AP Statistics Unit 4 Probability Rules: How to Compute and Approximate Probabilities

Estimating Probabilities Using Simulation

What simulation is (and what it is not)

A simulation is a method for estimating a probability by mimicking a chance process many times and using the long-run relative frequency of an outcome as an approximation of the true probability. You use simulation when the probability is difficult to compute exactly—often because the situation is too complex for a clean counting approach, or because events depend on many steps.

Simulation is not “making up” data. A good simulation is grounded in a clear probability model: you specify how randomness enters the situation, how each trial works, and exactly what counts as a success.

Why simulation matters in AP Statistics

AP Statistics emphasizes that probability can be approached in two complementary ways:

• Theoretical probability: compute exactly using rules and counting.
• Empirical probability: estimate using data or repeated random trials.

Simulation is your main tool for empirical probability when you can’t easily run the real process thousands of times (or when doing so would be too expensive, too slow, or unethical). It also helps you check whether a theoretical result seems reasonable.

How simulation works: the essential steps

A strong simulation solution is organized and explicit. Think of it as a recipe someone else could follow and get the same kind of results.

1. Define the chance process and assumptions.

   • What outcomes are possible?
   • Are outcomes equally likely? Are trials independent?
   • What probabilities are you assuming for each outcome?

2. Choose a randomization method.
   Common AP-friendly tools include:

   • Random digits (table of random digits)
   • A random number generator (calculator or software)
   • Physical devices (coins, dice, spinners) when they match the model

3. Design one trial.
   A trial is one complete repetition of the process. You must state:

   • How you will use random numbers to generate outcomes
   • When the trial ends
   • How you record whether the trial is a success

4. Repeat many trials and count successes.
   The estimate improves as the number of trials increases (because of the law of large numbers).

5. Compute the estimate as a relative frequency.
   If you run n trials and get s successes, your simulation estimate is:

   \hat{p} = \frac{s}{n}

6. Interpret the result in context.
   Use a sentence that ties the number back to the real event (and ideally mentions it’s an estimate).

Connecting simulation to the law of large numbers

The law of large numbers says that as you repeat a chance process many times, the sample proportion of successes tends to get closer to the true probability. This does not mean results will be “nice” in the short run—small simulations can bounce around a lot.

A common mistake is thinking a simulation with only 20 or 30 trials is automatically “good enough.” On the AP exam, you’re often expected to run at least a few dozen trials, and more is better when feasible.

Example 1: Simulation with random digits (worked)

Situation: A store says that 30% of its customers use a coupon. You observe 8 customers. Estimate the probability that at least 4 use a coupon.

Step 1: Model

Assume each customer independently uses a coupon with probability 0.30.

Step 2: Random digits mapping

Use two-digit numbers from 00–99.

• Coupon user: 00–29 (30 numbers)
• Not a coupon user: 30–99 (70 numbers)

Step 3: One trial

Read 8 two-digit numbers. Count how many are 00–29. Record “success” if the count is 4 or more.

Step 4: Repeat

Suppose you perform 100 trials and find 12 trials had at least 4 coupon users.

Step 5: Estimate

\hat{p} = \frac{12}{100} = 0.12

Interpretation

Based on the simulation, the probability that at least 4 of 8 customers use a coupon is approximately 0.12.

What can go wrong:

• Using an incorrect mapping (e.g., 00–30 would be 31%).
• Forgetting the independence assumption (which might fail if customers influence each other).

Example 2: Simulation for a “complicated” rule

Simulation is especially helpful when the event is defined by a rule that’s awkward to count.

Situation: A game flips a fair coin until it gets two heads in a row or until 6 flips occur, whichever comes first. Estimate the probability the game ends with “two heads in a row.”

You can simulate each trial by generating coin flips and applying the stopping rule. Even though the coin is simple, the stopping condition makes direct calculation harder than it looks.

Exam Focus

• Typical question patterns:
  • “Describe how you would use simulation to estimate…” (you must outline mapping, trials, repetition, and what counts as success).
  • “Using the following random digits…” then run a specified number of trials and compute an estimate.
  • “Is this simulation valid?” questions that target assumptions (independence, correct probabilities, realistic model).
• Common mistakes:
  • Not defining a clear trial and success condition (graders need specifics).
  • Using too few trials or forgetting to compute a relative frequency.
  • Mismatching random numbers to outcomes (probabilities in the mapping don’t match the context).

Mutually Exclusive Events and the Addition Rule

Events, unions, and intersections (the language you must speak)

In probability, an event is a set of outcomes. If A and B are events:

• A \cup B means “A or B (or both)” (the union).
• A \cap B means “A and B” (the intersection).

Students often misread “or” in everyday language as exclusive. In probability, “or” is usually inclusive unless stated otherwise.

Mutually exclusive events: what it means

Two events are mutually exclusive (also called disjoint) if they cannot happen at the same time. Formally, they have no overlap:

P(A \cap B) = 0

Why it matters: If events are mutually exclusive, the probability of “A or B” is simpler to compute.

The Addition Rule (general and special case)

The key probability rule for “or” is the general addition rule:

P(A \cup B) = P(A) + P(B) - P(A \cap B)

Why do we subtract? If you add P(A) and P(B), any outcomes that are in both A and B get counted twice—once in P(A) and once in P(B). Subtracting P(A \cap B) fixes the double-counting.

If A and B are mutually exclusive, then P(A \cap B) = 0 and the rule becomes:

P(A \cup B) = P(A) + P(B)

Example 1: Using the general addition rule (worked)

Situation: In a certain school, 55% of students play a sport, 40% are in band, and 20% do both. If you pick a student at random, what is the probability the student plays a sport or is in band?

Let S = plays a sport, B = in band.

Given:

• P(S) = 0.55
• P(B) = 0.40
• P(S \cap B) = 0.20

Compute:

P(S \cup B) = 0.55 + 0.40 - 0.20 = 0.75

Interpretation: There is a 0.75 probability a randomly chosen student plays a sport or is in band (or both).

What can go wrong: A very common error is to do 0.55 + 0.40 = 0.95 and forget to subtract the overlap.

Example 2: Recognizing mutually exclusive events

Situation: Draw one card from a standard 52-card deck. Let A be the event “the card is a heart” and B be the event “the card is a spade.”

A card cannot be both a heart and a spade, so A and B are mutually exclusive.

Thus:

P(A \cup B) = P(A) + P(B) = \frac{13}{52} + \frac{13}{52} = \frac{26}{52} = 0.5

A helpful interpretation tool: Venn diagrams

Venn diagrams are not just drawings—they’re a reasoning tool. The general addition rule corresponds to “add both circles, subtract the overlapping region once.” If your problem includes a “both” probability, you are almost certainly in general-addition-rule territory.

Exam Focus

• Typical question patterns:
  • Compute P(A \cup B) given P(A), P(B), and P(A \cap B).
  • Identify whether events are mutually exclusive from context, then decide whether you can use the simplified rule.
  • Use a Venn diagram or interpret a Venn diagram to find missing probabilities.
• Common mistakes:
  • Treating “or” as exclusive automatically and using P(A)+P(B) when overlap exists.
  • Confusing “mutually exclusive” with “independent” (they are very different; see next section).
  • Mixing up A \cup B and A \cap B (“or” vs “and”).

Conditional Probability and Independence

Conditional probability: updating the probability after learning information

Conditional probability is the probability of an event given that another event has occurred. Notation:

• P(A \mid B) means “the probability of A given B.”

Why it matters: Real decisions often happen with partial information. Once you know B happened, your “universe” shrinks to just the outcomes in B, and you reassess how likely A is inside that restricted set.

The conditional probability formula

When P(B) > 0, conditional probability is defined by:

P(A \mid B) = \frac{P(A \cap B)}{P(B)}

Read this as: “Of the outcomes where B happens, what fraction also have A?”

This definition is foundational because it connects conditional probability to the probability rules for intersections (which leads directly to the multiplication rule later).

Using two-way tables (a high-clarity method)

Many conditional probability questions are easiest with a two-way table because it forces you to keep track of counts or proportions.

If the problem gives counts, you can compute conditional probabilities as:

P(A \mid B) = \frac{\text{count}(A \cap B)}{\text{count}(B)}

(You are effectively dividing within the “given” group.)

Example 1: Conditional probability from a table (worked)

Situation: In a sample of 200 adults, 80 own a bike. Of those 80 bike owners, 50 also own a car. Find the probability an adult owns a car given that they own a bike.

Let B = owns a bike, C = owns a car.

We are told:

• \text{count}(B) = 80
• \text{count}(B \cap C) = 50

Compute:

P(C \mid B) = \frac{50}{80} = 0.625

Interpretation: Among bike owners, about 62.5% also own a car.

What can go wrong: Dividing by the wrong total (like dividing by 200 instead of by 80). The phrase “given that they own a bike” means your denominator must be the bike group.

Independence: when knowing one event doesn’t change the other

Two events A and B are independent if knowing one happened does not change the probability of the other. A core way to state this is:

P(A \mid B) = P(A)

(as long as P(B) > 0). Equivalently:

P(B \mid A) = P(B)

(as long as P(A) > 0).

Why it matters: Independence is one of the biggest “structure” ideas in probability. If events are independent, many computations become simpler (especially “and” probabilities). If they are not independent, you must account for how the first event changes the second.

Independence vs. mutually exclusive (do not mix these up)

These are commonly confused because both describe relationships between events.

• Mutually exclusive: cannot occur together.
  • Characterized by P(A \cap B) = 0.
• Independent: can occur together, and one does not affect the other.
  • Characterized by P(A \mid B) = P(A).

A crucial fact: if A and B are mutually exclusive and both have positive probability, they cannot be independent. If P(A) > 0 and P(B) > 0 but P(A \cap B) = 0, then knowing B happened forces A not to happen, so P(A \mid B) = 0, which is not equal to P(A).

Example 2: Checking independence with probabilities

Situation: Suppose P(A) = 0.3, P(B) = 0.5, and P(A \cap B) = 0.15. Are A and B independent?

One way to check is to see whether:

P(A \cap B) = P(A)P(B)

Compute:

P(A)P(B) = 0.3 \times 0.5 = 0.15

This matches P(A \cap B) = 0.15, so A and B are independent.

What can go wrong: Students sometimes assume events are independent just because the story “sounds unrelated.” On AP problems, you should justify independence using a definition or given information.

Notation reference (common equivalent tests for independence)

Exam Focus

• Typical question patterns:
  • Compute P(A \mid B) from a two-way table, counts, or given probabilities.
  • Decide whether events are independent by checking whether conditioning changes probabilities or by testing P(A \cap B) = P(A)P(B).
  • Interpret a conditional probability in context (what group you are restricting to).
• Common mistakes:
  • Using the wrong denominator for P(A \mid B) (you must divide by the “given” event).
  • Claiming independence because events are “different” without a probability-based justification.
  • Confusing independence with mutually exclusive events.

The Multiplication Rule

What the multiplication rule is really about

The multiplication rule is the main rule for calculating the probability of an intersection (“and”). It is not a separate idea from conditional probability—it’s essentially conditional probability rearranged.

When you want P(A \cap B), you are asking: “What is the probability that both happen?” In many real processes (medical testing, manufacturing quality checks, drawing cards, multi-step games), “both” occurs through a sequence: one thing happens, then another happens possibly with updated odds.

The general multiplication rule

From the definition of conditional probability:

P(A \mid B) = \frac{P(A \cap B)}{P(B)}

Multiply both sides by P(B) to get the general multiplication rule:

P(A \cap B) = P(B)P(A \mid B)

Similarly, you can also write:

P(A \cap B) = P(A)P(B \mid A)

How to choose which form: Use the one where the conditional probability is easier to compute from the story. If the process is naturally “first A, then B,” the second form often matches the narrative.

The multiplication rule for independent events (special case)

If A and B are independent, then P(B \mid A) = P(B), so:

P(A \cap B) = P(A)P(B)

This is a powerful shortcut—but only when independence is justified.

Tree diagrams as a multiplication rule organizer

A tree diagram helps you track multi-step processes:

• Each branch is a conditional probability at that step.
• The probability of a complete path is found by multiplying along the branches (that is the multiplication rule in action).

Students often draw trees but then add along a path or multiply across different paths. The reliable method is:

• Multiply along a path to get an “and.”
• Add across paths to get an “or.”

Example 1: Dependent events (without replacement) (worked)

Situation: A bag has 5 red and 3 blue marbles. Two marbles are drawn without replacement. Find the probability both are red.

Let R_1 = first draw is red, R_2 = second draw is red.

We want P(R_1 \cap R_2).

First draw red:

P(R_1) = \frac{5}{8}

Given the first was red, there are now 4 red out of 7 total:

P(R_2 \mid R_1) = \frac{4}{7}

Multiply:

P(R_1 \cap R_2) = \frac{5}{8} \times \frac{4}{7} = \frac{20}{56} = \frac{5}{14}

Interpretation: The probability of drawing two reds without replacement is \frac{5}{14}.

What can go wrong: Treating the draws as independent and computing \frac{5}{8} \times \frac{5}{8}. Without replacement, the second draw depends on the first.

Example 2: Independent events (with replacement or separate trials)

Situation: A fair coin is flipped 3 times. Find the probability of getting heads on all 3 flips.

Let H_i be the event that flip i is heads. Each flip is independent and P(H_i) = \frac{1}{2}.

Thus:

P(H_1 \cap H_2 \cap H_3) = \left(\frac{1}{2}\right)\left(\frac{1}{2}\right)\left(\frac{1}{2}\right) = \frac{1}{8}

Example 3: Using multiplication with conditional probability from context

Situation: 2% of items from a supplier are defective. A test correctly flags a defective item 95% of the time. Estimate the probability that a randomly selected item is defective and the test flags it.

Let D = item is defective, F = test flags defective.

Given:

• P(D) = 0.02
• P(F \mid D) = 0.95

We want P(D \cap F):

P(D \cap F) = P(D)P(F \mid D) = 0.02 \times 0.95 = 0.019

Interpretation: About 1.9% of all items are both defective and flagged by the test.

What can go wrong: Confusing P(F \mid D) with P(D \mid F). The probability the test flags given defective is not the same as the probability an item is defective given it was flagged.

Exam Focus

• Typical question patterns:
  • Compute P(A \cap B) using a given conditional probability such as P(B \mid A).
  • Multi-step “and then” probability using a tree diagram (with or without independence).
  • Identify whether the independence shortcut P(A \cap B) = P(A)P(B) is valid.
• Common mistakes:
  • Multiplying probabilities that should be conditional as if they were unconditional (especially in without-replacement settings).
  • Mixing up conditional directions (using P(A \mid B) when the problem provides P(B \mid A)).
  • Adding when the question asks for “and,” or multiplying when the question asks for “or.”