Lesson 3 of 8 · 16 min

Expected value

Calculate the mean of a discrete random variable as a probability-weighted average, interpret it as a long-run average, and use it to judge games and decisions.

What you will be able to do

  • Calculate E⁡(X)=μ=∑pixi\E(X) = \mu = \sum p_i x_i from a probability function.
  • Interpret E⁡(X)\E(X) as the long-run average of XX over many repetitions.
  • Use E⁡(X)\E(X) to find unknown probabilities and to decide whether a game is fair.

A school fete runs a game: roll a fair die, win $10 for a six, $2 for a four or a five, and nothing otherwise. Over one roll you can't predict what you'll win. Over 600 rolls you can: about 100 sixes, 200 fours or fives and 300 others, so the total paid out is about

100×10+200×2+300×0=1400 dollars,100 \times 10 + 200 \times 2 + 300 \times 0 = 1400 \text{ dollars},

or 1400600≈$2.33\frac{1400}{600} \approx \$2.33 per roll. That long-run average is the expected value.

Divide each count by 600 and the calculation becomes "each value times its probability, added up":

10×16+2×26+0×36=146=73≈2.33.10 \times \frac{1}{6} + 2 \times \frac{2}{6} + 0 \times \frac{3}{6} = \frac{14}{6} = \frac{7}{3} \approx 2.33.

Key idea

Expected value

The mean or expected value of a discrete random variable XX is

E⁡(X)=μ=∑pixi,\E(X) = \mu = \sum p_i x_i,

where pip_i is the probability of the value xix_i. It is a weighted average: each value counts in proportion to how likely it is.

Interactive

The balance point of a distribution

If the bars were weights on a see-saw, the mean is where it would balance. Find the marked mean, then check it with ∑pixi\sum p_i x_i.

Worked example

Calculate an expected value from a table

  • Simple familiar
  • Technology free

The number of goals, XX, that a netball goal shooter scores in a quarter has this probability function. Calculate E⁡(X)\E(X) and interpret it.

xx01234
P(X=x)P(X = x)0.10.30.350.150.1
4 steps

Try it yourself first, then check the working one step at a time.

Method

  1. Check that the probabilities add to 1.

  2. Multiply each value by its probability.

  3. Add the products, and interpret the result as a long-run average.

Watch out for

  • Averaging the values 0 to 4 to get 2, ignoring the probabilities.

  • Rounding the mean to a whole number.

Let XX be the score on a fair six-sided die. Find E⁡(X)\E(X).

Reads asYour answer shows here as maths.

Note

The mean need not be a possible value

You can never roll 3.5, and no family has 2.4 children. E⁡(X)\E(X) describes the centre of the distribution, the average over many repetitions, not a value you expect on any one trial. Leave it as an exact fraction or decimal; don't round it to a whole number.

For a uniform random variable the weights are equal, so the mean is the ordinary average of the values, which sits in the middle by symmetry. For a non-uniform one you must weight.

The random variable XX has this probability function, and E⁡(X)=2.1\E(X) = 2.1. Find aa and bb.

xx123
P(X=x)P(X = x)0.3aabb

Expected value is how games, insurance and investments are judged. If GG is a player's gain (winnings minus the cost of playing), then E⁡(G)\E(G) is the average gain per game. The game is fair when E⁡(G)=0\E(G) = 0; when E⁡(G)<0\E(G) < 0 the player loses money in the long run and the organiser profits.

Worked example

Is the fete game fair?

  • Complex familiar
  • Technology free

At a school fete it costs $3 to roll a fair die. A six wins $10, a four or a five wins $2, and any other score wins nothing. (The $3 is not returned.)

(a) Let GG be the player's gain in dollars (prize minus cost). Determine the probability function of GG.

(b) Determine E⁡(G)\E(G) and interpret it.

(c) Determine the cost to play that would make the game fair.

5 steps

Try it yourself first, then check the working one step at a time.

Method

  1. Define the gain as prize minus cost, and list its values.

  2. Find the probability of each value.

  3. Calculate E⁡(G)=∑pigi\E(G) = \sum p_i g_i.

  4. Fair means E⁡(G)=0\E(G) = 0: the cost equals the expected prize.

Watch out for

  • Using the prizes instead of the gains, then forgetting to subtract the cost.

  • Subtracting the cost only from the winning outcomes.

Common mistake

Averaging the values

The plain average of the values, 0+1+2+3+45=2\frac{0 + 1 + 2 + 3 + 4}{5} = 2, is not E⁡(X)\E(X) unless the values are equally likely. Always multiply each value by its probability.

Extension: scaling and shifting a random variable

If every value of XX is multiplied by aa and then has bb added, every value in the weighted average changes the same way, so

E⁡(aX+b)=a E⁡(X)+b.\E(aX + b) = a\,\E(X) + b.

For example, if a stall sells XX drinks an hour with E⁡(X)=12\E(X) = 12, earns $4 a drink and pays $20 an hour to run, its hourly profit is 4X−204X - 20, with mean 4×12−20=$284 \times 12 - 20 = \$28. You can always get the same answer by writing out the probability function of the new variable and using ∑pixi\sum p_i x_i.

Exam tip

Show the sum

Write the products before you add them, e.g. E⁡(X)=0(0.1)+1(0.3)+2(0.35)+3(0.15)+4(0.1)=1.85\E(X) = 0(0.1) + 1(0.3) + 2(0.35) + 3(0.15) + 4(0.1) = 1.85. A marker can award the method even if one product is wrong.

Summary

  • E⁡(X)=μ=∑pixi\E(X) = \mu = \sum p_i x_i: multiply each value by its probability and add.

  • E⁡(X)\E(X) is the long-run average of XX and the balance point of its graph. It need not be a possible value.

  • For a uniform random variable, E⁡(X)\E(X) is the ordinary average of the values.

  • With unknowns, use ∑pi=1\sum p_i = 1 and the given mean as two equations.

  • A game is fair when the expected gain is 0.

Finished the lesson?

Try the 2 quick checks above and the lesson completes itself, or mark it complete now.

Lesson 3 of 80% read

Save your progress with a free account