HSC Study

The Binomial Distribution

Counting successes in n independent Bernoulli trials, the binomial probability formula and the meaning of nCr in it, expected frequencies, and the mean np and variance np(1 − p).

50 min · Statistical Analysis

Assumes you have read Bernoulli Random Variables and Permutations and Combinations.

The full teaching. Start here.

The short version

  • counts successes in independent trials, each with two outcomes and the same .
  • for .
  • counts which trials are the successes; is the probability of each such sequence.
  • , .
  • At least one: .
  • Expected frequency of an outcome in repetitions: its probability.

Counting successes

Repeat a Bernoulli trial times and count the successes. If the trials are independent and stays the same, the count has a binomial distribution.

Binomial random variable

The number of successes in independent Bernoulli trials, each with success probability . We write . Its possible values are .

Before using the model, check the conditions:

  1. a fixed number of trials;
  2. each trial has two outcomes;
  3. the trials are independent;
  4. the probability of success is the same for each trial.

Sampling without replacement breaks conditions 3 and 4, because each draw changes what is left. When the population is much larger than the sample, the change is tiny, and the binomial model is still a good approximation.

CheckpointAnswer before reading on.

Which is not a condition for the number of successes to be binomial?

Select one answer

Hint 1Fixed n, independent, same p, two outcomes.

Hint 2p can be anything.

Hint 3Look for the extra condition.

CheckpointAnswer before reading on.

Which random variable is binomial?

Select one answer

Hint 1Fixed n?

Hint 2Independent with constant p?

Hint 3Counting successes?

CheckpointAnswer before reading on.

Five cards are dealt from a deck without replacement. Why is the number of hearts not binomial?

Select one answer

Hint 1Is p constant?

Hint 2Deck changes.

Hint 3Dependence.

CheckpointAnswer before reading on.

What does mean?

Select one answer

Hint 1First parameter n.

Hint 2Second parameter p.

Hint 312 trials, p = 0.25.

CheckpointAnswer before reading on.

About 12% of people are left-handed. Thirty people are chosen at random and X counts the left-handers. What is p?

Select one answer

Hint 1Probability of success.

Hint 2Success = left-handed.

Hint 30.12.

The probability formula

Take and ask for exactly successes. One way is SSFF, with probability . Every other arrangement, such as SFSF or FFSS, has the same probability. There are arrangements, one for each choice of which trials succeed. So . In general:

These probabilities add to , because by the binomial theorem they are the terms of .

0.10.20.3012345678910mean 3X ~ Bin(10, 0.3)0.10.20.3012345678910mean 5X ~ Bin(10, 0.5)
Two binomial distributions with . With the distribution is skewed to the right and peaks at ; with it is symmetric about . In both, the peak is at or next to the mean .

Worked example4 marks

Binomial probabilities in context

A vaccine produces immunity in of people, independently. Twelve people are vaccinated. Find the probability that (a) exactly become immune; (b) at least become immune; (c) at least one does not become immune.

Model

Check the conditions and name the distribution.

Twelve independent trials, success = immune, . Let be the number immune: .

(a) Formula

Exactly 10.

.

(b) Add the cases

At least 11 means 11 or 12.

and . Total .

(c) Complement

At least one not immune is the opposite of all immune.

.

Answer

(a) . (b) . (c) .

Is that answer sensible?

The mean is , so values near should be the most likely. is the largest single probability in the problem, as expected.

CheckpointAnswer before reading on.

In , what does count?

Select one answer

Hint 1Orders of successes.

Hint 2Which trials succeed.

Hint 3Positions.

CheckpointAnswer before reading on.

. Find to 4 decimal places.

Hint 1.

Hint 2.

Hint 3About 0.267.

CheckpointAnswer before reading on.

Eight fair coins are tossed. Find the probability of exactly 5 heads.

Hint 1.

Hint 2.

Hint 30.21875.

CheckpointAnswer before reading on.

Five dice are rolled. Find the probability of exactly three sixes, to 4 decimal places.

Hint 1.

Hint 2.

Hint 3About 0.032.

CheckpointAnswer before reading on.

. Find .

Hint 1Complement.

Hint 2.

Hint 3.

CheckpointAnswer before reading on.

. Find to 4 decimal places.

Hint 1.

Hint 2.

Hint 3About 0.736.

CheckpointAnswer before reading on.

. What is the most likely value of ?

Hint 1Near the mean.

Hint 2Mean is 3.

Hint 3Compare P(2), P(3), P(4).

Exam question

Harder · about 3 min

2 marks

Six dice are rolled. Find the probability of at least two sixes, to 4 decimal places.

Hint 1.

Hint 2.

Hint 3.

Where students lose marks on this one
  • Subtracting only P(0).

    Why it happens: Answering "at least one".

    At least two excludes 0 and 1.

Written for this site.

Exam question

Standard · about 3 min

2 marks

A basketballer makes 70% of free throws, independently. Find the probability of making at least 4 of the next 5, to 4 decimal places.

Hint 1.

Hint 2.

Hint 3About 0.528.

Where students lose marks on this one
  • Only computing P(4).

    Why it happens: Missing P(5).

    At least 4 includes 5.

Written for this site.

Exam question

Harder · about 3 min

2 marks

A student guesses all 10 questions of a multiple-choice test, each with 4 options. Find the probability of at least 5 correct, to 4 decimal places.

Hint 1.

Hint 2Sum for to .

Hint 3About 0.078.

Where students lose marks on this one
  • Using p = 0.5.

    Why it happens: Assuming right or wrong are equally likely.

    Guessing among 4 options gives p = 0.25.

Written for this site.

Exam question

Standard · about 3 min

2 marks

Each item from a machine is faulty with probability 0.05, independently. In a box of 20, find the probability that (a) no item is faulty; (b) at least one is faulty. (2 marks)

Hint 1.

Hint 2.

Hint 3Complement.

Where students lose marks on this one
  • Computing 20 × 0.05 = 1 for (b).

    Why it happens: Using the mean as a probability.

    Use the complement.

Written for this site.

Exam question

Harder · about 3 min

2 marks

with , and . Find . (2 marks)

Hint 1.

Hint 2.

Hint 3.

Where students lose marks on this one
  • Forgetting the binomial coefficients are equal.

    Why it happens: Not justifying cancellation.

    Use symmetry of nCr.

Written for this site.

Exam question

Standard · about 3 min

2 marks

Show that . (2 marks)

Hint 1Binomial expansion.

Hint 2.

Hint 3.

Where students lose marks on this one
  • Checking one example.

    Why it happens: Not proving generally.

    Use (a + b)ⁿ with a = p, b = 1 − p.

Written for this site.

Common mistake

P(exactly 2 sixes in 5 rolls) = (1/6)²(5/6)³.

Why it happens: That is the probability of one particular order, such as SSFFF.

There are orders for two sixes among five rolls, each with the same probability. So .

Change and below and watch the distribution. It is symmetric when , skewed when is near or , and the tallest bar always sits close to the mean .

The binomial distributionThe binomial distribution with n = 10 and p = 0.30. The mean is 3.00 and the standard deviation is 1.449.mean np = 3.0001234567891000.050.10.150.20.250.3number of successes, rP(X = r)
Push p towards 0 or 1 and the bars pile up against one end. Keep p near 0.5 and raise n, and the shape becomes a symmetric bell centred on np.
The numbers behind this graph
rP(X = r)
00.0282
10.1211
20.2335
30.2668
40.2001
50.1029
60.0368
70.0090
80.0014
90.0001
100.0000
10
0.30
E(X) = np
3.00
Var(X) = np(1 − p)
2.100
Standard deviation
1.449
P(X = 0) = (1 − p)ⁿ
0.0282

Every slider is a normal range input, so the arrow keys move it one step and Home and End jump to the extremes.

How many trials?

To find the least that makes "at least one success" likely enough, solve with logarithms. Dividing by , which is negative, reverses the inequality.

Exam question

Harder · about 4 min

3 marks

Each trial succeeds with probability 0.1. What is the least number of trials needed for the probability of at least one success to exceed 0.95? (3 marks)

Hint 1.

Hint 2.

Hint 3Logarithms; the inequality flips.

Where students lose marks on this one
  • Answering 28.

    Why it happens: Rounding down or not flipping the inequality.

    Dividing by a negative log flips the inequality.

Written for this site.

Mean, variance and expected frequencies

A binomial variable is a sum of independent Bernoulli variables, . Means always add, and variances add for independent variables, so:

If an experiment is repeated times, the expected frequency of an outcome is times its probability.

Worked example4 marks

Expected frequencies and parameters

(a) Four seeds, each germinating with probability , are planted in each of pots. Find the expected number of pots with seeds germinating. (b) A binomial variable has mean and standard deviation . Find and .

(a) Probabilities

X ~ Bin(4, 0.5) per pot.

: .

(a) Frequencies

Multiply by 64.

.

(b) Use both formulas

np = 4 and np(1 − p) = 3.

Dividing: , so and .

Answer

(a) . (b) , .

Is that answer sensible?

(a) The frequencies sum to , and their mean is . Correct.

CheckpointAnswer before reading on.

. Find .

Hint 1.

Hint 2.

Hint 36.

CheckpointAnswer before reading on.

. Find .

Hint 1.

Hint 2.

Hint 35.1.

CheckpointAnswer before reading on.

. Find the standard deviation of .

Hint 1Variance .

Hint 225.

Hint 3Square root.

CheckpointAnswer before reading on.

A student guesses all 10 questions, each with 4 options. What is the expected number correct?

Hint 1.

Hint 2.

Hint 32.5.

Exam question

Standard · about 3 min

2 marks

A binomial random variable has mean 6 and variance 4.2. Find n and p. (2 marks)

Hint 1, .

Hint 2Divide: .

Hint 3.

Where students lose marks on this one
  • Solving for p from the variance alone.

    Why it happens: Using one equation.

    Divide variance by mean.

Written for this site.

Exam question

Standard · about 3 min

2 marks

Three fair coins are tossed 80 times, and the number of heads is recorded each time. Find the expected frequency of 0, 1, 2 and 3 heads. (2 marks)

Hint 1Probabilities .

Hint 2Multiply by 80.

Hint 310, 30, 30, 10.

Where students lose marks on this one
  • Using 1/4 for each.

    Why it happens: Treating outcomes as equally likely.

    Counts of heads are not equally likely.

Written for this site.

Exam question

Harder · about 3 min

2 marks

Explain why has and . (2 marks)

Hint 1Sum of n Bernoulli variables.

Hint 2Means add.

Hint 3Variances add for independent variables.

Where students lose marks on this one
  • Not mentioning independence for the variance.

    Why it happens: Adding variances without justification.

    Variances add only for independent variables.

Written for this site.

What lifts this to a Band 6

Open every binomial answer with a sentence of the form "Let be the number of … Then with , ." Markers often award a mark for the model itself. It also forces you to check that the trials are independent with constant , which is where "without replacement" traps are caught.

Where the peak is

Comparing neighbouring probabilities shows where the distribution peaks:

This ratio is greater than exactly when . So the probabilities increase while is below and decrease after, and the most likely value is the integer part of . For , , so the peak is at . When is a whole number, two values tie: for , , and .