Introducing Probability — AP Statistics Study Guide

For: AP Statistics candidates sitting AP Statistics.

Covers: Basic probability notation, sample spaces, the law of large numbers, events, the complement rule, and probability calculation for equally likely outcomes, all core introductory topics for Unit 4 assessment.

You should already know: Basic set notation for subsets, descriptive statistics for sample proportions, basic counting principles for small sets.

A note on the practice questions: All worked questions in the "Practice Questions" section below are original problems written by us in the AP Statistics style for educational use. They are not reproductions of past College Board / Cambridge / IB papers and may differ in wording, numerical values, or context. Use them to practise the technique; cross-check with official mark schemes for grading conventions.

1. What Is Introducing Probability?

Introducing probability establishes the foundational framework for all of Unit 4: Probability, Random Variables, and Probability Distributions, which accounts for 10–15% of the total AP Statistics exam weight. This topic appears in both multiple-choice (MCQ) and free-response (FRQ) sections, most often as a foundational step for larger problems, with 2–3 standalone MCQ questions common on every exam. Probability is the study of chance and uncertainty, and the AP Statistics curriculum uses the frequentist definition: the probability of an event is the proportion of times the event would occur in a very large number of repeated identical trials of a random process. Standard notation writes the probability of event $A$ as $P(A)$, with all probabilities falling between 0 (an impossible event) and 1 (a certain event). This topic is the backbone of all later probability work and statistical inference, as inference relies on probability to quantify uncertainty about population parameters from sample data.

2. The Law of Large Numbers

The Law of Large Numbers (LLN) is the core conceptual rule that connects long-run behavior to probability. The LLN states that as the number of repetitions of a random process increases, the sample proportion of times an event occurs approaches the true, fixed probability of that event. The key intuition here is that short runs of a random process can have highly variable results that do not reflect the true probability, but this variation fades as more trials are added. A common misconception associated with the LLN is the gambler’s fallacy: the false belief that short-run streaks of one outcome will be "corrected" by the opposite outcome in the near future to balance out the proportion. In reality, the LLN only works by diluting past streaks with many more new trials, it does not compensate for past variation. Independent trials (like coin flips or roulette spins) have no memory of past outcomes, so past streaks do not change the probability of future outcomes.

Worked Example

A slot machine player has lost 12 straight spins on a fair slot machine and claims that they are "due for a win" because the law of large numbers says the proportion of wins will approach the advertised 10% win rate. Is the player’s reasoning correct? Explain what the LLN actually predicts for this scenario.

1. The player’s reasoning is incorrect, and they are falling for the gambler's fallacy.
2. The LLN only applies to the long-run average over thousands of trials, not to the next few spins after a short losing streak. Each spin of the slot machine is independent, so a past losing streak does not change the 10% win probability for the next spin.
3. If the player continues to play thousands of additional spins, the 12 initial losses will become a tiny fraction of the total number of spins.
4. Over this large number of total spins, the overall proportion of wins will approach the true 10% probability, regardless of the early losing streak.

Exam tip: Any AP Exam question about streaks or the LLN is almost always testing for the gambler’s fallacy — always explicitly state that LLN does not correct short-run streaks in independent trials.

3. Sample Spaces and Equally Likely Outcomes

A random process is any process with uncertain outcomes before it occurs. A sample space $S$ is the set of all possible outcomes of a random process. An event is any collection of outcomes (any subset of the sample space), typically labeled with a capital letter like $A$. When every outcome in the sample space is equally likely to occur, the probability of an event is calculated as the ratio of the number of favorable outcomes to the total number of outcomes in the sample space: $$P(A)=\frac{n(A)}{n(S)}$$ where $n(A)$ is the number of outcomes in event $A$ and $n(S)$ is the total number of outcomes in the sample space. Two core axioms always hold for this calculation: all probabilities are between 0 and 1, and the sum of probabilities for all outcomes in the sample space equals 1. A common mistake here is treating groups of outcomes (like color groups of marbles) as equally likely, when only individual outcomes are equally likely.

Worked Example

You flip a fair coin three times. What is the probability that you get exactly two heads?

1. First, list all outcomes in the sample space, counting ordered flips (since each flip is independent): $(HHH),(HHT),(HTH),(HTT),(THH),(THT),(TTH),(TTT)$. This gives $n(S)=8$ equally likely outcomes.
2. Count the number of outcomes with exactly two heads: $(HHT),(HTH),(THH)$, so $n(A)=3$.
3. Calculate the probability: $P(\text{exactly two heads})=\frac{3}{8}=0.375$.
4. Confirm the result is between 0 and 1, which satisfies the basic probability axiom.

Exam tip: Always count ordered outcomes for independent sequential trials (like multiple coin flips or multiple draws with replacement), even if the problem does not specify order — ordered outcomes are always equally likely for these processes.

4. The Complement Rule

The complement of an event $A$, written $A^c$, is the event that $A$ does not occur. Since $A$ and $A^c$ together cover the entire sample space, the sum of their probabilities equals 1. Rearranging this gives the complement rule: $$P(A^c)=1-P(A)$$ This rule is one of the most useful tools in introductory probability because it drastically simplifies calculations for common problems like finding the probability of "at least one" of an event occurring. Instead of counting all the cases where at least one event occurs (which can require adding many terms), you only need to calculate the probability of the complement (no events occurring) and subtract from 1. This reduces counting errors and saves significant time on the exam.

Worked Example

A box of 20 donuts has 3 donuts with cream filling. If you select 2 donuts at random without replacement, what is the probability that at least one of the donuts you select has cream filling?

1. Define event $A$: "at least one donut has cream filling". The complement $A^c$ is "neither donut has cream filling".
2. Calculate $P(A^c)$: The probability the first donut is not cream-filled is $\frac{17}{20}$, and the probability the second is also not cream-filled (after removing one non-cream donut) is $\frac{16}{19}$. So $P(A^c)=\frac{17}{20}\times \frac{16}{19}=\frac{272}{380}=\frac{68}{95}\approx 0.7158$.
3. Apply the complement rule: $P(A)=1-P(A^c)=1-\frac{68}{95}=\frac{27}{95}\approx 0.2842$.
4. Direct counting would require adding the probability of exactly one cream-filled and exactly two cream-filled, which gives the same result but requires twice as much calculation.

Exam tip: If a problem asks for the probability of "at least one", the complement rule is always the fastest, most error-free approach — never start by adding multiple terms unless the complement is more complicated.

5. Common Pitfalls (and how to avoid them)

• Wrong move: Interpreting a 20% probability of rain tomorrow as "it will rain for 20% of the day". Why: Students confuse the probability of an event occurring with a proportion of time or area, misapplying the frequentist definition. Correct move: Interpret probability as the proportion of times the event would occur if the same conditions were repeated many times — a 20% rain chance means rain occurs in 2 out of 10 identical weather scenarios.
• Wrong move: Claiming that after 6 straight heads in coin flips, tails is more likely next flip because the law of large numbers will balance the proportion. Why: Students confuse long-run convergence with short-run correction of streaks, falling for the gambler's fallacy. Correct move: Always note that independent trials have no memory; LLN only dilutes past streaks with many future trials, it does not change the probability of the next outcome.
• Wrong move: When rolling two identical dice, calculating $P(\text{sum}=3)=\frac{1}{21}$ by counting unordered outcomes. Why: Students assume identical objects mean unordered outcomes are equally likely, but even identical dice are independent, so ordered outcomes are equally likely. Correct move: Always count ordered outcomes for separate independent trials, even if the objects are not visibly distinguishable, so $P(\text{sum}=3)=\frac{2}{36}=\frac{1}{18}$.
• Wrong move: Calculating $P(\text{at least one defective in 5 items})$ by adding $P(1)+P(2)+...+P(5)$ and making a counting error. Why: Students forget the complement rule simplifies "at least one" problems significantly, leading to unnecessary extra work. Correct move: Always calculate $P(\text{at least one})=1-P(\text{none defective})$ to reduce the problem to one simple calculation.
• Wrong move: Reporting a final probability of 1.3 or -0.2 after calculation. Why: Students rush through calculations and forget the fundamental axiom that all probabilities are between 0 and 1. Correct move: Always check that your final probability falls between 0 and 1 before moving on; any value outside this range confirms you made an error.

6. Practice Questions (AP Statistics Style)

Question 1 (Multiple Choice)

A bakery makes chocolate chip cookies, and 15% of all cookies have fewer than 5 chocolate chips. A customer buys 5 random cookies. What is the probability that at least one of the cookies has fewer than 5 chocolate chips? A) $0.15^5\approx 0.0001$ B) $5\times 0.15=0.75$ C) $1-(0.85{)}^5\approx 0.5563$ D) $(0.85{)}^5\approx 0.4437$

Worked Solution: We need the probability of at least one cookie with fewer than 5 chips. By the complement rule, the complement of "at least one" is "all 5 cookies have 5 or more chips". The probability a single cookie has 5 or more chips is $1-0.15=0.85$, so for independent cookies, $P(\text{all good})=(0.85{)}^5$. Subtracting from 1 gives $P(\text{at least one bad})=1-(0.85{)}^5\approx 0.5563$. Option A is the probability all 5 are bad, option B incorrectly multiplies number of cookies by individual probability, and option D is the complement of the desired result. The correct answer is C.

Question 2 (Free Response)

A drawer contains 4 white socks, 3 black socks, and 2 gray socks. All socks are equally likely to be drawn. (a) One sock is drawn at random. What is the probability the sock is gray? (b) One sock is drawn at random. What is the probability the sock is not black? (c) A student claims that since there are 3 colors, the probability of drawing each color is $\frac{1}{3}$. Explain why the student is incorrect.

Worked Solution: (a) First count total socks: $4+3+2=9$ total equally likely outcomes. Number of gray socks is 2. So $P(\text{gray})=\frac{2}{9}\approx 0.222$. (b) Let $B$ be the event the sock is black. $P(B)=\frac{3}{9}=\frac{1}{3}$. By the complement rule, $P(\text{not }B)=1-P(B)=1-\frac{1}{3}=\frac{2}{3}\approx 0.667$. (c) The student incorrectly assumes that each color is an equally likely outcome. Equally likely outcomes apply to individual socks, not to color groups. Since there are different numbers of socks per color, each color does not have the same probability of being drawn, so the $\frac{1}{3}$ calculation is wrong.

Question 3 (Application / Real-World Style)

A wildlife biologist estimates that 70% of adult deer in a forest have been exposed to a specific parasite. The biologist randomly selects 6 adult deer for testing. What is the probability that at least one of the selected deer has not been exposed to the parasite? Round your answer to four decimal places, and interpret the result in context.

Worked Solution: Let $A$ be the event that at least one selected deer is not exposed. The complement $A^c$ is the event that all 6 deer are exposed. The probability a single deer is exposed is 0.7, so $P(A^c)=(0.7{)}^6\approx 0.1176$. By the complement rule, $P(A)=1-0.1176=0.8824$. Interpretation: If we repeatedly select 6 random adult deer from this forest, approximately 88.24% of these selections will include at least one deer that has not been exposed to the parasite.

7. Quick Reference Cheatsheet

8. What's Next

This chapter lays the foundational vocabulary and rules you need for all subsequent probability topics in Unit 4. Immediately after this, you will learn how to calculate probabilities for compound events using the addition rule and multiplication rule for independent and dependent events, which rely directly on the complement rule and basic probability interpretation you learned here. Without mastering the basics of probability, the law of large numbers, and the complement rule, you will not be able to correctly calculate probabilities for random variables or interpret sampling distributions, which are core to statistical inference later in the course. This topic also feeds into expected value calculation and confidence interval interpretation, both heavily tested on the AP Exam.

• Addition and Multiplication Rules of Probability
• Conditional Probability
• Introducing Random Variables
• Binomial Probability Distributions