (defective = success, or non-defective = success — we choose). ✓
The labels are arbitrary; what matters is that there are exactly two categories.

Condition 4 — Constant Probability of Success

The probability of success, denoted p, is the same for every trial. Consequently, the probability of failure q = 1 − p is also constant.

Example: Each seed from a batch has a 0.8 probability of germinating (assuming independence). p = 0.8 throughout. ✓
Counter-example: Testing items from a production line where quality changes over time — p is not constant. ✗

Identifying Binomial Situations

Checklist: For any context, ask:
• Is n fixed? → If not, NOT binomial.
• Are trials independent? → If not, NOT binomial.
• Are there exactly two outcomes? → If not, consider other distributions.
• Is p constant? → If not, NOT binomial.

Only if all four hold can you write X ~ B(n, p).

When Binomial Does NOT Apply — Sampling Without Replacement

Key Scenario: Selecting items from a small, finite population without replacement violates independence — once an item is removed, the probabilities change for remaining draws.

Example: A box contains 5 red and 3 blue balls. Draw 3 without replacement. X = number of red balls.
P(1st red) = 5/8, but P(2nd red | 1st red) = 4/7 ≠ 5/8. Not constant → NOT binomial.

Correct model: Hypergeometric distribution.

Exception: If the population is very large relative to sample size, the probabilities barely change and we treat it as approximately binomial. A common rule: if sample < 10% of population, binomial is a good approximation.

Worked Identification Examples

Example 1: A biased coin with P(Head) = 0.6 is tossed 8 times. X = number of heads.
n = 8 (fixed) ✓ | independent ✓ | two outcomes: head/tail ✓ | p = 0.6 constant ✓
X ~ B(8, 0.6)
Example 2: A factory makes bolts; 5% are defective. 20 bolts chosen at random (with replacement). X = number defective.
n = 20 ✓ | independent (with replacement) ✓ | defective/not ✓ | p = 0.05 ✓
X ~ B(20, 0.05)
Example 3: A bag has 4 green and 6 white balls. 3 are drawn without replacement. X = green balls drawn.
n = 3 ✓ | NOT independent (without replacement, small population) ✗
Binomial does NOT apply. Use hypergeometric.

Learn 2 — The P(X = r) Formula

The Binomial Probability Formula

If X ~ B(n, p), then the probability that X takes the value r (exactly r successes in n trials) is:

P(X = r) = nCr × pr × (1−p)n−r    for r = 0, 1, 2, …, n

Where nCr = n! / (r!(n−r)!) is the binomial coefficient — the number of ways to choose which r of the n trials are successes.

Understanding Each Part

nCr: the number of arrangements of r successes among n trials (e.g. SSFFS... in different orders)
pr: probability of getting exactly r successes (each success has probability p)
(1−p)n−r: probability of getting exactly n−r failures (each failure has probability q = 1−p)

We multiply them because each specific arrangement has probability pr × qn-r, and there are nCr such arrangements.

Step-by-Step Calculation

Example: X ~ B(5, 0.3). Find P(X = 2).

Step 1: Identify n = 5, p = 0.3, q = 0.7, r = 2.

Step 2: Calculate 5C2 = 5!/(2!×3!) = (5×4)/(2×1) = 10

Step 3: Calculate pr = (0.3)2 = 0.09

Step 4: Calculate qn−r = (0.7)3 = 0.343

Step 5: Multiply: P(X=2) = 10 × 0.09 × 0.343 = 0.3087

Special Cases: P(X = 0) and P(X = n)

P(X = 0): All trials are failures.
P(X=0) = nC0 × p0 × qn = 1 × 1 × qn = (1−p)n

P(X = n): All trials are successes.
P(X=n) = nCn × pn × q0 = 1 × pn × 1 = pn
Example: X ~ B(4, 0.2). Find P(X=0) and P(X=4).
P(X=0) = (0.8)4 = 0.4096
P(X=4) = (0.2)4 = 0.0016

Symmetry when p = 0.5

When p = 0.5, the binomial distribution is perfectly symmetric: P(X = r) = P(X = n−r) for all r.

This is because nCr = nCn−r and prqn−r = qrpn−r when p = q = 0.5.

Example: X ~ B(6, 0.5): P(X=2) = P(X=4) and E(X) = 3. The distribution is symmetric about 3.

Another Full Example

Example: A multiple-choice test has 10 questions, each with 4 options. A student guesses randomly. Find the probability of getting exactly 3 correct.

X ~ B(10, 0.25) (p = 1/4 for correct guess, n = 10)

P(X=3) = 10C3 × (0.25)3 × (0.75)7
= 120 × 0.015625 × 0.13348...
= 120 × 0.002086... = 0.2503 (4 s.f.)
Always write down X ~ B(n, p) first. State n, p and the value of r explicitly before applying the formula. This earns method marks in the exam even if arithmetic goes wrong.

Learn 3 — Mean & Variance of B(n, p)

The Formulas

E(X) = np     Var(X) = np(1−p) = npq     SD(X) = √(npq)

where q = 1 − p is the probability of failure. These formulas are given in the Cambridge formula booklet, but you must be able to apply them quickly and reverse them to find n or p.

Worked Examples

Example 1: X ~ B(20, 0.4). Find E(X), Var(X) and SD(X).

E(X) = np = 20 × 0.4 = 8
Var(X) = npq = 20 × 0.4 × 0.6 = 4.8
SD(X) = √4.8 ≈ 2.191
Example 2: X ~ B(50, 0.15). Find E(X) and Var(X).

E(X) = 50 × 0.15 = 7.5
Var(X) = 50 × 0.15 × 0.85 = 6.375

Interpretation

• E(X) = np is the expected (average) number of successes in n trials with probability p each.
• If you toss a fair coin 100 times, you expect E(X) = 100 × 0.5 = 50 heads.
• Var(X) = npq measures how variable the count of successes is. Larger p(1−p) means more spread.
The product p(1−p) is maximised when p = 0.5, giving maximum spread. For p near 0 or 1, the variance is small — the distribution is concentrated at one end.

Finding n and p from Given Information

Example 3: X ~ B(n, p). Given E(X) = 6 and Var(X) = 4.2, find n and p.

Step 1: Var(X) / E(X) = npq / np = q = 4.2 / 6 = 0.7
So q = 0.7, giving p = 0.3

Step 2: E(X) = np ⟹ n = E(X) / p = 6 / 0.3 = 20
Check: Var(X) = 20 × 0.3 × 0.7 = 4.2 ✓
Example 4: X ~ B(n, 0.4). Given Var(X) = 7.2, find n.

Var(X) = n × 0.4 × 0.6 = 0.24n = 7.2
n = 7.2 / 0.24 = 30

Standard Deviation as a Spread Measure

Example 5: A production line produces items, each 2% defective. 500 items inspected. Find E(X) and SD(X) for the number of defective items.

X ~ B(500, 0.02)
E(X) = 500 × 0.02 = 10
Var(X) = 500 × 0.02 × 0.98 = 9.8
SD(X) = √9.8 ≈ 3.130

Interpretation: on average 10 defective items, typically varying by about 3.
Note: Var(X) = npq, NOT np²q and NOT np(1+p). The most common error is writing np² instead of npq. Always double-check by ensuring q = 1 − p.

Learn 4 — Cumulative Probabilities

Key Cumulative Probability Rules

P(X ≤ r) = Σk=0r P(X=k)     P(X ≥ r) = 1 − P(X ≤ r−1)     P(X = r) = P(X ≤ r) − P(X ≤ r−1)
P(a ≤ X ≤ b) = P(X ≤ b) − P(X ≤ a−1)

P(X ≤ r) — Cumulative Probability

P(X ≤ r) is the sum of all probabilities from 0 up to r:
P(X ≤ r) = P(X=0) + P(X=1) + … + P(X=r)

Example: X ~ B(4, 0.3). Find P(X ≤ 2).

P(X=0) = (0.7)4 = 0.2401
P(X=1) = 4C1(0.3)(0.7)3 = 4 × 0.3 × 0.343 = 0.4116
P(X=2) = 4C2(0.3)2(0.7)2 = 6 × 0.09 × 0.49 = 0.2646
P(X ≤ 2) = 0.2401 + 0.4116 + 0.2646 = 0.9163

P(X ≥ r) — Using the Complement

Critical formula: P(X ≥ r) = 1 − P(X ≤ r−1)

NOT 1 − P(X ≤ r). The complement of "at least r" is "at most r−1", NOT "at most r".

Example: X ~ B(4, 0.3). Find P(X ≥ 2).
P(X ≥ 2) = 1 − P(X ≤ 1) = 1 − [0.2401 + 0.4116] = 1 − 0.6517 = 0.3483

P(a ≤ X ≤ b) — Range Probabilities

Example: X ~ B(10, 0.4). Find P(3 ≤ X ≤ 6).

P(3 ≤ X ≤ 6) = P(X ≤ 6) − P(X ≤ 2)

Using calculator or summing:
P(X ≤ 6) ≈ 0.9452
P(X ≤ 2) ≈ 0.1672
P(3 ≤ X ≤ 6) ≈ 0.9452 − 0.1672 = 0.7780

Step-by-Step Cumulative Method

Strategy for any cumulative probability:
1. Write down X ~ B(n, p)
2. Convert the event to one involving P(X ≤ k) or its complement
3. Add individual probabilities or use calculator/tables
4. Show all working for marks

Useful conversions:
• P(X < r) = P(X ≤ r−1)
• P(X > r) = 1 − P(X ≤ r)
• P(X ≥ r) = 1 − P(X ≤ r−1)
• P(X = r) = P(X ≤ r) − P(X ≤ r−1)

Finding Critical Regions (Introductory)

A critical region is the set of values of X for which the observed result would be considered statistically significant at a given significance level α.

For a one-tailed lower test at 5%: find the largest r such that P(X ≤ r) < 0.05.
For a one-tailed upper test at 5%: find the smallest r such that P(X ≥ r) < 0.05, i.e. 1 − P(X ≤ r−1) < 0.05.

Example: X ~ B(20, 0.5) under H0. Find the critical region for an upper-tail test at 5%.
Need smallest r with P(X ≥ r) < 0.05, i.e. P(X ≤ r−1) > 0.95.
P(X ≤ 13) ≈ 0.9423 < 0.95; P(X ≤ 14) ≈ 0.9793 > 0.95
Critical region: X ≥ 15 (since P(X ≥ 15) = 1 − P(X ≤ 14) ≈ 0.0207 < 0.05).
When summing individual binomial probabilities by hand, build a table row by row. Round intermediate values to at least 5 s.f. to avoid accumulation of rounding errors.

Learn 5 — Using Binomial Tables & Calculator

Using a Scientific Calculator

Most modern scientific calculators and graphical calculators have built-in binomial functions:

binompdf(n, p, r) — gives P(X = r) directly
binomcdf(n, p, r) — gives P(X ≤ r) directly (cumulative)

On a Casio fx-CG50 or similar: MENU → STAT → DIST → BINM → Bpd (pdf) or Bcd (cdf)
Enter n, p, x (= r), then Execute.

Using Binomial Cumulative Distribution Tables

Cambridge 9709 exams sometimes provide a booklet of cumulative binomial tables. These give P(X ≤ r) for selected values of n and p.

How to read the table:
1. Find the section for your n (e.g. n = 10)
2. Find the column for your p (e.g. p = 0.30)
3. Find the row for your r (e.g. r = 4)
4. Read off P(X ≤ 4) directly from the table

Deriving other probabilities from table values:
• P(X = r) = table(r) − table(r−1)
• P(X ≥ r) = 1 − table(r−1)
• P(X > r) = 1 − table(r)

Strategy: When to Use Which Method

Single exact probability, small n: Use the formula directly. Show all working.
Single exact probability, large n: Use binompdf on calculator or tables.
Cumulative probability: Use binomcdf on calculator or tables.
Range P(a ≤ X ≤ b): binomcdf(n,p,b) − binomcdf(n,p,a−1)
Complement P(X ≥ r): 1 − binomcdf(n,p,r−1)

Complementary Events to Simplify Calculations

When computing P(X ≥ r) for large r, complement reduces work significantly.

Example: X ~ B(15, 0.4). Find P(X ≥ 8).
Direct: P(X=8)+P(X=9)+…+P(X=15) — 8 terms
Complement: 1 − P(X ≤ 7) — one cumulative value
P(X ≤ 7) using binomcdf(15, 0.4, 7) ≈ 0.7869
P(X ≥ 8) = 1 − 0.7869 = 0.2131

Worked Calculator Example

Example: X ~ B(12, 0.35). Find (a) P(X = 4), (b) P(X ≤ 5), (c) P(X ≥ 4), (d) P(3 ≤ X ≤ 7).

(a) binompdf(12, 0.35, 4) = 0.2367 (4 s.f.)
(b) binomcdf(12, 0.35, 5) = 0.7873 (4 s.f.)
(c) 1 − binomcdf(12, 0.35, 3) = 1 − 0.3467 = 0.6533 (4 s.f.)
(d) binomcdf(12, 0.35, 7) − binomcdf(12, 0.35, 2) = 0.9764 − 0.1513 = 0.8251

p > 0.5: Using the Complementary Approach

If p > 0.5 and your tables only go up to p = 0.5, use the symmetry of the binomial:

If X ~ B(n, p) then (n − X) ~ B(n, 1−p)
So P(X = r) with parameter p = P(Y = n−r) with parameter 1−p, where Y = n−X.

Example: X ~ B(8, 0.7). Find P(X ≤ 3) using tables for p = 0.3.
P(X ≤ 3) = P(Y ≥ 5) where Y ~ B(8, 0.3)
= 1 − P(Y ≤ 4) = 1 − 0.9420 = 0.0580
In the exam, always state what function and parameters you use: "Using binomcdf(10, 0.3, 4)" earns the method mark. Never just write a decimal answer from a calculator without referencing the method.

Worked Examples

Eight fully worked examples covering all key binomial skills. Study the method, not just the answer.

Example 1 — Verifying Binomial Conditions

A survey asks 15 people whether they support a policy. Each person responds independently. Previous surveys show 60% support. Let X = number who support the policy. Verify that X ~ B(15, 0.6) and state any assumptions.

Step 1 — Fixed n: 15 people surveyed, so n = 15. ✓
Step 2 — Independence: We are told each person responds independently. ✓ (Assumption: this holds in practice.)
Step 3 — Two outcomes: Each person either supports (success) or does not (failure). ✓
Step 4 — Constant p: We assume each person has the same probability 0.6 of supporting, as given by historical data. ✓
Conclusion: All four conditions satisfied. X ~ B(15, 0.6). Assumption: the probability applies equally to all individuals in this sample.

Example 2 — P(X = 3) Exact Calculation

X ~ B(8, 0.25). Find P(X = 3) exactly.

Step 1: n = 8, p = 0.25, q = 0.75, r = 3.
Step 2: 8C3 = 8!/(3!5!) = (8×7×6)/(3×2×1) = 336/6 = 56
Step 3: p3 = (0.25)3 = 0.015625; q5 = (0.75)5 = 0.2373046875
Step 4: P(X=3) = 56 × 0.015625 × 0.2373046875 = 56 × 0.003707... = 0.2076 (4 s.f.)
Exact fraction = 56 × (1/4)3 × (3/4)5 = 56 × (35/48) = 56 × 243/65536 = 13608/65536 = 1701/8192

Example 3 — P(X ≤ 4) Cumulative Probability

X ~ B(10, 0.3). Find P(X ≤ 4) by summing individual probabilities.

Step 1: Calculate each P(X = k) for k = 0, 1, 2, 3, 4.
P(X=0) = (0.7)10 = 0.02825
P(X=1) = 10×(0.3)(0.7)9 = 10×0.3×0.04035 = 0.12106
P(X=2) = 45×(0.3)2×(0.7)8 = 45×0.09×0.05765 = 0.23347
P(X=3) = 120×(0.3)3×(0.7)7 = 120×0.027×0.08235 = 0.26683
P(X=4) = 210×(0.3)4×(0.7)6 = 210×0.0081×0.11765 = 0.20012
Step 2: Sum: P(X ≤ 4) = 0.02825 + 0.12106 + 0.23347 + 0.26683 + 0.20012 = 0.8497 (4 s.f.)
Check Using binomcdf(10, 0.3, 4) = 0.8497 ✓

Example 4 — P(X ≥ 5) Using Complement

X ~ B(10, 0.4). Find P(X ≥ 5).

Step 1: P(X ≥ 5) = 1 − P(X ≤ 4)   [complement: "at least 5" is complement of "at most 4"]
Step 2: Using binomcdf(10, 0.4, 4):
P(X ≤ 4) = P(X=0)+P(X=1)+P(X=2)+P(X=3)+P(X=4)
= 0.00605 + 0.04031 + 0.12093 + 0.21499 + 0.25082 = 0.63310 (5 s.f.)
Step 3: P(X ≥ 5) = 1 − 0.63310 = 0.3669 (4 s.f.)
Key point NEVER write P(X ≥ 5) = 1 − P(X ≤ 5). This omits P(X=5) from both sides. Always subtract P(X ≤ r−1).

Example 5 — Find n Given Probability Condition

X ~ B(n, 0.2). Find the smallest n such that P(X ≥ 1) > 0.95.

Step 1: P(X ≥ 1) = 1 − P(X = 0) = 1 − (0.8)n
Step 2: We need 1 − (0.8)n > 0.95   ⇒   (0.8)n < 0.05
Step 3: Take natural logs: n ln(0.8) < ln(0.05)   ⇒   n > ln(0.05)/ln(0.8)
ln(0.05) = −2.9957...,   ln(0.8) = −0.22314...
n > 2.9957/0.22314 = 13.43...
Step 4: Since n must be a whole number and n > 13.43, the smallest n is 14.
Check: P(X ≥ 1) = 1 − (0.8)14 = 1 − 0.04398 = 0.9560 > 0.95 ✓
n = 13: P(X ≥ 1) = 1 − (0.8)13 = 0.9450 < 0.95 ✗

Example 6 — Find p Given Probability

X ~ B(3, p). Given P(X = 0) = 0.125, find p.

Step 1: P(X=0) = (1−p)3 = 0.125
Step 2: Take cube roots: 1−p = (0.125)1/3 = 0.5
Step 3: p = 1 − 0.5 = 0.5
Check P(X=0) = (0.5)3 = 0.125 ✓

Example 7 — Mean and Variance

X ~ B(n, p) with E(X) = 4 and Var(X) = 3.2. Find n, p, and P(X = 4).

Step 1: q = Var(X)/E(X) = 3.2/4 = 0.8   ⇒   p = 0.2
Step 2: n = E(X)/p = 4/0.2 = 20
Step 3: X ~ B(20, 0.2). Find P(X=4):
P(X=4) = 20C4 × (0.2)4 × (0.8)16
= 4845 × 0.0016 × 0.02815... = 4845 × 0.00004504... = 0.2182 (4 s.f.)

Example 8 — Critical Region for Hypothesis Test (Intro)

Under H0, X ~ B(20, 0.3). For a one-tailed lower-tail test at 5% significance, find the critical region.

Step 1: Critical region is the set of small values of X where P(X ≤ c) < 0.05.
Step 2: Calculate cumulative probabilities:
P(X ≤ 0) = (0.7)20 = 0.0008 < 0.05 ✓
P(X ≤ 1) = 0.0008 + 20(0.3)(0.7)19 = 0.0008 + 0.0068 = 0.0076 < 0.05 ✓
P(X ≤ 2) ≈ 0.0355 < 0.05 ✓
P(X ≤ 3) ≈ 0.1071 > 0.05 ✗
Step 3: Critical region: X ≤ 2. The actual significance level is P(X ≤ 2) ≈ 0.0355 = 3.55%.
Note Cambridge S1 does not require formal hypothesis test procedures — just finding the critical region is sufficient at this stage.

Common Mistakes

These mistakes appear repeatedly in Cambridge S1 scripts. Learn them now and avoid losing marks.

Mistake 1 — Wrong Complement for P(X ≥ r)

WRONG: P(X ≥ 4) = 1 − P(X ≤ 4)
CORRECT: P(X ≥ 4) = 1 − P(X ≤ 3)   [subtract P of "at most r−1"]

The complement of "at least 4" is "fewer than 4" = "at most 3". Writing 1 − P(X ≤ 4) gives 1 minus the event that includes P(X=4), so you're leaving P(X=4) out from nowhere — and double-subtracting it.

Mistake 2 — Not Checking All Four Conditions

WRONG: "There are two outcomes, so it's binomial."
CORRECT: Check ALL four conditions — fixed n, independence, two outcomes, constant p.

Drawing without replacement from a small population typically fails the independence and/or constant p conditions. "Two outcomes" is necessary but not sufficient.

Mistake 3 — Wrong Variance Formula

WRONG: Var(X) = np   or   Var(X) = np2
CORRECT: Var(X) = np(1−p) = npq   where q = 1−p

A very common slip is copying E(X) = np and confusing it with the variance formula. Always write q = 1 − p explicitly first.

Mistake 4 — Treating Dependent Trials as Binomial

WRONG: "Draw 5 cards from a deck without replacement. X ~ B(5, 1/4) for hearts."
CORRECT: Binomial does not apply — trials are dependent (sampling without replacement). Use hypergeometric or direct counting.

Each card drawn changes the composition of the remaining deck, so p changes with every draw.

Mistake 5 — Forgetting nCr in the Formula

WRONG: P(X=3) = p3(1−p)n−3
CORRECT: P(X=3) = nC3 × p3 × (1−p)n−3

Without nCr, you only account for one specific arrangement of successes and failures, not all possible arrangements. This gives the probability of a specific sequence, not the total probability of r successes.

Mistake 6 — Using p Instead of q in the Exponent

WRONG: P(X=2) = nC2 × p2 × pn−2 = nC2 × pn
CORRECT: P(X=2) = nC2 × p2 × (1−p)n−2

Failures have probability q = 1−p, not p. Confusing p and q is extremely common when p is close to 0.5.

Mistake 7 — Wrong Inequality in Finding n or p

WRONG: P(X ≥ 1) > 0.95 ⇒ (0.8)n > 0.05 ⇒ n < 13.4 ⇒ n = 13
CORRECT: P(X ≥ 1) = 1 − (0.8)n > 0.95 ⇒ (0.8)n < 0.05 ⇒ n > 13.4 ⇒ n = 14

When taking logarithms of an inequality with a negative base (ln of a number less than 1 is negative), the inequality sign flips. Always check by substituting your answer back in.

Mistake 8 — Rounding Intermediate Values

WRONG: Rounding each P(X=k) to 2 d.p. before summing, then getting an incorrect total.
CORRECT: Keep at least 5 significant figures in intermediate steps, and round only the final answer.

Cumulative binomial probabilities require summing several small terms. Early rounding compounds errors and can cause a final answer that differs by 0.01 or more — losing accuracy marks.

Key Formulas — Binomial Distribution

Core Probability Formula

P(X = r) = nCr × pr × (1−p)n−r    r = 0, 1, 2, …, n
nCr = n! / [r!(n−r)!]

Mean and Variance

E(X) = np     Var(X) = np(1−p) = npq     SD(X) = √(npq)

Cumulative Probability Identities

P(X ≤ r) = Σk=0r P(X=k)
P(X ≥ r) = 1 − P(X ≤ r−1)
P(X > r) = 1 − P(X ≤ r)
P(X < r) = P(X ≤ r−1)
P(a ≤ X ≤ b) = P(X ≤ b) − P(X ≤ a−1)
P(X = r) = P(X ≤ r) − P(X ≤ r−1)

Special Values

P(X = 0) = qn = (1−p)n     P(X = n) = pn

Finding n and p from Moments

Given E(X) and Var(X):   q = Var(X)/E(X)  ⇒  p = 1−q  ⇒  n = E(X)/p

Full Formula Reference Table

QuantityFormulaNotes
NotationX ~ B(n, p)n trials, prob p of success
P(X = r)nCr pr qn-rq = 1−p
E(X)npMean / expected number of successes
Var(X)npqq = 1−p; always ≥ 0
SD(X)√(npq)Same units as X
P(X ≥ r)1 − P(X ≤ r−1)Complement with r−1 NOT r
P(X > r)1 − P(X ≤ r)Strict inequality
P(a≤X≤b)P(X≤b) − P(X≤a−1)Both endpoints included
Binomial coefficientnCr = n!/(r!(n−r)!)Also written C(n,r) or (n choose r)

Conditions Checklist (for exam questions)

When asked to justify a binomial model, state all four:
• Fixed number of trials: n = ___ (state the value)
• Trials are independent: (state the reason from context)
• Only two outcomes: success = ___, failure = ___
• Constant probability: p = ___ for each trial

Proof Bank

These proofs are not required for Cambridge S1 but deepen your understanding and support Further Mathematics. Each is written with full logical steps.

Proof 1 — P(X = r) from Combinatorial Argument

Claim: If X counts the number of successes in n independent Bernoulli trials each with probability p, then P(X = r) = nCr pr(1−p)n−r.


Proof:
Consider a specific sequence of n trial outcomes with exactly r successes and n−r failures, for example: SS…SFF…F (r S's followed by n−r F's).

By independence, the probability of this specific sequence is:
   P(SS…SFF…F) = p × p × … × p × (1−p) × … × (1−p) = pr(1−p)n−r

Any sequence with exactly r successes in n trials has this same probability pr(1−p)n−r (by independence and commutativity of multiplication).

The number of distinct sequences with exactly r successes in n trials equals the number of ways to choose which r positions (out of n) are successes = nCr.

Since all these sequences are mutually exclusive (they are distinct outcomes):
   P(X = r) = nCr × pr(1−p)n−r   □

Proof 2 — Deriving E(X) = np Using Indicator Variables

Claim: If X ~ B(n, p), then E(X) = np.


Proof (Method 1 — Indicator Variables):
Define indicator variables I1, I2, …, In where Ik = 1 if trial k is a success, 0 if it is a failure.

Then X = I1 + I2 + … + In (total number of successes).

For each k: E(Ik) = 1×P(Ik=1) + 0×P(Ik=0) = p.

By linearity of expectation (which holds regardless of independence):
   E(X) = E(I1) + E(I2) + … + E(In) = p + p + … + p = np   □

Proof (Method 2 — Direct Summation):
E(X) = Σr=0n r × nCr pr qn-r
The r = 0 term vanishes. For r ≥ 1: r×nCr = n × n−1Cr−1.
So E(X) = np Σr=1n n−1Cr−1 pr−1 qn−r = np Σj=0n−1 n−1Cj pj qn−1−j = np(p+q)n−1 = np × 1 = np   □

Proof 3 — Deriving Var(X) = npq

Claim: If X ~ B(n, p), then Var(X) = np(1−p).


Proof (Using Indicator Variables):
Using the same indicator variables Ik as above, X = ΣIk.

Each Ik is a Bernoulli(p) variable: Var(Ik) = E(Ik2) − [E(Ik)]2 = p − p2 = p(1−p) = pq.

Since I1, I2, …, In are independent (trials are independent), variances add:
   Var(X) = Var(I1) + Var(I2) + … + Var(In) = pq + pq + … + pq = npq   □

Alternative — using E(X2):
E(X(X−1)) = Σr=2n r(r−1)nCrprqn−r. Using r(r−1)nCr = n(n−1)n−2Cr−2:
E(X(X−1)) = n(n−1)p2(p+q)n−2 = n(n−1)p2
E(X2) = E(X(X−1)) + E(X) = n(n−1)p2 + np
Var(X) = E(X2) − (np)2 = n(n−1)p2 + np − n2p2 = np(1−p) = npq   □

Proof 4 — Σ P(X=r) = 1 (Probabilities Sum to 1)

Claim: The binomial probabilities sum to 1.


Proof:
By the Binomial Theorem: (p + q)n = Σr=0n nCr pr qn−r

Since q = 1−p: (p + q) = p + (1−p) = 1

Therefore: Σr=0n P(X=r) = Σr=0n nCr pr qn−r = (p+q)n = 1n = 1   □

Binomial Distribution Visualiser

Enter n (1–30) and p (0–1) to draw the binomial distribution. Enter r to highlight P(X=r).

Enter values above and click Draw Chart.

Exercise 1 — Binomial Conditions (10 Questions)

For numerical answers use 1=Yes binomial applies, 0=No. For probability answers give to 4 decimal places.

Exercise 2 — Calculate P(X = r) (10 Questions)

Give answers to 4 decimal places unless stated.

Exercise 3 — Mean & Variance (10 Questions)

Give exact decimal answers.

Exercise 4 — Cumulative Probabilities P(X ≤ r), P(X ≥ r) (10 Questions)

Give answers to 4 decimal places. Remember: P(X≥r) = 1 − P(X≤r−1).

Exercise 5 — Find n or p from Given Information (10 Questions)

Give n as a whole number and p to 4 decimal places unless stated.

Practice — 30 Mixed Questions

Mixed practice covering all binomial topics. Aim for 100% — confetti awaits!

Challenge — 15 Harder Questions

Harder questions involving multi-step reasoning, finding n and p, and critical regions.

Exam Style Questions (8 Questions)

These questions mirror Cambridge A-Level 9709 S1 style. Show all working. Click "Show Mark Scheme" to reveal the answer.

Question 1 [5 marks]

A biased coin has P(Head) = 0.35. The coin is tossed 12 times. X = number of heads.
(a) State the distribution of X.   [1]
(b) Find P(X = 4).   [2]
(c) Find P(X ≤ 3).   [2]

(a) X ~ B(12, 0.35) [B1]
(b) P(X=4) = 12C4 × (0.35)4 × (0.65)8 = 495 × 0.015006 × 0.031939 = 0.2367 [M1 A1]
(c) P(X≤3) = P(X=0)+P(X=1)+P(X=2)+P(X=3)
= 0.6512 + 12(0.35)(0.65)11 + 66(0.35)2(0.65)10 + 220(0.35)3(0.65)9
= 0.00569 + 0.03678 + 0.10876 + 0.19458 = 0.3458 (4 s.f.) [M1 A1]

Question 2 [5 marks]

X ~ B(n, 0.3). Given that E(X) = 7.5, find:
(a) n   [1]
(b) Var(X)   [1]
(c) P(X ≥ 8)   [3]

(a) E(X) = 0.3n = 7.5 ⇒ n = 25 [B1]
(b) Var(X) = 25 × 0.3 × 0.7 = 5.25 [B1]
(c) P(X ≥ 8) = 1 − P(X ≤ 7)
P(X ≤ 7) = binomcdf(25, 0.3, 7) = 0.5118 (4 s.f.)
P(X ≥ 8) = 1 − 0.5118 = 0.4882 [M1 M1 A1]

Question 3 [6 marks]

In a multiple-choice test there are 20 questions, each with 5 options. A student guesses every answer. X = number of correct answers.
(a) State the distribution of X and justify each condition.   [4]
(b) Find P(X ≥ 6).   [2]

(a) X ~ B(20, 0.2) [B1]
Fixed n = 20 questions [B1] | Each guess independent (student guesses randomly) [B1]
Two outcomes: correct/incorrect | p = 1/5 = 0.2 constant for each question (5 equally likely options) [B1]
(b) P(X ≥ 6) = 1 − P(X ≤ 5)
P(X ≤ 5) = binomcdf(20, 0.2, 5) = 0.8042 (4 s.f.)
P(X ≥ 6) = 1 − 0.8042 = 0.1958 (4 s.f.) [M1 A1]

Question 4 [6 marks]

X ~ B(n, p) with E(X) = 6 and Var(X) = 4.8.
(a) Find n and p.   [3]
(b) Find P(X ≥ 7).   [3]

(a) Var/E = q ⇒ q = 4.8/6 = 0.8 ⇒ p = 0.2 [M1 A1]
n = E(X)/p = 6/0.2 = 30 [A1]
(b) X ~ B(30, 0.2). P(X ≥ 7) = 1 − P(X ≤ 6)
P(X ≤ 6) = binomcdf(30, 0.2, 6) = 0.6065 (4 s.f.)
P(X ≥ 7) = 1 − 0.6065 = 0.3935 [M1 M1 A1]

Question 5 [5 marks]

Each day the probability that it rains in a particular city is 0.45, independently of other days. Find the probability that, in a week of 7 days:
(a) it rains on exactly 3 days   [2]
(b) it rains on more than 4 days   [3]

Let X ~ B(7, 0.45).
(a) P(X=3) = 7C3(0.45)3(0.55)4 = 35 × 0.091125 × 0.09151 = 0.2918 (4 s.f.) [M1 A1]
(b) P(X > 4) = P(X ≥ 5) = 1 − P(X ≤ 4)
P(X ≤ 4) = binomcdf(7, 0.45, 4) = 0.7667
P(X > 4) = 1 − 0.7667 = 0.2333 (4 s.f.) [M1 M1 A1]

Question 6 [5 marks]

A factory tests light bulbs. The probability a bulb is faulty is p. A sample of 8 bulbs is tested. The probability that none are faulty is 0.1678 (to 4 s.f.).
(a) Show that p = 0.2.   [2]
(b) Find the probability that at most 2 are faulty.   [3]

(a) P(X=0) = (1−p)8 = 0.1678
1−p = (0.1678)1/8 = 0.8001 ≈ 0.8 [M1]
p = 1 − 0.8 = 0.2 (shown) [A1]
(b) X ~ B(8, 0.2). P(X ≤ 2) = P(X=0) + P(X=1) + P(X=2)
= (0.8)8 + 8(0.2)(0.8)7 + 28(0.2)2(0.8)6
= 0.16777 + 0.33554 + 0.29360 = 0.7969 (4 s.f.) [M1 M1 A1]

Question 7 [6 marks]

X ~ B(15, 0.4).
(a) Find P(4 ≤ X ≤ 8).   [3]
(b) Find the smallest value of k such that P(X ≥ k) < 0.1.   [3]

(a) P(4 ≤ X ≤ 8) = P(X ≤ 8) − P(X ≤ 3) [M1]
P(X ≤ 8) = binomcdf(15, 0.4, 8) = 0.9050
P(X ≤ 3) = binomcdf(15, 0.4, 3) = 0.0905 [M1]
P(4 ≤ X ≤ 8) = 0.9050 − 0.0905 = 0.8145 [A1]
(b) Need P(X ≥ k) < 0.1 ⇒ 1 − P(X ≤ k−1) < 0.1 ⇒ P(X ≤ k−1) > 0.9 [M1]
P(X ≤ 9) = 0.9662 > 0.9 ⇒ k−1 = 9 works
P(X ≤ 8) = 0.9050 > 0.9 ⇒ k−1 = 8 also works ⇒ k = 9
Check: P(X ≤ 7) = 0.7869 < 0.9 ⇒ k = 8 gives P(X≥8) = 0.2131 > 0.1 ✗
Smallest k = 9 [M1 A1]

Question 8 [7 marks]

Under H0, X ~ B(20, 0.5). A one-tailed upper test is conducted at the 5% significance level.
(a) Find the critical region for X.   [3]
(b) State the actual significance level of the test.   [1]
(c) An observation of x = 14 is obtained. State the conclusion of the test.   [3]

(a) Need P(X ≥ c) < 0.05, i.e. 1 − P(X ≤ c−1) < 0.05 ⇒ P(X ≤ c−1) > 0.95 [M1]
P(X ≤ 13) = binomcdf(20, 0.5, 13) = 0.9423 < 0.95 ✗
P(X ≤ 14) = binomcdf(20, 0.5, 14) = 0.9793 > 0.95 ✓ [M1]
Critical region: X ≥ 15 [A1]
(b) Actual significance level = P(X ≥ 15) = 1 − 0.9793 = 2.07% [B1]
(c) x = 14 does not lie in the critical region (X ≥ 15). [M1]
There is insufficient evidence at the 5% level to reject H0. [M1]
We conclude that the data is consistent with X ~ B(20, 0.5). [A1]

Past Paper Questions

Representative Cambridge 9709 S1 past paper questions on the Binomial Distribution. Attempt fully before revealing the solution.

Past Paper Q1 — (Nov 2018 style) [6 marks]

The random variable X ~ B(12, p). Given that P(X = 0) = 0.05314 (to 4 s.f.), find:
(a) The value of p, to 2 decimal places.   [2]
(b) P(X ≤ 3).   [2]
(c) The mean and variance of X.   [2]

(a) P(X=0) = (1−p)12 = 0.05314
1−p = (0.05314)1/12 = 0.7500
p = 0.25
(b) X ~ B(12, 0.25).
P(X≤3) = (0.75)12 + 12(0.25)(0.75)11 + 66(0.25)2(0.75)10 + 220(0.25)3(0.75)9
= 0.03168 + 0.12671 + 0.23261 + 0.25811 = 0.6488 (4 s.f.)
(c) E(X) = 12 × 0.25 = 3; Var(X) = 12 × 0.25 × 0.75 = 2.25

Past Paper Q2 — (Jun 2019 style) [5 marks]

A quality control inspector examines items from a production line where 8% are defective. Items are selected independently. The inspector examines 25 items. Find the probability that:
(a) No items are defective.   [1]
(b) At least 3 are defective.   [3]
(c) The expected number of defective items.   [1]

X ~ B(25, 0.08).
(a) P(X=0) = (0.92)25 = 0.1244 (4 s.f.)
(b) P(X ≥ 3) = 1 − P(X ≤ 2)
P(X=1) = 25(0.08)(0.92)24 = 0.2704
P(X=2) = 300(0.08)2(0.92)23 = 0.2821
P(X ≤ 2) = 0.1244 + 0.2704 + 0.2821 = 0.6769
P(X ≥ 3) = 1 − 0.6769 = 0.3231 (4 s.f.)
(c) E(X) = 25 × 0.08 = 2

Past Paper Q3 — (Nov 2020 style) [7 marks]

Each trial in an experiment has probability p of success, independently. An experiment consists of 10 trials.
(a) Given Var(X) = 1.96, find the two possible values of p.   [3]
(b) Given also that E(X) > 5, state which value of p applies and find P(X ≥ 8).   [4]

(a) X ~ B(10, p). Var(X) = 10p(1−p) = 1.96
10p − 10p2 = 1.96 ⇒ 10p2 − 10p + 1.96 = 0
p = [10 ± √(100 − 78.4)] / 20 = [10 ± √21.6] / 20 = [10 ± 4.648] / 20
p = 0.7324 or p = 0.2676
(b) E(X) = 10p > 5 ⇒ p > 0.5 ⇒ p = 0.7324
X ~ B(10, 0.7324). P(X ≥ 8) = 1 − P(X ≤ 7)
P(X ≤ 7) = binomcdf(10, 0.7324, 7) ≈ 0.4439
P(X ≥ 8) = 1 − 0.4439 = 0.5561 (4 s.f.)

Past Paper Q4 — (Jun 2021 style) [6 marks]

X ~ B(20, 0.35).
(a) Find P(X = 7).   [2]
(b) Find P(5 ≤ X ≤ 9).   [3]
(c) Find the smallest n such that P(X ≥ n) < 0.05.   [1] (where X ~ B(20, 0.35))

(a) P(X=7) = 20C7(0.35)7(0.65)13
= 77520 × 0.0006434 × 0.01029 = 0.1844 (4 s.f.)
(b) P(5 ≤ X ≤ 9) = P(X ≤ 9) − P(X ≤ 4)
P(X ≤ 9) = binomcdf(20, 0.35, 9) = 0.8782
P(X ≤ 4) = binomcdf(20, 0.35, 4) = 0.1182
P(5 ≤ X ≤ 9) = 0.8782 − 0.1182 = 0.7600 (4 s.f.)
(c) Need P(X ≥ n) < 0.05 ⇒ P(X ≤ n−1) > 0.95.
P(X ≤ 12) = 0.9840 > 0.95; P(X ≤ 11) = 0.9468 > 0.95; P(X ≤ 10) = 0.9468... test: binomcdf(20,0.35,10) = 0.9468, binomcdf(20,0.35,9) = 0.8782 < 0.95
So n−1 = 10 ⇒ n = 11 (P(X≥11) = 1−0.9468 = 0.0532... try n=12: P(X≥12)=1−0.9840=0.0160<0.05) → n = 12

Past Paper Q5 — (Nov 2022 style) [8 marks]

A factory produces components; each is independently faulty with probability 0.05. Components are packed in boxes of 30.
(a) Find the probability that a box contains no faulty components.   [1]
(b) Find the probability that a box contains at most 2 faulty components.   [3]
(c) A company buys 10 boxes. Find the probability that exactly 3 boxes contain at most 2 faulty components.   [4]

X ~ B(30, 0.05) for number of faulty components per box.
(a) P(X=0) = (0.95)30 = 0.2146 (4 s.f.)
(b) P(X ≤ 2) = P(X=0)+P(X=1)+P(X=2)
P(X=1) = 30(0.05)(0.95)29 = 0.3389
P(X=2) = 435(0.05)2(0.95)28 = 0.2586
P(X ≤ 2) = 0.2146 + 0.3389 + 0.2586 = 0.8122 (4 s.f.)
(c) Let Y = number of boxes (out of 10) with at most 2 faulty. Y ~ B(10, 0.8122).
P(Y=3) = 10C3(0.8122)3(0.1878)7
= 120 × 0.5361 × 0.0000374 × ... let me use exact calc:
= 120 × (0.8122)3 × (0.1878)7
(0.8122)3 = 0.5360; (0.1878)7 = 0.000003764
P(Y=3) = 120 × 0.5360 × 0.000003764 = 0.000242 (3 s.f.)