Discrete random variables

S1 Topic 5, spec point by spec point (Discrete random variables). Each spec point has the method, worked examples written the way the mark scheme wants them, how the marks go and the mistakes that cost the most marks.

Spec5.1Discrete random variables

Understand the concept of a discrete random variable.

The idea

A random variable \(X\) is a quantity whose value depends on chance. It's discrete if it takes separate values (usually whole numbers: a score, a count). Its probability distribution lists each value with its probability, and the probabilities add to 1.

Capital \(X\) is the random variable; lower-case \(x\) is a particular value.

building a distribution

A coin with \(P(\text{head}) = 0.4\) is spun until a head appears, up to a maximum of 4 spins. \(X\) is the number of spins. Find the distribution.

\(P(X = 1) = 0.4\), \(P(X = 2) = 0.6 \times 0.4 = 0.24\), \(P(X = 3) = 0.6^2 \times 0.4 = 0.144\).

\(P(X = 4) = 0.6^3 = 0.216\) (three tails means the fourth spin happens, whatever it shows).

Check: \(0.4 + 0.24 + 0.144 + 0.216 = 1\).

  • B1 cso for showing a probability with the full product visible (\(0.6^2 \times 0.4\); "\(0.6 \times 0.24\)" alone is not enough) (WST01 Jun 2017 Q6a).
  • Events from inequalities: e.g. \(Y = 5 - 3X > 0\) means \(X < \frac53\); then add the relevant probabilities (WST01 Jun 2014 Q5f).

Often asked as: build a distribution from a game or experiment; then the mean and variance (5.3).

Forgetting the "last" case (here \(P(X = 4)\) includes four tails).

Spec5.2Probability functions and cumulative distribution functions

Use the probability function \(p(x) = P(X = x)\) and the cumulative distribution function \(F(x_0) = P(X \leqslant x_0) = \sum_{x \leqslant x_0} p(x)\).

The idea

To find an unknown constant, use "probabilities add to 1" (and any other given condition, such as the mean).

finding k and using F

\(P(X = x) = kx\) for \(x = 1, 2, 3, 4\). Find \(k\), \(F(3)\) and \(P(X > 1)\).

\(k(1 + 2 + 3 + 4) = 1\), so \(k = \frac{1}{10}\).

\(F(3) = \frac{1 + 2 + 3}{10} = 0.6\). \(P(X > 1) = 1 - F(1) = 0.9\).

from F to p

\(F(1) = 0.1\), \(F(2) = 0.2\), \(F(3) = 0.3\), \(F(4) = 0.5\), \(F(5) = 0.7\), \(F(6) = 1\).

\(P(X = x)\): 0.1, 0.1, 0.1, 0.2, 0.2, 0.3 for \(x = 1, \ldots, 6\).

  • B1 for reading \(F(x)\) (WST01A Jan 2026 Q1a); B1, M1, A1 for a distribution from \(F\) by differences (WST01 Oct 2024 Q6b).
  • Distributions in terms of \(p\): B1 for the right values of the variable, M1 for most probabilities, A1 (WST01 Jan 2020 Q6e).

Often asked as: find constants; write down \(F(x)\); probability tables from context.

\(P(X < 3) = F(3)\) (it's \(F(2)\) for whole-number values).

Spec5.3Mean and variance of a discrete random variable

Find \(E(X)\) and \(\text{Var}(X)\) using \(E(X)\) and \(E(X^2)\), and use \(E(aX + b) = aE(X) + b\) and \(\text{Var}(aX + b) = a^2\text{Var}(X)\).

The formulas

\[E(X) = \sum x\,P(X = x) \qquad E(X^2) = \sum x^2\,P(X = x) \qquad \text{Var}(X) = E(X^2) - [E(X)]^2\]

\(E(X)\) is the long-run average value. Don't divide by the number of values: the probabilities already do the averaging.

Linear functions

\[E(aX + b) = aE(X) + b \qquad \text{Var}(aX + b) = a^2\text{Var}(X)\]

Adding \(b\) shifts the values but doesn't spread them, so it doesn't change the variance. Multiplying by \(a\) multiplies the variance by \(a^2\) (even if \(a\) is negative).

mean, variance and a linear function

For the coin distribution in 5.1, find \(E(X)\), \(\text{Var}(X)\), and the mean and variance of \(Y = 5 - 8X\).

\(E(X) = 1(0.4) + 2(0.24) + 3(0.144) + 4(0.216) = 2.176\).

\(E(X^2) = 1(0.4) + 4(0.24) + 9(0.144) + 16(0.216) = 6.112\).

\(\text{Var}(X) = 6.112 - 2.176^2 = 1.377\).

\(E(Y) = 5 - 8(2.176) = -12.4\). \(\text{Var}(Y) = 64 \times 1.377 = 88.1\).

  • M1 for \(E(X)\) with at least three correct products (dividing by \(n\) is M0), A1 (WST01A Jan 2026 Q1b).
  • Variance: M1 for \(E(X^2)\), M1 for \(E(X^2) - [E(X)]^2\), A1 (WST01 Jun 2017 Q6d).
  • Linear functions: B1 for \(E(Y)\), M1 for \(a^2\text{Var}(X)\), A1 (WST01 Jun 2021 Q5c).

Often asked as: the most examined S1 point (104 questions). Find unknown probabilities from a given mean; \(E(X^2)\) "show that"; variance; linear functions; set \(E(Y)\) equal to a given value to find a constant.

\(\text{Var}(5 - 8X) = -8\text{Var}(X)\) or \(5 - 64\text{Var}(X)\).

\(E(X^2) = [E(X)]^2\).

Spec5.4The discrete uniform distribution

Use the discrete uniform distribution, with its mean and variance.

The idea

Every value \(1, 2, \ldots, n\) is equally likely: \(P(X = x) = \frac1n\). Example: the score on a fair die.

\[E(X) = \frac{n + 1}{2} \qquad \text{Var}(X) = \frac{n^2 - 1}{12}\]

(These are for values \(1\) to \(n\). For other equally spaced values, use a linear function: if \(Y = 2X + 3\), then \(E(Y) = 2E(X) + 3\).)

a die and an inequality

(a) A fair six-sided die: \(E(X) = 3.5\), \(\text{Var}(X) = \frac{35}{12}\).

(b) \(X\) is uniform on \(1, \ldots, 5\). Find \(P(X + 2 < 3X - 4)\).

\(6 < 2X\), so \(X > 3\): \(X = 4\) or \(5\). Probability \(\frac25\).

  • M1 for rearranging to the right inequality, A1. A probability for the wrong inequality scores 0 (WST01 Jan 2026 Q6a).
  • Name the distribution: "(discrete) uniform"; "continuous uniform" is wrong (WST01 Oct 2017 Q6a).

Often asked as: name the distribution; probabilities of ranges like \(P(2 \leqslant W < 3.5)\) (WST01 Jan 2022 Q4a); mean and variance.

Using the continuous uniform formulas.