Expected Value — I can compute expected value in a simple discrete context and interpret it as a long-run average. - Math 3

SECTION 1

Historical Context & Motivation

Humans have been gambling for thousands of years, but it wasn't until the 1600s that mathematicians began to develop a formal way to measure whether a game was fair. The central question was surprisingly simple: if you play a game over and over again, how much should you expect to win (or lose) on average? This question gave birth to one of the most powerful ideas in all of probability — the concept of expected value.

The story begins with a famous puzzle known as the Problem of Points. Two players are in the middle of a game of chance when it gets interrupted. How should they divide the prize money based on each player's probability of winning? This seemingly simple question stumped thinkers for decades, until a legendary exchange of letters cracked it open.

1494 • Pacioli Poses the Problem

Italian mathematician Luca Pacioli publishes the Problem of Points, asking how to fairly divide stakes in an interrupted game. His proposed solution was flawed, but the problem itself sparked centuries of debate.

1654 • Pascal and Fermat Correspond

Blaise Pascal and Pierre de Fermat exchange a series of letters that solve the Problem of Points. Their work lays the foundation for probability theory and introduces the idea of weighting outcomes by how likely they are.

1657 • Huygens Formalizes Expected Value

Dutch scientist Christiaan Huygens publishes the first formal treatment of expected value in his treatise on games of chance, calling it the "value of the expectation." This marks the birth of expected value as a mathematical tool.

1713 • Bernoulli's Law of Large Numbers

Jacob Bernoulli proves that as you repeat an experiment many times, the average result converges to the expected value. This theorem gives expected value its powerful interpretation as a long-run average.

Today, expected value is used far beyond gambling. Insurance companies use it to set premiums, businesses use it to evaluate risky investments, and scientists use it to model everything from weather patterns to genetics. The core question remains the same: if a random process repeats many times, what single number best predicts the average outcome?

SECTION 2

Core Principles & Definitions

Before diving into calculations, you need a few foundational ideas. Expected value builds on concepts you already know from basic probability — outcomes, probabilities, and the idea that some outcomes happen more often than others. The magic of expected value is that it combines all of these into a single, meaningful number.

Discrete Random Variable

A discrete random variable is a variable whose value is determined by the outcome of a random process. It can only take on a countable number of distinct values — like the result of rolling a die (1, 2, 3, 4, 5, or 6).

Probability Distribution

A probability distribution is a table or function that assigns a probability to each possible value of the random variable. All probabilities must be between 0 and 1, and they must add up to exactly 1.

Weighted Average

Expected value is a weighted average — not all outcomes contribute equally. Outcomes that are more probable carry more weight in the calculation, just like a heavily weighted test category affects your grade more.

Long-Run Average

Expected value is not a prediction for a single trial. Instead, it tells you the long-run average — the value the average result approaches if you repeat the experiment thousands of times.

✦

KEY TAKEAWAY

Think of expected value like your GPA. Your GPA isn't the grade you got in any single class — it's a weighted average that accounts for how many credits each class is worth. In the same way, expected value weights each possible outcome by its probability. A rare jackpot gets less weight than a common small payout, just like a 1-credit elective has less influence on your GPA than a 4-credit core course.

SECTION 3

Visualizing Expected Value

One of the best ways to understand expected value is to see it as the balance point of a probability distribution. Imagine placing weights on a number line — heavier weights at more probable outcomes. The expected value is the point where the number line would perfectly balance, like a seesaw's fulcrum.

This bar chart shows a symmetric probability distribution for outcomes 1 through 5. The height of each bar represents the probability of that outcome. The green dashed line marks the expected value E(X) = 3.0 — the exact balance point. Because this distribution is symmetric around 3, the expected value falls right in the center. Notice that outcome 3, with the highest probability (0.40), pulls the balance point toward it most strongly.

In the diagram above, the distribution is perfectly symmetric, so the expected value lands right at the center. But what happens when the distribution is skewed? If the probability of outcome 5 were much higher than the others, the balance point would shift to the right. The expected value always "slides" toward outcomes that are more probable — they have more weight pulling the fulcrum in their direction.

SECTION 4

The Mathematical Framework

Now that you have the intuition, let's translate it into a formula. Computing expected value is straightforward: multiply each outcome by its probability, then add all the products together. The result is a single number that represents the theoretical long-run average.

EXPECTED VALUE FORMULA

E(X) = x₁ · P(x₁) + x₂ · P(x₂) + … + xₙ · P(xₙ)

Where x₁, x₂, …, xₙ are the possible outcomes and P(x₁), P(x₂), …, P(xₙ) are their corresponding probabilities. The Greek letter sigma (Σ) is often used to write this more compactly.

COMPACT NOTATION

E(X) = Σ xᵢ · P(xᵢ)

This notation means "sum up every product of an outcome times its probability." The subscript i runs from 1 through n, covering every possible outcome.

VALIDITY CHECK

P(x₁) + P(x₂) + … + P(xₙ) = 1

Before computing E(X), always verify that all probabilities sum to 1. If they don't, the probability distribution is incomplete or contains an error.

⚠

IMPORTANT NOTE

The expected value does not have to be one of the possible outcomes. For example, the expected value of rolling a fair six-sided die is 3.5, even though you can never actually roll a 3.5. This makes sense because E(X) represents a theoretical average, not an actual result.

SECTION 5

Step-by-Step Breakdown

Let's walk through how to set up and compute expected value using a concrete scenario. Imagine a carnival game where a spinner lands on different dollar amounts. The following diagram shows the computation process from start to finish.

This flowchart shows the three-step process for computing expected value. Step 1 (amber): list all outcomes. Step 2 (violet): assign probabilities and verify they sum to 1. Step 3 (emerald): multiply each outcome by its probability and sum the products. The final expected value of $3.00 means that over many spins, you'd average $3.00 per spin.

Outcome (x)	Probability P(x)	x · P(x)
$0	0.50	$0.00
$1	0.25	$0.25
$5	0.15	$0.75
$20	0.10	$2.00
Total	1.00	E(X) = $3.00

Expected value calculation table for the carnival spinner

Notice how the $20 outcome, despite being the least likely at only 10%, contributes $2.00 to the expected value — the largest single contribution. This illustrates a key idea: even rare outcomes can significantly influence the expected value if their payoff is large enough. Meanwhile, the most common outcome ($0 with 50% probability) contributes nothing because its value is zero.

SECTION 6

Worked Example — Is This Game Fair?

A local fundraiser offers a dice game. You pay $4 to play. You roll a standard six-sided die, and the payout depends on the result: roll a 6 and win $15, roll a 4 or 5 and win $6, roll anything else and win $0. Is this game worth playing? Let's find the expected value of your net gain (winnings minus the cost to play).

Dice Game — Expected Net Gain

1
Step 1 — Identify All Outcomes and Payouts

There are three distinct payout groups. Rolling a 1, 2, or 3 pays $0. Rolling a 4 or 5 pays $6. Rolling a 6 pays $15. Since it costs $4 to play, the net gains are: $0 − $4 = −$4, $6 − $4 = $2, and $15 − $4 = $11.

Net gains: −$4, +$2, +$11
2
Step 2 — Assign Probabilities

A fair die has 6 equally likely outcomes. The probability of rolling 1, 2, or 3 is 3/6 = 1/2. The probability of rolling 4 or 5 is 2/6 = 1/3. The probability of rolling 6 is 1/6. Let's verify: 1/2 + 1/3 + 1/6 = 3/6 + 2/6 + 1/6 = 6/6 = 1. ✓

P(−$4) = 1/2, P(+$2) = 1/3, P(+$11) = 1/6
3
Step 3 — Compute Each Product

Multiply each net gain by its probability. First product: (−$4) × (1/2) = −$2.00. Second product: ($2) × (1/3) ≈ $0.67. Third product: ($11) × (1/6) ≈ $1.83.

Products: −$2.00, +$0.67, +$1.83
4
Step 4 — Sum the Products

Add all three products: E(X) = −$2.00 + $0.67 + $1.83 = $0.50.

E(X) = +$0.50
5
Step 5 — Interpret the Result

The expected net gain is +$0.50 per game. This means that if you played this game many, many times, you would average a profit of 50 cents per game. Since the expected value is positive, this game is in your favor — it is worth playing from a purely mathematical standpoint. A fair game would have E(X) = 0, meaning neither side has an advantage.

The game favors the player in the long run.

SECTION 7

Strengths & Limitations of Expected Value

Expected value is an incredibly useful tool, but like any mathematical concept, it has both strengths and limitations. Understanding these will help you know when to rely on expected value and when you need additional information to make good decisions.

Strengths	Limitations
Reduces a complex probability distribution to a single, easy-to-interpret number.	Tells you nothing about variability — two distributions can have the same E(X) but very different spreads.
Enables direct comparison between different options (e.g., which game is more profitable).	Only meaningful for repeated trials. For a one-time decision, the long-run average may not reflect your actual experience.
Guaranteed to match the actual average as trials increase (Law of Large Numbers).	Can be misleading if the distribution is heavily skewed by extreme outliers (e.g., lottery jackpots).
Foundation for more advanced concepts like variance, standard deviation, and decision theory.	Doesn't account for human risk preferences — losing $100 often "feels" worse than winning $100 feels good.

✦

KEY TAKEAWAY

Think of expected value like a GPS estimated arrival time. It gives you the best single prediction for your trip based on average conditions. But on any given day, you might arrive earlier or later due to traffic, weather, or construction. The estimate is most reliable when you make the same trip many times — just like expected value is most meaningful when a random process repeats.

SECTION 8

Connection to Advanced Concepts

Expected value is the starting point for a rich family of statistical concepts. Once you master E(X), you'll be ready to tackle tools that measure not just the center of a distribution, but its shape and spread. Here's a preview of how expected value connects to what you'll learn next.

Concept	What It Measures	How It Relates to E(X)
Variance	How spread out outcomes are from the expected value	Computed as E[(X − E(X))²] — the expected value of squared deviations
Standard Deviation	Average distance from the mean, in the original units	Square root of variance — makes spread interpretable
Law of Large Numbers	The theoretical guarantee behind long-run averages	Proves that sample averages converge to E(X) as trials increase
Continuous E(X)	Expected value when outcomes form a continuous range	Uses integration instead of summation — requires calculus

In your current course, you'll work with discrete distributions where the outcomes are countable and the probabilities sum to 1. Later in AP Statistics or college-level courses, you'll extend expected value to continuous distributions using integral calculus. The good news is that the core intuition — weighting outcomes by how likely they are — stays exactly the same.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

The expected value of rolling a fair six-sided die is 3.5. Does this mean you should expect to roll a 3 or a 4 most often? Explain what E(X) = 3.5 actually tells you.

Show answer

No — E(X) = 3.5 does not mean you'll roll a 3 or 4 more often. Every face (1 through 6) is equally likely with probability 1/6. The expected value of 3.5 is the long-run average: if you roll the die thousands of times, the mean of all your results will approach 3.5. It is not a prediction for any single roll, and it doesn't even have to be a possible outcome. The formula confirms this: E(X) = (1 + 2 + 3 + 4 + 5 + 6)/6 = 21/6 = 3.5.

Question 2: BASIC CALCULATION

A bag contains 3 red marbles worth 2 points, 5 blue marbles worth 4 points, and 2 gold marbles worth 10 points. You draw one marble at random. What is the expected point value of your draw?

Show answer

There are 3 + 5 + 2 = 10 marbles total. The probabilities are P(red) = 3/10, P(blue) = 5/10, and P(gold) = 2/10. Applying the expected value formula: E(X) = 2 × (3/10) + 4 × (5/10) + 10 × (2/10) = 6/10 + 20/10 + 20/10 = 46/10 = 4.6 points. On average, you'd expect to earn 4.6 points per draw over many repetitions. Notice that 4.6 is not any marble's actual point value — it's a weighted average reflecting the mix of marbles in the bag.

Question 3: INTERMEDIATE

A raffle sells 200 tickets at $5 each. There is one grand prize of $400, three second prizes of $50 each, and five third prizes of $10 each. You buy one ticket. Calculate the expected net gain and determine whether the raffle is favorable for you.

Show answer

First, find the net gains (prize minus $5 ticket cost): Grand prize net gain = $400 − $5 = $395 with P = 1/200. Second prize net gain = $50 − $5 = $45 with P = 3/200. Third prize net gain = $10 − $5 = $5 with P = 5/200. No prize net gain = −$5 with P = 191/200. Now compute E(X) = $395 × (1/200) + $45 × (3/200) + $5 × (5/200) + (−$5) × (191/200) = $1.975 + $0.675 + $0.125 − $4.775 = −$2.00. The expected net gain is −$2.00, meaning the raffle is not favorable for you. On average, you lose $2 per ticket. This makes sense because the total prize pool ($400 + $150 + $50 = $600) is less than the total revenue ($1,000), with the $400 difference going to the organizer.

Question 4: APPLIED

A delivery company knows that 2% of its packages are damaged in transit. Each damaged package costs the company $150 in refunds and reshipping. The company can purchase insurance that costs $4 per package and covers the full $150 damage cost. Using expected value, advise the company on whether insurance is a good deal.

Show answer

Without insurance, the expected cost per package is: E(cost) = $150 × 0.02 + $0 × 0.98 = $3.00. With insurance, the cost is a flat $4.00 per package regardless of damage. Since the expected cost without insurance ($3.00) is less than the insurance premium ($4.00), the company would save money in the long run by not purchasing insurance, from a pure expected-value perspective. However, this analysis only considers the average — it doesn't account for the risk of an unusually bad month with many damaged packages. A company that cannot absorb occasional large losses might rationally choose insurance despite the higher expected cost, because it provides predictability.

Question 5: CRITICAL THINKING

Game A costs $10 to play and pays $0 with probability 0.99 and $1,000 with probability 0.01. Game B costs $10 to play and pays $10 with probability 0.50 and $10 with probability 0.50 (i.e., you either break even or break even). Calculate the expected net gain for both games. Then explain why someone might prefer Game B over Game A despite Game A having a higher expected value.

Show answer

Game A: E(net) = ($0 − $10) × 0.99 + ($1,000 − $10) × 0.01 = (−$10)(0.99) + ($990)(0.01) = −$9.90 + $9.90 = $0.00. Game B: E(net) = ($10 − $10) × 0.50 + ($10 − $10) × 0.50 = $0 + $0 = $0.00. Both games have an expected net gain of exactly $0, so they are equally fair in the long run. However, someone might prefer Game B because it has zero variability — you always break even. Game A is extremely volatile: you lose $10 almost every time (99% of the time), and you only occasionally win big. A risk-averse person would strongly prefer the guaranteed $0 outcome of Game B. This shows that expected value alone doesn't capture all the information needed for decision-making — you also need to consider variance and personal risk tolerance.

SUMMARY

Lesson Summary

Summary

Expected value is the weighted average of all possible outcomes of a discrete random variable, where each outcome is weighted by its probability. The formula E(X) = Σ xᵢ · P(xᵢ) tells you to multiply each outcome by how likely it is, then add the products together. The result is a single number representing the long-run average — the value the actual average approaches as the experiment is repeated many times.

Key things to remember: expected value does not need to be a possible outcome (like 3.5 on a die). Always check that your probabilities sum to 1 before computing E(X). A fair game has an expected net gain of zero. Positive expected value favors the player; negative expected value favors the house. While E(X) is powerful, it doesn't capture variability or risk — for that, you'll need variance and standard deviation.