3.3 Two Basic Rules of Probability

The two rules that do the arithmetic of the chapter, and the two special cases that simplify them. The multiplication rule computes an AND probability as P(B) times the conditional P(A given B), and collapses to a plain product exactly when the events are independent. The addition rule computes an OR probability as the sum of the two probabilities minus their intersection, and loses the subtraction exactly when the events are mutually exclusive. Both general forms are always valid, so section 3.2's two properties turn out to be permissions to drop a term rather than separate techniques — which is why identifying them was worth the effort.

Subject: Statistics · 66 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. Two Basic Rules of Probability

Title

Statistics · Chapter 3 — Probability Topics

One rule for AND, one rule for OR, and the two conditions that shorten them

2. What this lesson gives you

Objectives

Section 3.2 taught you to recognise two properties a pair of events can have: independence and mutual exclusivity. It never told you what those properties are for. This section answers that. Almost every probability question in the rest of the chapter asks for one of two things, the chance that two events both happen or the chance that at least one of them does, and there is exactly one general rule for each.

Keep one idea in view the whole way through. The two general rules are always true, for any events at all. Independence and mutual exclusivity are not extra rules; they are permissions to delete one term. That is why checking them carefully matters: delete a term you were not entitled to delete and the answer is wrong, even though the arithmetic is perfect.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 179 — the two rules and their special cases

3. Before anything new: the conditional formula

Warm-up

Everything in this lesson grows out of one formula from Section 3.1. Write it down from memory before you read the answer, together with the two definitions from Section 3.2.

Discussion prompt

Write the formula for the probability of A given B. Then say, in one sentence each, what it means for A and B to be independent and what it means for them to be mutually exclusive.

Notice that the two definitions talk about different things. Independence is about a conditional probability staying the same; mutual exclusivity is about an intersection being empty. Hold on to that difference, because the two rules in this lesson each use exactly one of them, and mixing them up is the most common error in the whole chapter.

Look also at what the conditional formula has in its numerator: the probability of A AND B. Turn that formula around and it tells you how to compute an AND probability. That rearrangement is the first rule of the lesson.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 179 — the rule written in its rearranged form

4. The multiplication rule

Section

Part 1

5. Rearrange the conditional formula and you have the rule

Concept

Start from the definition you just wrote down. It expresses a conditional probability as a fraction whose top is the AND probability you want.

\[ P(A \mid B) = \frac{P(A\text{ AND }B)}{P(B)} \]

Multiply both sides by the probability of B. That is the only move: the B in the denominator cancels on the right and appears as a factor on the left.

\[ P(B)\,P(A \mid B) = P(A\text{ AND }B) \]

Read it right to left and you have the rule as the book states it. For any two events, the chance that both happen is the chance of one of them times the chance of the other given the first.

\[ P(A\text{ AND }B) = P(B)\,P(A \mid B) \]

multiplication rule — For any two events A and B on the same sample space, the probability of A AND B equals the probability of B times the probability of A given B. It needs no assumption about the events.

Nothing has been assumed here. The rule is the definition of conditional probability, rearranged, so it holds for every pair of events you will ever meet, dependent or not. That is what makes it safe to reach for first.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 179 — The Multiplication Rule

6. Reading the multiplication rule piece by piece

Notation

Each symbol in the rule does one job. Knowing which job protects you from the classic mistake of multiplying the wrong two numbers.

Annotate

On: \( P(A\text{ AND }B) = P(B)\;P(A \mid B) \)

  • Both events happen. This is the intersection, the outcomes that sit in A and in B at the same time. It is what you are asked to find.
  • The chance of the event you condition on. Think of it as the first gate: B has to happen before the second factor even comes into play.
  • The chance of A inside the world where B has already happened. This is the ingredient that carries any dependence between the events. It is usually the number the problem hands you with the words given, of those, or after.
  • A fraction of a fraction. First keep the share of outcomes in B, then keep the share of those that are also in A.

The vertical bar is the part people misread. The event on the right of the bar is the one assumed to have happened; the event on the left is the one whose chance you are measuring. So the factor outside the conditional must be the event to the right of the bar, never the other one.

Why a fraction of a fraction? Imagine 1000 repetitions. About 1000 times the probability of B of them have B happen. Among just those, the share where A also happens is the conditional. So the count of both is the count of B times that share, and dividing by 1000 turns counts back into probabilities.

A quick self-test: if a problem gives you the probability of B given A, which plain probability do you multiply it by? The one to the right of the bar, the probability of A. The next slide says why either order is allowed.

7. Either order works, so use the conditional you are given

Concept

AND does not care about order: A AND B is the same set of outcomes as B AND A. So you may condition on either event, and the rule has a mirror image.

\[ P(A\text{ AND }B) = P(B)\,P(A \mid B) \]

\[ P(A\text{ AND }B) = P(A)\,P(B \mid A) \]

Both lines are true at once, and they give the same number. Which one you use is decided entirely by which conditional probability the problem gives you. Given the chance of B after A, use the second line; given the chance of A after B, use the first.

Here is the habit that makes this automatic. Find the conditional in the problem first. Whatever sits to the right of its bar is the plain probability that multiplies it.

the problem gives youusemultiply by
P(B given A)P(A AND B) = P(A) P(B given A)P(A)
P(A given B)P(A AND B) = P(B) P(A given B)P(B)
P(M given S) = 0.25 (Felicity)P(M AND S) = P(S) P(M given S)P(S) = 0.65

The Carlos example on the next slides gives the chance of the second goal given the first, so there the first goal is the plain factor. Felicity's example gives math given speech, so there speech is.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — Example 3.15(a) uses the mirror form

8. The rule as an area: a fraction of a fraction

Picture it

Here is a picture that makes the rule obvious, using Felicity's numbers from Example 3.17. The whole square has area 1 and stands for every possible way her enrolment could go.

Figure (svg): A unit square split into two columns: 'speech' of width 0.65 and 'no speech' of width 0.35. In the speech column the bottom quarter is shaded and labelled M AND S, 0.1625; in the other column a thinner band is shaded, 0.0375.

The shaded yellow block is a rectangle: its width is the probability of speech and its height is the chance of math within speech. Its area is the product, and that product is the multiplication rule.

Cut the square into two columns. The left one has width 0.65, the probability that she takes speech. Inside that column, shade the bottom quarter: 0.25 is the probability of math given speech, so it is a height measured within the speech column only.

\[ \text{area} = 0.65 \times 0.25 = 0.1625 \]

The yellow block is exactly the set of outcomes where she takes both. Its area is width times height, and width times height is the probability of speech times the probability of math given speech. The multiplication rule is nothing more than the area of a rectangle.

The picture also shows why the plain probability of math, 0.2, is the wrong height to use. That 0.2 is spread over both columns; inside the speech column the proportion is 0.25. Using 0.2 would draw the block the wrong height.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 182 — Example 3.17

9. Example 3.15(a): Carlos makes both goals

Worked example

Carlos makes 65 percent of his shots, and he tends to shoot in streaks: after a made first goal, his chance on the second rises to 0.90. Find the probability that he makes both.

\[ P(A) = 0.65, \quad P(B) = 0.65, \quad P(B \mid A) = 0.90 \]

Sort the three numbers before you use any of them. Two are plain probabilities, one for each shot taken on its own. The third is conditional: it is measured only on the occasions when the first shot went in. The words given that he made the first goal are what tell you so, and they also tell you which event sits to the right of the bar.

Name what is asked

Why: both goals means the first AND the second, an intersection.

\[ \text{find } P(A\text{ AND }B) \]

Choose the form that uses the given conditional

Why: the problem gives B given A, so A is the plain factor.

\[ P(A\text{ AND }B) = P(A)\,P(B \mid A) \]

Substitute

Why: 0.65 for the first goal, 0.90 for the second given the first.

\[ P(A\text{ AND }B) = (0.65)(0.90) \]

Multiply

Why: the product of the two decimals.

\[ P(A\text{ AND }B) = 0.585 \]

Figure (svg): A probability tree. From the start, a branch 'make 1st: 0.65' and a branch 'miss 1st: 0.35'. After a make, 'make 2nd: 0.90' and 'miss 2nd: 0.10'; after a miss, 'make 2nd: 0.186' and 'miss 2nd: 0.814'. The four path ends read 0.585, 0.065, 0.065 and 0.285.

Every path's probability is a product: the first branch times the conditional branch that follows it. The four ends add to 1, as the four ways the two shots can go must.

The tree shows the same calculation as a path: out along the make-first branch, then along the make-second branch drawn after it. The other three paths use numbers derived from the given ones, for example 0.35 for a missed first shot.

Check that the answer can be right

Why: both goals is a subset of the first goal, so its probability must not exceed 0.65, and it does not; the four path ends of the tree also add to exactly 1.

\[ 0.585 \le 0.65, \qquad 0.585 + 0.065 + 0.065 + 0.285 = 1 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — Example 3.15(a)

10. Why the streak has to be in the calculation

Intuition

It is tempting to say Carlos makes each shot 65 percent of the time, so both is 0.65 times 0.65. That reasoning treats the second shot as if it had no memory of the first. The problem says it does: he shoots in streaks.

\[ P(B \mid A) = 0.90 \ne 0.65 = P(B) \]

A streak means success raises the chance of further success. So the paths where he makes the first shot are more likely to continue with a make, and the both-goals path gets heavier than a memoryless product would make it.

\[ 0.585 > 0.4225 = (0.65)(0.65) \]

The general rule captures this automatically, because the second factor is whatever the conditional actually is. You never need to decide in advance whether the streak matters; you only need to use the conditional you were given.

Notice where the information about the streak lives: entirely in that one number, 0.90. Change it to 0.65 and the streak vanishes; change it to 0.50 and you have a shooter who gets rattled after a make. The rule handles every one of these without changing shape.

11. Try It 3.15: Helen at the free-throw line

Prediction

Helen makes 75 percent of her free throws, and 85 percent of her second shots after making the first. Before any arithmetic, place the answer.

Predict first

Compared with 0.75 times 0.75, which is 0.5625, the true probability that Helen makes both is

  • larger than 0.5625
  • smaller than 0.5625
  • exactly 0.5625
  • impossible to place without more data

Correct: larger than 0.5625

Why: Like Carlos, Helen does better after a make: her chance on the second shot rises from 0.75 to 0.85. The both-shots path therefore carries more weight than a memoryless product would give it. The true value is 0.75 times 0.85, which is 0.6375.

The reasoning is the same comparison you just made for Carlos. The plain product assumes the second shot ignores the first. When the conditional is bigger than the plain probability, the joint probability is bigger than the plain product, and when it is smaller, the joint is smaller.

Making a rough prediction before computing is worth the three seconds it takes. If your final answer lands on the wrong side of the plain product, you know at once that you multiplied the wrong pair of numbers.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 181 — Try It 3.15

12. Try It 3.15: Helen makes both free throws

Worked example

Now compute it. C is making the first free throw and D is making the second.

\[ P(C) = 0.75, \quad P(D) = 0.75, \quad P(D \mid C) = 0.85 \]

The shape is identical to Carlos, and that is the point: once you can see the pattern, the names and numbers stop mattering. Two plain probabilities, one conditional, and a question about both. The conditional is measured on the occasions she made the first shot, so the first shot is the gate you pass through first.

Name what is asked

Why: both free throws is C AND D.

\[ \text{find } P(C\text{ AND }D) \]

Choose the form

Why: the given conditional is D given C, so C is the plain factor.

\[ P(C\text{ AND }D) = P(C)\,P(D \mid C) \]

Substitute

Why: first shot 0.75, second given first 0.85.

\[ P(C\text{ AND }D) = (0.75)(0.85) \]

Multiply

Why: seventy-five hundredths of eighty-five hundredths.

\[ P(C\text{ AND }D) = 0.6375 \]

The answer sits where the prediction put it, above 0.5625, and below both 0.75 and 0.85, as an intersection must.

Check with the other conditional you can now derive

Why: if she makes the second shot 75 percent of the time overall and 0.6375 of the time together with the first, then after a miss she makes it 0.1125 out of 0.25 of the time, which is 0.45. A player who is worse after a miss is exactly the streaky player described.

\[ P(D \mid C') = \frac{0.75 - 0.6375}{0.25} = \frac{0.1125}{0.25} = 0.45 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 181 — Try It 3.15

13. Example 3.17(a): Felicity takes math and speech

Worked example

The probability that Felicity enrols in math is 0.2, in speech 0.65, and in math given that she enrols in speech 0.25. Find the probability that she takes both.

\[ P(M) = 0.2, \quad P(S) = 0.65, \quad P(M \mid S) = 0.25 \]

This time the conditional runs the other way from Carlos: it measures math among the students who take speech. So speech is the gate and math is what you look for inside it. The area picture two slides back is this exact calculation, so keep it in mind as you go.

Name what is asked

Why: math and speech is the intersection.

\[ \text{find } P(M\text{ AND }S) \]

Choose the form

Why: the given conditional is M given S, so S is the plain factor; that is exactly the book's hint.

\[ P(M\text{ AND }S) = P(S)\,P(M \mid S) \]

Substitute

Why: speech 0.65, math given speech 0.25.

\[ P(M\text{ AND }S) = (0.65)(0.25) \]

Multiply

Why: a quarter of 0.65.

\[ P(M\text{ AND }S) = 0.1625 \]

Notice the number you did not use: the plain probability of math, 0.2. It is not needed for an AND probability here. It will be needed in part (b), which asks for an OR, and in part (c), which asks about independence.

Check against the other conditional

Why: the joint probability divided by the probability of math gives the chance of speech given math. It must be a genuine probability, between 0 and 1, and it is.

\[ P(S \mid M) = \frac{0.1625}{0.2} = 0.8125 \le 1 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 182 — Example 3.17(a)

14. Try It 3.17(a): a book and a DVD

Fill the middle

A student at the library checks out a book with probability 0.40 and a DVD with probability 0.30, and a DVD given a book with probability 0.5. Fill in the two blanks to find the chance of both.

Fill in the blanks

P(B\text0.40D) = P(B)\,P(D \mid B) = (0.2)(0.5) = ___

Why: The given conditional is D given B, so the plain factor is the probability of B, the book, which is 0.40. Half of 0.40 is 0.20. The probability of a DVD, 0.30, is not part of this product at all.

Read the conditional aloud as a sentence: of the visits where a book is taken, half also take a DVD. The group it is measured in, the book-takers, is the group whose probability multiplies it.

The trap on this one is the number 0.30. It is sitting right there in the problem and it is the probability of the other event, so it looks like it belongs. But the conditional has the book to the right of its bar, so the book is the one that multiplies.

A good sanity check is the ceiling. The answer 0.2 is below 0.40 and below 0.30, the chances of each event alone, as the chance of both must be. The rest of Try It 3.17 and Try It 3.19 come back to this library later in the lesson.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 182 — Try It 3.17(a)

15. Trap: multiplying the two plain probabilities

Trap

The trap

Asked for the chance that Carlos makes both goals, a common first line:

\[ P(A\text{ AND }B) = P(A)\,P(B) \]

\[ = (0.65)(0.65) = 0.4225 \]

Wrong. The arithmetic is fine; the formula is not the general rule.

It is an easy line to write because the two probabilities are both printed right there, both equal to 0.65, and the conditional 0.90 is tucked into a later sentence.

The fix

The plain product is the rule only when the second event ignores the first. Carlos shoots in streaks, so it does not. The general rule uses the conditional:

\[ P(A\text{ AND }B) = P(A)\,P(B \mid A) \]

\[ = (0.65)(0.90) = 0.585 \]

The habit that prevents this: always write the general rule first, with its conditional. Replace the conditional by a plain probability only after you have a reason, such as a stated or checked independence.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — Example 3.15

16. Exercise 80: which number multiplies which?

Step zero

A California survey reported that 61 percent of registered voters supported same-sex marriage, 60 percent said the coming court ruling was important to them, and of those who supported same-sex marriage, 75 percent said the ruling was important.

\[ P(C) = 0.61, \quad P(B) = 0.60, \quad P(B \mid C) = 0.75 \]

Discussion prompt

Part (g) asks for the probability of C AND B. Before you multiply anything, say which plain probability goes with the 0.75, and why the other one does not.

This is the whole skill in miniature. The phrase of those who support names the group the 75 percent was measured in, and that group's probability is the one that multiplies it. The words of those, among, given and after all do the same job: they point at the event to the right of the bar.

The survey also gave a fourth number, 78 percent support among 18 to 39 year olds. That is a different conditional, C given A, and it plays no part in part (g). Deciding which numbers a question does not need is part of the work.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3 homework, pp. 207-208 — Exercise 80

17. Exercise 80(g): support and importance together

Worked example

Finish the calculation set up on the last slide, then use it to say something about the voters.

Write the rule with the given conditional

Why: the conditional is B given C, so C is the plain factor.

\[ P(C\text{ AND }B) = P(C)\,P(B \mid C) \]

Say the product in words before computing it, because the words are the check on the setup. Of all voters, 61 percent support; of those supporters, three quarters say the ruling matters. Three quarters of 61 percent is the share of all voters who do both.

Substitute

Why: support 0.61, important given support 0.75.

\[ P(C\text{ AND }B) = (0.61)(0.75) \]

Multiply

Why: three quarters of 0.61.

\[ P(C\text{ AND }B) = 0.4575 \]

In context: about 46 percent of California registered voters both supported same-sex marriage and said the ruling mattered to them.

Compare with the plain product

Why: if support and importance were unrelated, the joint would be 0.61 times 0.60. It is noticeably larger, so supporters were more likely than average to care about the ruling.

\[ (0.61)(0.60) = 0.366 < 0.4575 \]

Check the answer against both ceilings

Why: the joint cannot exceed either single probability, and 0.4575 is below 0.61 and below 0.60.

\[ 0.4575 \le 0.60 \le 0.61 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3 homework, p. 208 — Exercise 80(g)

18. An AND probability can never beat either event alone

Concept

Every outcome in A AND B is also in A, and also in B. So the intersection is a piece of each event, and a piece can never be bigger than the whole.

\[ P(A\text{ AND }B) \le P(A) \quad \text{and} \quad P(A\text{ AND }B) \le P(B) \]

The multiplication rule shows the same thing algebraically. The second factor is a probability, so it is at most 1, and multiplying by a number no bigger than 1 can only shrink or keep.

\[ P(A\text{ AND }B) = P(A)\,P(B \mid A) \le P(A) \cdot 1 = P(A) \]

Use this as a free check on every AND answer. It catches the two most common slips: dividing where you should multiply, and adding where you should multiply. Either one tends to produce an answer bigger than one of the ingredients.

It also becomes a check on the data itself. If a problem hands you numbers whose AND probability comes out bigger than one of the single probabilities, the problem is inconsistent, however official it looks. You will meet exactly that in one of the book's own exercises later in this lesson.

19. When independence simplifies it

Section

Part 2

20. Independence turns the conditional into a plain probability

Concept

Independence has a precise meaning from Section 3.2: knowing B happened does not change the chance of A.

\[ A, B \text{ independent} \iff P(A \mid B) = P(A) \]

Now take the general rule, which is always true, and substitute that equality into it. The conditional factor is replaced by the plain one.

\[ P(A\text{ AND }B) = P(B)\,P(A \mid B) \]

\[ P(A\text{ AND }B) = P(B)\,P(A) \]

So the familiar product rule is not a separate fact to memorise. It is the general rule with one factor simplified, and it is valid exactly when that simplification is valid: when the events are independent.

multiplication rule, independent case — If A and B are independent, the probability of A AND B equals the probability of A times the probability of B.

When do you know? Sometimes the problem says so outright, or the physical setup guarantees it: separate coin flips, draws with replacement, spins of a fair wheel. Sometimes you have to check it from data, which is what the three tests coming up are for.

Keep the direction straight. You may use the plain product after you know the events are independent. You may not use it to make them independent. The next slides turn the equation around and use it as a test.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 179 — the independent special case

21. What independence looks like in the square

Picture it

Go back to the area picture, now for two shots. Split the square by the first shot, then shade the chance of making the second shot inside each column.

Figure (svg): Two unit squares side by side. Left, labelled 'no streaks': both columns shaded to the same height 0.65, so the shaded band is level. Right, labelled 'Carlos, streaky': the make column shaded to 0.90 and the miss column only to 0.186, so the band steps down.

Shaded height is the chance of making the second shot. On the left it does not care what happened first, so the band is level; on the right it jumps after a make. A level band is what independence looks like.

On the left is a shooter with no streaks. The chance of a make on the second shot is 0.65 whichever way the first shot went, so the shaded band sits at the same height in both columns. That level band is independence, drawn.

On the right is Carlos. After a make, the band jumps to 0.90; after a miss it drops to about 0.186. The band steps, and a stepped band means the first shot carries information about the second: dependence.

Where does 0.186 come from? His overall second-shot rate is 0.65, of which 0.585 comes through the make column, so the remaining 0.065 has to come through the miss column of width 0.35. Dividing gives 0.065 over 0.35, about 0.186.

\[ P(B \mid A') = \frac{0.65 - 0.585}{0.35} = \frac{0.065}{0.35} \approx 0.186 \]

Notice that both squares have the same total shaded area, 0.65, because both shooters make the second shot 65 percent of the time overall. What differs is how that area is distributed between the columns. Independence is not about how likely an event is; it is about whether its likelihood shifts when you learn something.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — Example 3.15

22. Reading the three equivalent tests for independence

Notation

Section 3.2 listed three ways to show independence. With the multiplication rule in hand you can see why they are the same test in three costumes.

Annotate

On: \( P(A \mid B) = P(A) \iff P(A\text{ AND }B) = P(A)\,P(B) \)

  • Learning that B happened leaves the chance of A unchanged. Use this one when a conditional with B to the right of its bar is given or easy to count.
  • P(B given A) = P(B) is the same statement with the roles swapped, and it is equivalent too. Carlos's problem gives this conditional, so it is the natural test there.
  • Plug either conditional equality into the general multiplication rule and this comes out. It is the most convenient test when you are given the joint probability, or counts in a table.
  • If and only if: when one of the three holds they all hold, and when one fails they all fail. You only ever need to check one.

The choice of test is practical, not logical. Pick the one whose ingredients the problem hands you. Example 3.15 gives a conditional, so compare conditionals or compute a product; Example 3.16 gives counts, so the product is the fastest.

Why must they agree? If the chance of A given B equals the chance of A, the multiplication rule turns into the product, and dividing the product by the probability of A gives the chance of B given A equal to the chance of B. Each form can be rewritten as the others in one line.

One warning about rounding. When the numbers are decimals rounded to three places, two sides of a test can differ in the last digit even for independent events. Compare with a little sense: 0.585 against 0.4225 is a real difference, 0.2667 against 0.2666 is rounding.

23. Example 3.15(c): are Carlos's two shots independent?

Worked example

Use the product test: compute what independence would predict and compare it with the joint probability you found in part (a).

\[ P(A) = 0.65, \quad P(B) = 0.65, \quad P(A\text{ AND }B) = 0.585 \]

The logic of a test is worth stating plainly. If the events were independent, the joint probability would have to equal the plain product. So compute the product and see whether the real joint matches it. A match means independent; a clear mismatch means dependent.

Compute what independence would predict

Why: multiply the two plain probabilities.

\[ P(A)\,P(B) = (0.65)(0.65) \]

Multiply

Why: the book prints 0.423, which is this value rounded to three places.

\[ P(A)\,P(B) = 0.4225 \]

Compare with the actual joint probability

Why: the two numbers are far apart, not a rounding wobble.

\[ P(A\text{ AND }B) = 0.585 \ne 0.4225 \]

So the events are not independent. In context: whether Carlos makes the first goal changes his chance of making the second, which is exactly what shooting in streaks means.

Check with the conditional test

Why: the conditional test must agree with the product test, and it does: the chance of the second goal after a make is 0.90, not the plain 0.65.

\[ P(B \mid A) = 0.90 \ne 0.65 = P(B) \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — Example 3.15(c)

24. Example 3.16(a) to (c): counting the swim team

Worked example

A swim team has 150 members: 75 advanced, 47 intermediate and the rest novice. Forty advanced, thirty intermediate and ten novice swimmers practise four times a week. One member is chosen at random.

Figure (svg): A grid of 150 dots, 15 to a row. The first five rows are blue (advanced), the next 47 dots orange (intermediate), the last 28 purple (novice). Within each colour the first 40, 30 and 10 dots are filled to show the swimmers who practise four times a week; the rest are hollow.

One dot per swimmer. Every probability in Example 3.16 is a count of dots over 150, so you can check each answer by counting.

Because the member is chosen at random, every swimmer is equally likely, and each probability is simply a count of dots divided by 150. That makes this example a good place to see the rules working on real counts, where you can check every answer by looking.

Count the novices

Why: everyone who is not advanced or intermediate.

\[ 150 - 75 - 47 = 28 \]

(a) Probability of a novice

Why: equally likely members, so count over total.

\[ P(\text{novice}) = \frac{28}{150} \approx 0.1867 \]

Count who practises four times a week

Why: add the filled dots of all three colours.

\[ 40 + 30 + 10 = 80 \]

(b) Probability of practising four times

Why: eighty filled dots out of 150.

\[ P(\text{four}) = \frac{80}{150} \approx 0.5333 \]

(c) Probability of advanced AND four times

Why: the filled blue dots only.

\[ P(\text{adv}\text{ AND }\text{four}) = \frac{40}{150} \approx 0.2667 \]

Part (c) could also be done with the multiplication rule: half the team is advanced, and 40 of the 75 advanced swimmers practise four times. The two routes must agree.

Check part (c) with the multiplication rule

Why: the probability of advanced times the chance of four times within the advanced group gives the same number as counting.

\[ \frac{75}{150} \cdot \frac{40}{75} = \frac{40}{150} \approx 0.2667 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 181 — Example 3.16(a)-(c)

25. Example 3.16(e): are novice and practising four times independent?

Worked example

The book asks whether being a novice and practising four times a week are independent. Use the product test with the counts from the last slide.

The actual joint probability

Why: ten novices practise four times a week.

\[ P(\text{nov}\text{ AND }\text{four}) = \frac{10}{150} \approx 0.0667 \]

What independence would predict

Why: multiply the two single probabilities.

\[ P(\text{nov})\,P(\text{four}) = \frac{28}{150} \cdot \frac{80}{150} \]

Evaluate the product

Why: the fraction is 2240 over 22500.

\[ \frac{2240}{22500} \approx 0.0996 \]

Compare

Why: the actual joint is two thirds of what independence predicts.

\[ 0.0667 \ne 0.0996 \]

Is a gap from 0.0667 to 0.0996 big enough to matter, or just rounding? Here the numbers come from exact counts, so there is no rounding to blame: if the events were independent, about 15 of the 150 swimmers would be novices who practise four times, and in fact only 10 are.

\[ 150 \times 0.0996 \approx 14.9 \quad \text{expected, against } 10 \text{ observed} \]

Figure (svg): A bar chart of the chance of practising four times a week for each level: advanced 0.533, intermediate 0.638, novice 0.357, with a dashed line across at 0.533 for the whole team.

Where a bar meets the dashed line, knowing the level tells you nothing about practice: independence. The novice bar falls well short of it, so novice and practising four times are dependent.

The bar chart tells the same story in conditional form. Among novices only 10 of 28, about 36 percent, practise four times, against 53 percent of the whole team. In context: novices practise four times a week noticeably less often than the team as a whole, so knowing a swimmer is a novice changes the chance.

Check with the advanced swimmers, where the answer flips

Why: for advanced swimmers the conditional equals the team rate exactly, so being advanced and practising four times are independent, and the product test agrees to the last digit.

\[ \frac{75}{150} \cdot \frac{80}{150} = \frac{40}{150} = P(\text{adv}\text{ AND }\text{four}) \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 181 — Example 3.16(e)

26. One question, three tests: fill in the table

Comparison

Each row is a pair of events from this section. Fill the blanks, then read across: the three columns always agree, because they are the same test in three forms.

Comparison matrix

eventsP(A given B)P(A)independent?
Felicity: math given speech0.250.2no
Carlos: 2nd goal given 1st0.900.65no
swim: four times given advanced0.5330.533yes
swim: four times given novice0.3570.533no

The Felicity row is Example 3.17(c). Math given speech is 0.25 but math overall is 0.2, so taking speech makes math a little more likely, and the events are dependent. The product test agrees: 0.2 times 0.65 is 0.13, not 0.1625.

The Carlos row runs the conditional test instead of the product test. The two always agree, so pick whichever needs the fewest new calculations: here the conditional 0.90 is given outright, so comparing it with 0.65 is immediate.

The advanced-swimmer row is the surprise of the section. Forty of 75 is exactly 80 of 150, so the conditional equals the plain probability and the events are independent. The same team, the same practice question, and the answer depends on which level you look at.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, pp. 180-182 — Examples 3.15(c), 3.16(e), 3.17(c)

27. Exercise 82(g): even numbers and the first dozen at roulette

Worked example

An American roulette wheel has 38 pockets: 1 to 36, plus 0 and 00, which count as neither even nor odd. Are the events even and first dozen, the numbers 1 to 12, independent?

Figure (svg): The 38 pockets of an American roulette wheel laid out as on the betting table: 0 and 00 in green on the left, then 1 to 36 in twelve columns of three. Even numbers are shaded blue. A thick yellow box surrounds 1 to 12, the first dozen.

Inside the yellow box exactly half the numbers are shaded. Across the whole wheel slightly fewer than half are, because the two green pockets pay out on neither.

Your instinct probably says independent: being even and being small feel unrelated. The table is the place to test an instinct. Three counts will settle it, and you can read each one straight off the layout.

Probability of even

Why: 18 of the numbers 1 to 36 are even, and neither green pocket counts.

\[ P(E) = \frac{18}{38} \approx 0.4737 \]

Probability of the first dozen

Why: the twelve numbers 1 to 12.

\[ P(D) = \frac{12}{38} \approx 0.3158 \]

Probability of both

Why: the even numbers inside the yellow box: 2, 4, 6, 8, 10, 12.

\[ P(E\text{ AND }D) = \frac{6}{38} \approx 0.1579 \]

What independence would predict

Why: the product of the two single probabilities.

\[ P(E)\,P(D) = \frac{18}{38} \cdot \frac{12}{38} = \frac{216}{1444} \approx 0.1496 \]

Compare the two lines. Six thirty-eighths against two hundred sixteen over fourteen hundred forty-four: put both over the same denominator and the first is 228 over 1444, bigger than 216 over 1444.

The two numbers are close but not equal, and they are exact fractions, so this is not rounding. In context: learning that the ball landed in the first dozen raises the chance it is even from about 47 percent to exactly half, because the first dozen contains no green pockets.

Check by removing the green pockets

Why: on a wheel with only 1 to 36, even has probability one half and the first dozen one third, and the product equals the joint exactly: the zeros alone caused the dependence.

\[ \frac12 \cdot \frac13 = \frac16 = \frac{6}{36} \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3 homework, p. 208 — Exercise 82(g)

28. May you multiply the plain probabilities?

Sorting

For each pair of events, decide whether the plain product gives the AND probability, or whether you need a conditional. Think about whether the first event changes the chance of the second.

Sort into buckets

Sort each pair by whether the plain product is valid for the probability of both.

Independent: the plain product works
two flips of a fair coin: heads, then heads; two balls drawn with replacement: white, then white; swim team: advanced, and practises four times a week
Dependent: use the conditional
two cards dealt without replacement: ace, then ace; Carlos's two goals, when he shoots in streaks; roulette: even, and in the first dozen
prod
A coin has no memory, a replaced ball restores the box exactly, and the advanced swimmers practise at exactly the team rate of 80 out of 150. In each case the conditional equals the plain probability, so the product is the general rule with nothing lost.
cond
Dealing an ace leaves 3 aces in 51 cards, not 4 in 52. Carlos is streaky. The first dozen contains no green pocket, so even is more likely inside it. In each case the second factor must be the conditional.

The card case is worth one line of arithmetic. Two aces without replacement is four fifty-seconds times three fifty-firsts, about 0.00452; the plain product would give four fifty-seconds squared, about 0.00592, too large because it pretends the first ace went back in the deck.

\[ \frac{4}{52} \cdot \frac{3}{51} \approx 0.00452 \quad \text{vs} \quad \left(\frac{4}{52}\right)^2 \approx 0.00592 \]

The roulette pair is the one most people sort wrongly, and it is worth knowing why. The dependence there is small, 0.1579 against 0.1496, and it comes entirely from the two green pockets. Small dependence is still dependence: the product rule would give a slightly wrong answer, and a casino lives on slightly wrong answers.

Notice that two of the independent cases, the coin and the replaced ball, are independent by physics, while the swimmers are independent only because the counts happen to line up. Independence can be built into a situation or discovered in data; either way it has to be established before the product is used.

29. The addition rule

Section

Part 3

30. Adding two probabilities counts the overlap twice

Concept

Now the other question: the chance that at least one of two events happens, A OR B. In probability, OR is inclusive. It means A, or B, or both, which is why the obvious move of adding the two probabilities goes wrong.

Figure (svg): A rectangle for all outcomes of Carlos's two shots, containing two overlapping circles: A, first goal, and B, second goal. The region in A only is labelled 0.065, the overlap 0.585, the region in B only 0.065, and the outside 0.285.

Circle A holds 0.065 plus 0.585; circle B holds 0.585 plus 0.065. Add the two circles and the middle 0.585 has been counted twice.

Look at Carlos's two goals as regions. Circle A is everything where he makes the first goal: the 0.065 where he makes only the first, plus the 0.585 where he makes both. Circle B is 0.585 plus 0.065 the other way round. The union is the three pieces inside the circles, each counted once.

\[ P(A\text{ OR }B) = 0.065 + 0.585 + 0.065 \]

Adding the two circles instead adds the middle piece twice, once as part of A and once as part of B.

\[ P(A) + P(B) = (0.065 + 0.585) + (0.585 + 0.065) \]

So the fix is to take the double-counted piece away once. That piece is the intersection, the AND probability, which the multiplication rule has just taught you to find.

\[ P(A\text{ OR }B) = P(A) + P(B) - P(A\text{ AND }B) \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 179 — The Addition Rule

31. Reading the addition rule piece by piece

Notation

Like the multiplication rule, this one holds for every pair of events, with no assumption at all. Each term earns its place.

Annotate

On: \( P(A\text{ OR }B) = P(A) + P(B) - P(A\text{ AND }B) \)

  • At least one of the two happens. Inclusive: the outcomes where both happen belong here too. The words either, or and at least one all ask for this.
  • Each event's whole probability, overlap included. On its own this sum counts every both-outcome twice, and it can even exceed 1.
  • Removes the second copy of the overlap. Subtract it once, not twice: after the subtraction the overlap is still counted, exactly once.
  • Often from the multiplication rule, as in Examples 3.15 and 3.17, so an OR question usually needs an AND answer first.

There is a useful asymmetry between the two rules. The multiplication rule needs one conditional and one plain probability. The addition rule needs two plain probabilities and one AND. So an OR problem usually runs through the multiplication rule on the way, while an AND problem never needs the addition rule.

The minus sign is the part people drop, and it is also the part people overdo. A common muddle is to subtract the intersection twice, reasoning that it was in both events. But the goal is to count it once, and it was counted twice, so one subtraction is exactly right.

You can confirm the logic with the extreme case where A and B are the same event. Then the rule gives the probability of A plus the probability of A minus the probability of A, which is the probability of A, as it must be.

32. Example 3.15(b): Carlos makes the first goal or the second

Worked example

Find the probability that Carlos makes either the first goal or the second, using the joint probability from part (a).

\[ P(A) = 0.65, \quad P(B) = 0.65, \quad P(A\text{ AND }B) = 0.585 \]

The word either can sound as if it means one or the other but not both. In probability it never does. Unless a problem says exactly one, an OR includes the both case, and that is the case the addition rule must not count twice.

Name what is asked

Why: either goal means at least one: the inclusive OR.

\[ \text{find } P(A\text{ OR }B) \]

Write the addition rule

Why: the general form, since the events clearly can both happen.

\[ P(A\text{ OR }B) = P(A) + P(B) - P(A\text{ AND }B) \]

Substitute

Why: the two single probabilities and the overlap from part (a).

\[ P(A\text{ OR }B) = 0.65 + 0.65 - 0.585 \]

Add the two single probabilities

Why: on its own this 1.30 is impossible, which is the double count showing itself.

\[ 0.65 + 0.65 = 1.30 \]

Subtract the overlap once

Why: the answer is now a genuine probability.

\[ 1.30 - 0.585 = 0.715 \]

In context: Carlos makes at least one of the two goals with probability 0.715.

Check through the complement

Why: the only way to fail at least one goal is to miss both, the 0.285 path at the foot of the tree; one minus that must give the same answer, and it does.

\[ 1 - P(\text{miss both}) = 1 - 0.285 = 0.715 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — Example 3.15(b)

33. The rule in counts: the swim team again

Concept

The addition rule works just as well with counts as with probabilities, and counts make the double counting visible. Ask for the chance that a random swimmer is advanced OR practises four times a week.

There are 75 advanced swimmers and 80 who practise four times, but 40 people are in both lists. Add the lists and those 40 are counted twice.

\[ 75 + 80 - 40 = 115 \]

\[ P(\text{adv}\text{ OR }\text{four}) = \frac{115}{150} \approx 0.7667 \]

Now count directly, from the dot grid earlier. Every blue dot qualifies, 75 of them. Among the other colours, only the filled dots qualify: 30 intermediate and 10 novice. That is 115 again, with nobody counted twice.

\[ 75 + 30 + 10 = 115 \]

The same structure is in the probability version. Divide every count by 150 and you get the rule term by term: 75 over 150 plus 80 over 150 minus 40 over 150. The rule is just careful counting, divided through by the size of the sample space.

Counts also make the plain-sum error concrete. Seventy-five plus eighty is 155, more swimmers than the team has. No one would believe a team of 150 has 155 members who are advanced or practise often; the same absurdity hides in a probability sum that exceeds 1.

Counting this way is also the fastest route when a problem gives you a table, which is how Section 3.4 will present its data.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 181 — Example 3.16 data

34. Example 3.17(b): Felicity takes math or speech

Worked example

With the probabilities of math 0.2 and speech 0.65, and the joint probability 0.1625 from part (a), find the probability that Felicity takes math or speech.

Notice how the parts of the book's example chain together. Part (a) used the multiplication rule to find the overlap, and part (b) cannot be done without it. This is the most common shape for an OR question: the AND probability is not given, so you build it first.

Write the addition rule

Why: she can take both, so the general form is required.

\[ P(M\text{ OR }S) = P(M) + P(S) - P(M\text{ AND }S) \]

Substitute

Why: math 0.2, speech 0.65, both 0.1625.

\[ P(M\text{ OR }S) = 0.2 + 0.65 - 0.1625 \]

Add

Why: the double-counted total.

\[ 0.2 + 0.65 = 0.85 \]

Subtract the overlap once

Why: leaving each outcome counted once.

\[ 0.85 - 0.1625 = 0.6875 \]

Figure (svg): Three bars over a scale from 0 to 1. The top bar, math, runs from 0 to 0.2. The middle bar, speech, runs from 0.0375 to 0.6875, so it shares the stretch from 0.0375 to 0.2 with math. The bottom bar, math OR speech, runs from 0 to 0.6875. Dashed lines mark the shared stretch, labelled counted twice: 0.1625.

Lay the two probabilities end to end and they would reach 0.85; laid where they really fall, they share 0.1625 and together cover only 0.6875.

The bars show the answer as a length. Math covers 0.2, speech covers 0.65, and they share a stretch of 0.1625. Together they cover 0.6875 of the probability line. In context: Felicity takes at least one of the two classes with probability 0.6875, matching the book's answer.

Check by adding the three separate pieces

Why: math only is 0.2 minus 0.1625, speech only is 0.65 minus 0.1625, and those two plus the overlap must reproduce the union.

\[ 0.0375 + 0.1625 + 0.4875 = 0.6875 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 182 — Example 3.17(b)

35. Example 3.18: breast cancer and a negative test

Worked example

About one woman in seven who lives to 90 develops breast cancer: 0.143. Of those who do, the test is negative 2 percent of the time. Across all women the test is negative 85 percent of the time. B is developing cancer, N is testing negative.

(a) and (b): read off the given probabilities

Why: one in seven is 0.142857, which the book rounds to 0.143; the 2 percent is a conditional, measured among women with cancer.

\[ P(B) = 0.143, \quad P(N) = 0.85, \quad P(N \mid B) = 0.02 \]

The two percentages about negative tests are easy to confuse, so pin them down. The 85 percent is measured across all women. The 2 percent is measured only among women who have cancer, so it is the probability of N given B. Writing each number in notation is what keeps them apart.

(c) Multiplication rule for B AND N

Why: the conditional has B to the right of its bar, so B is the plain factor.

\[ P(B\text{ AND }N) = P(B)\,P(N \mid B) = (0.143)(0.02) \]

Multiply

Why: the book rounds this to 0.0029.

\[ P(B\text{ AND }N) = 0.00286 \]

(d) Addition rule for B OR N

Why: the events overlap, so the general form.

\[ P(B\text{ OR }N) = 0.143 + 0.85 - 0.00286 \]

Evaluate

Why: with the book's rounded 0.0029 you get its printed 0.9901.

\[ P(B\text{ OR }N) = 0.99014 \]

Figure (svg): One long bar standing for 100,000 women, split into four pieces: 14,014 with cancer who test positive, a sliver of 286 with cancer who test negative, 84,714 without cancer who test negative, and 986 without cancer who test positive. A bracket above the first two pieces is labelled cancer, 14,300; a bracket below the middle two is labelled tests negative, 85,000; a long bracket over the first three is labelled B OR N, 99,014.

Out of 100,000 women, only the 986 at the far right are in neither group. The tiny yellow sliver is the overlap the addition rule subtracts once.

The figure turns the probabilities into 100,000 women. The overlap is a sliver of 286, and almost everyone is in the union: the only women outside it are the 986 without cancer who test positive. Parts (e) and (f) follow at once: 0.02 is nothing like 0.85, so B and N are dependent, and 0.00286 is not zero, so they are not mutually exclusive.

Check the union by counting who is left out

Why: the complement of B OR N is no cancer AND a positive test; its size, 15,000 positives minus the 14,014 with cancer, gives the same answer.

\[ 1 - \frac{15000 - 14014}{100000} = 1 - 0.00986 = 0.99014 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, pp. 182-183 — Example 3.18

36. Two probabilities that add to more than one

Anomaly

Carlos's two goals each have probability 0.65, and 0.65 plus 0.65 is 1.30. A probability cannot be bigger than 1. Something has to give.

Predict first

Any two events with probability 0.65 each, whatever they describe, must have an overlap of at least

  • 0
  • 0.30
  • 0.4225
  • 0.65

Correct: 0.30

Why: The union can be at most 1, so 1.30 minus the overlap must be at most 1, which forces the overlap to be at least 0.30. The 0.4225 would be the overlap only if the events were independent, and 0.65 only if they were the same event; but the smallest the overlap can ever be is 0.30.

Here is the algebra. The addition rule says the union is the sum minus the overlap, and the union is a probability, so it is at most 1.

\[ P(A) + P(B) - P(A\text{ AND }B) \le 1 \]

\[ P(A\text{ AND }B) \ge P(A) + P(B) - 1 = 0.30 \]

This connects to the homework question about two probabilities that add to 1.2. By the same two lines, their overlap is at least 0.2, so the events are certainly not exclusive, and you know that without being told anything else about them.

Carlos's actual overlap, 0.585, clears this floor easily. The point is what the floor rules out: two events whose probabilities add to more than 1 cannot be mutually exclusive, whatever else is true. They are forced to overlap by at least the excess over 1.

37. Slide the two probabilities and watch the union

Tweak it

The line below is the addition rule drawn as a graph. Across the bottom is the overlap, the AND probability; up the side is the union it produces. Every legitimate answer has to sit at height 1 or below.

Parameter explorer

Set both sliders to 0.65. Read off the union when the overlap is 0, and find the overlap at which the line comes down to height 1. Then lower both sliders to 0.4 and say what changes.

\[ P(A\text{ OR }B) = {a} + {b} - P(A\text{ AND }B) \]

  • a — from 0 to 1: P(A)
  • b — from 0 to 1: P(B)

At 0.65 each, the line starts at 1.30 when the overlap is zero, which is above the ceiling, so zero overlap is impossible. It comes down to exactly 1 when the overlap reaches 0.30, the floor you found on the last slide.

Try Carlos's real overlap, 0.585, on the horizontal axis: the line is at 0.715 there, his answer from Example 3.15(b). And try Felicity's numbers, 0.2 and 0.65, with the overlap at 0.1625: the height is 0.6875. Every OR answer in this section is a point on one of these lines.

At 0.4 each, the line starts at 0.8, safely below 1, so these events could be mutually exclusive. The line also cannot run past an overlap equal to the smaller probability, since an intersection is never bigger than either event: that is the ceiling from Part 1. Between the floor and that ceiling, every overlap is possible, and the data decide which one is true.

38. Trap: adding without subtracting

Trap

The trap

Asked for the chance that Carlos makes at least one goal:

\[ P(A\text{ OR }B) = P(A) + P(B) \]

\[ = 0.65 + 0.65 = 1.30 \]

Wrong, and it announces itself: no probability is bigger than 1.

The instinct behind it is that or means add, which is true for counting separate things. It fails as soon as the things can coincide.

The fix

The plain sum is the addition rule only when the events cannot happen together. Carlos can make both goals, so the overlap was counted twice and must come off once:

\[ 0.65 + 0.65 - 0.585 = 0.715 \]

The dangerous version of this error is the one that does not give an answer above 1. Felicity's plain sum is 0.85, which looks perfectly respectable, yet the true answer is 0.6875. Always ask whether the events can overlap before dropping the subtraction.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, pp. 180-182 — Examples 3.15(b) and 3.17(b)

39. Exercises 44 to 53: the book's own numbers are impossible

Error analysis

The section's practice exercises describe a restaurant: 87 percent of customers order pizza, 32 percent order salad, and of the customers who order pizza, 55 percent also order a salad. Exercise 48 asks for the probability of both. Look at what the multiplication rule returns.

Annotate

On: \( P(Z\text{ AND }S) = (0.87)(0.55) = 0.4785 > 0.32 = P(S) \)

  • The multiplication rule, correctly applied: the conditional is salad given pizza, so pizza is the plain factor.
  • Nearly 48 percent of all customers would order both a pizza and a salad.
  • But only 32 percent of customers order a salad at all. The customers who order both are a subset of those who order salad, so this is impossible.
  • For salad to be 0.32 overall, salad given pizza can be at most 0.32 divided by 0.87, about 0.368. The printed 0.55 cannot be right together with the other two numbers.

Figure (svg): A bar chart with three bars: P(Z), pizza, 0.87; P(S), salad, 0.32; and in red the value the multiplication rule gives for P(Z AND S), 0.4785. A dashed line at 0.32 marks the ceiling that the intersection may not pass; the red bar pokes above it.

Customers who order both are a subset of customers who order salad, so the red bar can never stand taller than the salad bar. With the book's numbers it does.

You can see the impossibility without any rule at all. Picture 100 customers. Only 32 order a salad. But 87 order pizza, and 55 percent of those 87, about 48 people, are supposed to order a salad as well. Forty-eight salad orders cannot come from a group of 32 salad eaters.

This is the ceiling check from Part 1 earning its keep. The arithmetic is correct; the data are inconsistent, and no answer to the later parts can be trusted. The same exercise set also asks whether P and S are independent, a misprint for Z and S. Statistics is partly the habit of asking whether numbers could possibly be true before computing with them.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3 practice, p. 204 — Exercises 44-53

40. When mutual exclusivity simplifies it

Section

Part 4

41. No overlap, nothing to subtract

Concept

Mutually exclusive events, from Section 3.2, cannot happen together: their intersection is empty, so its probability is zero.

\[ A, B \text{ mutually exclusive} \iff P(A\text{ AND }B) = 0 \]

Take the general addition rule, always true, and substitute that zero into it. The subtracted term vanishes.

\[ P(A\text{ OR }B) = P(A) + P(B) - 0 \]

\[ P(A\text{ OR }B) = P(A) + P(B) \]

Figure (svg): A rectangle for Klaus's choice containing two separate circles that do not touch: A, New Zealand, labelled 0.6, and B, Alaska, labelled 0.35. The space outside both is labelled no vacation: 0.05.

Nothing sits in both circles, so adding the two circles counts every outcome exactly once. There is no overlap to subtract.

In the picture the two circles do not touch, so adding their probabilities counts every outcome exactly once: there is no middle piece to double count. The shortcut is valid exactly when the circles are apart.

The condition has to come from the situation, not from a wish. Klaus has money for one trip; a swimmer has one level; a die shows one face. Sentences like these are what establish exclusivity. Two events that merely seem different, like taking math and taking speech, can still overlap.

addition rule, mutually exclusive case — If A and B cannot both happen, the probability of A OR B equals the probability of A plus the probability of B.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 179 — the mutually exclusive special case

42. Example 3.14: Klaus chooses a vacation

Worked example

Klaus can afford only one vacation. He picks New Zealand with probability 0.6 and Alaska with probability 0.35. Find the probability that he picks one of the two, and the probability that he goes nowhere.

\[ P(A) = 0.6, \quad P(B) = 0.35 \]

The key sentence in the problem is the one about money, not the one with numbers in it. Klaus can afford only one vacation. That sentence is what tells you the two events cannot both happen, and it is the only thing that entitles you to drop the subtraction.

Establish the overlap

Why: he can take only one trip, so both is impossible; this is a fact about the situation, not an assumption for convenience.

\[ P(A\text{ AND }B) = 0 \]

Write the addition rule in its general form

Why: start from the safe form, then simplify.

\[ P(A\text{ OR }B) = P(A) + P(B) - P(A\text{ AND }B) \]

Substitute

Why: including the zero overlap.

\[ P(A\text{ OR }B) = 0.6 + 0.35 - 0 \]

Add

Why: nothing to subtract.

\[ P(A\text{ OR }B) = 0.95 \]

In context: Klaus goes to New Zealand or Alaska with probability 0.95. Writing the zero into the general rule, rather than jumping straight to the plain sum, costs one line and records why the subtraction disappeared.

The book adds one more observation: the probability that Klaus goes on neither trip. Going nowhere is the complement of going somewhere.

Check that the whole sample space adds to 1

Why: New Zealand, Alaska and no vacation are the only three possibilities and no two can happen together, so their probabilities must total 1.

\[ 0.6 + 0.35 + 0.05 = 1 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, pp. 179-180 — Example 3.14

43. Try It 3.14: Anna buys one car

Elimination

Anna can afford only one car. She buys car A with probability 0.25 and car B with probability 0.65. The Try It asks for the probability of A AND B and of A OR B. Knock out the lines that are wrong.

Eliminate the wrong options

Which line correctly answers both parts of Try It 3.14?

  • k. P(A AND B) = 0 and P(A OR B) = 0.25 + 0.65 = 0.90
  • p. P(A AND B) = (0.25)(0.65) = 0.1625
  • s. P(A OR B) = 0.25 + 0.65 − 0.1625 = 0.7375
  • c. P(A OR B) = 1, since she must buy one of them

Survives elimination: k

Why: Only one car is affordable, so buying both is impossible and its probability is zero. With no overlap, the addition rule loses its subtraction and the answer is 0.25 plus 0.65, which is 0.90. The remaining 0.10 is the chance she buys neither car.

The tempting wrong line is the product, 0.1625. It comes from treating the two purchases as independent, but independence and mutual exclusivity are different properties, and here the situation settles which one applies: she cannot do both.

The line that subtracts 0.1625 makes two errors at once: it invents an overlap by the product rule, then subtracts it. Each step is a correct rule used where its condition fails, which is exactly the mistake this lesson is designed to prevent.

The last wrong line is a quieter mistake: assuming the two named options are the only possibilities. Like Klaus with his 0.05, Anna has probability left over for doing neither.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — Try It 3.14

44. Example 3.16(d): advanced and intermediate at once

Worked example

Part (d) asks for the probability that a swimmer is advanced and intermediate, and whether those events are mutually exclusive. Then use the answer to find an OR probability the fast way.

Count the swimmers in both groups

Why: each member has exactly one level, so nobody is in both.

\[ P(\text{adv}\text{ AND }\text{int}) = \frac{0}{150} = 0 \]

Conclude

Why: a zero intersection is the definition of mutually exclusive.

\[ \text{advanced and intermediate are mutually exclusive} \]

In context: a swimmer cannot be advanced and intermediate at the same time. Now use that. What is the chance a random member is advanced or novice? Those are also different levels, so they are mutually exclusive too.

Use the mutually exclusive addition rule

Why: no overlap, so just add.

\[ P(\text{adv}\text{ OR }\text{nov}) = \frac{75}{150} + \frac{28}{150} \]

Add

Why: same denominator.

\[ P(\text{adv}\text{ OR }\text{nov}) = \frac{103}{150} \approx 0.6867 \]

Compare this with advanced OR practises four times, which needed the subtraction of 40 shared swimmers. The difference is not in the arithmetic but in the events: levels are exclusive by construction, while a level and a practice habit can overlap. The rule you use follows from what the events are.

Check with the complement

Why: not advanced or novice means intermediate, and one minus the intermediate probability must match.

\[ 1 - \frac{47}{150} = \frac{103}{150} \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 181 — Example 3.16(d)

45. Try Its 3.16 and 3.18: 200 seniors and their plans

Worked example

A school has 200 seniors: 140 going to college, 40 going to work, the rest taking a gap year. Fifty of the college-bound, thirty of the workers and five of the gap-year seniors play sports.

Figure (svg): Three horizontal bars drawn to scale, one per plan: college 140 seniors, work 40, gap year 20. The darker left part of each bar is the seniors who play sports: 50, 30 and 5.

The three bars cannot overlap: every senior has exactly one plan. That is what lets the three plan probabilities simply add.

Before computing, notice that the sports numbers are irrelevant to the first question. Try It 3.16 is only about plans, and the three plans are mutually exclusive and cover every senior. That makes the complement the natural route.

Try It 3.16: probability of a gap year

Why: the plans are mutually exclusive, so college OR work is a plain sum, and a gap year is its complement.

\[ P(\text{gap}) = 1 - \left(\frac{140}{200} + \frac{40}{200}\right) = 1 - 0.9 \]

Simplify

Why: one tenth of the seniors.

\[ P(\text{gap}) = 0.1 \]

Try It 3.18: college AND sports

Why: fifty seniors are both; count over the total.

\[ P(\text{col}\text{ AND }\text{sp}) = \frac{50}{200} = 0.25 \]

One more: college OR sports

Why: these overlap, so the general rule, with 85 players in all.

\[ P(\text{col}\text{ OR }\text{sp}) = \frac{140 + 85 - 50}{200} = \frac{175}{200} = 0.875 \]

In context: a tenth of the seniors are taking a gap year, a quarter are college-bound athletes, and seven in eight are college-bound, athletes, or both.

Notice the contrast inside one problem. The three plans never overlap, so their probabilities just add. College and sports overlap in 50 people, so their sum has to lose those 50 once.

Check the gap-year answer by counting

Why: 200 minus 140 minus 40 leaves 20 seniors, and 20 over 200 is one tenth.

\[ \frac{200 - 140 - 40}{200} = \frac{20}{200} = 0.1 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, pp. 181-183 — Try Its 3.16 and 3.18

46. Exclusive events are never independent

Intuition

Independence and mutual exclusivity sound like they might be the same thing, since both say the events have nothing to do with each other. They are close to opposites.

Suppose A and B are mutually exclusive, each with a positive probability. Then the moment you learn A happened, you know for certain B did not. Learning A changed the chance of B from something positive to zero. That is the opposite of independence.

\[ P(B \mid A) = 0 \ne P(B) \]

The product test says the same thing. Mutual exclusivity makes the joint zero, while the product of two positive probabilities is positive, so the two can never be equal.

\[ P(A\text{ AND }B) = 0 < P(A)\,P(B) \]

Klaus is the example: if he goes to New Zealand, the chance he goes to Alaska drops from 0.35 to zero. Mutually exclusive events are strongly dependent. The one loophole is an event with probability zero, which is independent of everything and exclusive of everything, and which never matters in practice.

47. Trap: treating independent and exclusive as the same

Trap

The trap

Example 3.15(d) asks whether Carlos's two goals are mutually exclusive. A common answer:

\[ \text{No, they are not independent,} \]

\[ \text{so they are not mutually exclusive} \]

Wrong reasoning, even though the verdict happens to be right.

The mistake is treating the two words as synonyms for unrelated. Run the same reasoning on two coin flips and it gives a wrong verdict.

The fix

Each property has its own test, and neither test mentions the other property. Mutually exclusive is about the intersection being zero:

\[ P(A\text{ AND }B) = 0.585 \ne 0 \]

That settles it: he can make both goals, so they are not mutually exclusive. Independence is a separate question with a separate test, and events can be independent and overlapping, like two coin flips, or exclusive and dependent, like Klaus's trips.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — Example 3.15(d)

48. Can independent events happen together?

Counterexample

Someone claims that independent events can never both happen, because independent means they have nothing to do with each other.

Discussion prompt

Give a counterexample using two flips of a fair coin, and show with numbers that your two events are independent and overlap.

The root of the confusion is the everyday phrase has nothing to do with. In probability, independence means one event carries no information about the other. Two events can carry no information about each other and still happen at the same time, just as the weather in two distant cities can both be sunny today.

Keep this pair in your head as the standard example of independent but overlapping, just as Klaus is the standard example of exclusive but dependent. Between them they show that neither property implies the other in the direction people expect.

The swim team supplies a third case worth remembering: advanced and practising four times are independent and overlap in 40 swimmers, a real-data version of the coin flips.

49. Match each pair of events to its relationship

Matching

For each pair, decide whether the events overlap and whether one changes the chance of the other. The two questions are separate.

Match the pairs

  • a. Klaus: New Zealand, Alaska
  • b. fair coin: heads on flip 1, heads on flip 2
  • c. Carlos: first goal, second goal
  • d. swim team: novice, not novice
  • w. exclusive, and therefore dependent; another outcome is possible
  • x. independent, and they overlap
  • y. neither: they overlap and are dependent
  • z. exclusive, and together they are certain

Why: Klaus cannot take both trips, and he might take neither, so the trips are exclusive but do not cover everything. The coin flips overlap in one outcome of four and pass the product test. Carlos's goals overlap in 0.585 and fail the product test. Novice and not novice are complements: exclusive, and between them certain.

The last pair is a special kind of exclusive pair, a complement pair. It is exclusive and exhaustive, so its two probabilities add to exactly 1, which is the complement rule. Klaus's pair is exclusive but not exhaustive, which is why his two probabilities leave 0.05 over.

Carlos is the pair that shows both properties can fail at once. His goals overlap, since he makes both 58.5 percent of the time, and they are dependent, since the streak lifts the second shot after a make. Neither property is the default; each has to be checked.

If you found yourself matching by gut feeling, go back to the two tests. Overlap is the question: is the AND probability zero? Dependence is the question: does the conditional differ from the plain probability?

50. Neither is the complement of OR

Concept

Many problems ask for the chance that neither event happens. Neither is exactly the outcome that falls outside the union, so it is one minus the OR probability.

\[ P(\text{neither}) = 1 - P(A\text{ OR }B) \]

Klaus's 0.05 is this: one minus 0.95. For Carlos, one minus 0.715 gives 0.285, which is exactly the miss-both path at the foot of his tree. For Felicity, one minus 0.6875 gives 0.3125, the chance she takes neither math nor speech.

\[ 1 - 0.715 = 0.285, \qquad 1 - 0.6875 = 0.3125 \]

Read the other way, this gives a useful shortcut for at least one. When the chance of none is easy to find, as it often is with repeated independent trials, subtract it from 1 rather than using the addition rule.

A warning about words. Neither A nor B is the complement of A OR B. It is not the complement of A AND B: not both includes the outcomes where exactly one happens. For Carlos, not both is one minus 0.585, which is 0.415, a very different number from 0.285.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 180 — the note after Example 3.14

51. Choosing the rule

Section

Part 5

52. The wording tells you which rule

Concept

Every question in this section is one of three shapes, and the shape is announced by the words. Learn to hear them before you look at any numbers.

Figure (svg): A flow diagram with three rows. Both or AND leads to the multiplication rule, and then, only if the events are independent, to the plain product. Either, OR or at least one leads to the addition rule, and then, only if the events are mutually exclusive, to the plain sum. Not or neither leads to one minus the probability of the event.

The wording picks the row; the general rule in the middle is always safe; the shortcut on the right needs a reason.

Both, and, all of: the multiplication rule, which needs a conditional. Either, or, at least one: the addition rule, which needs an AND probability, so often the multiplication rule first. Not, neither, fails to: one minus the probability of the event described.

Then comes the second decision, in the right-hand column. Is there a reason to simplify? Independence must be stated or checked before the conditional becomes a plain probability. Mutual exclusivity must follow from the situation, like one vacation or one level, before the overlap becomes zero.

Some words need care. At least one is an OR. Exactly one is not: it is the OR minus the AND, since the both case is excluded. Not both is the complement of the AND, while neither is the complement of the OR. Translate the phrase into notation before choosing a rule.

Finally, name the missing ingredient. An AND needs a conditional from the problem; an OR needs an AND. If the problem does not give you the ingredient, look for a way to derive it, from a complement, from counts, or from the other form of the rule.

53. The complement rule is the addition rule in disguise

Concept

Example 3.19 leans on complements, so it is worth seeing where the complement rule comes from. An event A and its complement, A prime, are mutually exclusive and together they cover the whole sample space.

\[ P(A\text{ AND }A') = 0, \qquad A\text{ OR }A' = \text{everything} \]

Apply the mutually exclusive addition rule. The union is certain, so its probability is 1.

\[ P(A) + P(A') = 1 \]

\[ P(A') = 1 - P(A) \]

The same works inside a conditional world. Among women with cancer, a test is either negative or positive, never both, so the two conditional probabilities add to 1. That is how Example 3.19 gets the chance of a positive test given cancer from the chance of a negative one.

\[ P(P \mid B) = 1 - P(N \mid B) = 1 - 0.02 = 0.98 \]

Be careful which way the conditional runs: the complement is taken on the left of the bar, never on the right. The chance of a negative test given no cancer is not one minus the chance of a negative test given cancer; those are measured in two different groups of women, and nothing forces them to add to 1.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 183 — Example 3.19(a) and (c)

54. Example 3.19: the positive side of the test

Worked example

Refer to Example 3.18. P now stands for testing positive. Find the four probabilities the book asks for, each with the rule it needs.

(a) Positive given cancer

Why: complement within the cancer branch.

\[ P(P \mid B) = 1 - P(N \mid B) = 1 - 0.02 = 0.98 \]

Part (a) is a complement taken inside the cancer group: a woman with cancer tests either negative or positive, so the two conditionals add to 1. Every later part reuses a number found earlier, so work them in order.

(b) Cancer AND positive

Why: the multiplication rule, with B as the plain factor.

\[ P(B\text{ AND }P) = P(P \mid B)\,P(B) = (0.98)(0.143) \]

Multiply

Why: the book rounds to 0.1401.

\[ P(B\text{ AND }P) = 0.14014 \]

(c) No cancer

Why: complement of B.

\[ P(B') = 1 - 0.143 = 0.857 \]

(d) A positive test

Why: complement of N.

\[ P(P) = 1 - 0.85 = 0.15 \]

Figure (svg): A probability tree. First branches: cancer B, 0.143, and no cancer, 0.857. After cancer: negative 0.02 and positive 0.98. After no cancer: negative 0.9885 and positive 0.0115. The path ends read 0.00286, 0.14014, 0.84714 and 0.00986.

The two positive paths add to 0.15, the book's positive rate. The lower branches were not given; they are what the given numbers force.

The tree puts all four answers in place and fills in the branches the book never gives. The lower branches come from dividing: the women without cancer who test positive are 0.15 minus 0.14014, and that piece divided by 0.857 is about 0.0115.

Check that the two positive paths add to the positive rate

Why: every positive test lies on one of the two P paths, so their total must be the 0.15 from part (d).

\[ 0.14014 + 0.00986 = 0.15 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, p. 183 — Example 3.19

55. What a positive result actually means

Real world

A woman takes this screening test and gets a positive result. She wants one number: the chance she actually has cancer. None of the book's four answers is that number, but the rules give it in one line.

Discussion prompt

Use the multiplication rule, rearranged, to find the probability of cancer given a positive test. Then say in one sentence what it means for the woman.

Notice how different this is from the 0.98 of Example 3.19(a). The chance of a positive test given cancer and the chance of cancer given a positive test are different questions, with different denominators. Confusing them is so common in medicine and law that it has a name, the prosecutor's fallacy.

The calculation is the multiplication rule read backwards: you know the AND probability and the probability of the condition, and you divide to get the conditional. It is the same move as the original definition in Section 3.1, and it is the heart of Bayes' theorem, which later courses build on exactly this tree.

The gap here is small because this population has a high rate of cancer, one in seven. For a rarer condition the same test would give a far smaller chance of disease after a positive result, because the false alarms from the huge healthy group would swamp the true positives. Chapter 3's contingency tables make that easy to see.

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, pp. 182-183 — Examples 3.18 and 3.19

56. Try Its 3.17(b) and 3.19: the library, every question

Worked example

Back to the library: a book with probability 0.40, a DVD with probability 0.30, and a DVD given a book 0.5. You already found the chance of both, 0.20.

3.17(b): book OR DVD

Why: they overlap, so the general addition rule.

\[ P(B\text{ OR }D) = 0.40 + 0.30 - 0.20 = 0.50 \]

Five more questions follow, and each one is a single rule applied once. Say the rule to yourself before each line: addition, complement, multiplication read backwards, subtraction of a piece, division by a piece.

3.19(a): no book

Why: complement.

\[ P(B') = 1 - 0.40 = 0.60 \]

3.19(b) and (c): both, then book given DVD

Why: AND does not care about order; divide the joint by the DVD probability for the conditional.

\[ P(D\text{ AND }B) = 0.20, \quad P(B \mid D) = \frac{0.20}{0.30} \approx 0.667 \]

3.19(d): DVD AND no book

Why: the DVD visits split into with-book and without-book, so subtract.

\[ P(D\text{ AND }B') = 0.30 - 0.20 = 0.10 \]

3.19(e): DVD given no book

Why: divide by the probability of no book.

\[ P(D \mid B') = \frac{0.10}{0.60} \approx 0.167 \]

Figure (svg): A unit square split into two columns: book, width 0.40, and no book, width 0.60. In the book column the bottom half is shaded, labelled B AND D = 0.20; in the no-book column a band of height one sixth is shaded, labelled 0.10.

The two shaded blocks together are all the DVD visits, 0.30. The left block is 0.20 of that, which is why a DVD borrower has taken a book two times in three.

In the mosaic, the DVD visits are the two shaded blocks. The left block is two thirds of the DVD area, which is part (c), and the right block's height is part (e). The band steps from one half down to one sixth: borrowing a book and borrowing a DVD are dependent.

Check the conditional answers against the plain DVD probability

Why: the overall DVD rate must be a weighted average of the two conditionals, weighted by the book and no-book probabilities.

\[ (0.40)(0.5) + (0.60)(0.1667) = 0.20 + 0.10 = 0.30 \]

OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability §3.3, pp. 182-183 — Try Its 3.17 and 3.19

57. At least one six in two rolls

Estimation

Roll a fair die twice. The chance of a six on each roll is one sixth, so it is tempting to say the chance of at least one six is one sixth plus one sixth, one third. Estimate before you compute.

Predict first

The true probability of at least one six in two rolls is

  • exactly 1/3
  • a little less than 1/3
  • a little more than 1/3
  • 1/36

Correct: a little less than 1/3

Why: Adding one sixth and one sixth counts the double six twice. The rolls are independent, so the double six has probability one sixth times one sixth, one thirty-sixth, and the addition rule subtracts it once: eleven thirty-sixths, about 0.306, a little under a third.

\[ P(\text{at least one six}) = \frac16 + \frac16 - \frac16 \cdot \frac16 = \frac{11}{36} \approx 0.306 \]

This one problem uses both rules and both special cases in the right places. The rolls are independent, so the AND is a plain product. The events are not exclusive, since both rolls can be sixes, so the OR keeps its subtraction.

The error in one third is small for two rolls, but watch it grow. For six rolls, adding six sixths would give a probability of 1, a certainty of rolling a six, which is plainly false. The more events you add, the more overlaps the plain sum double counts.

The complement gives the same answer another way: no six on either roll is five sixths times five sixths, twenty-five thirty-sixths, and one minus that is eleven thirty-sixths. When two methods agree, you can be confident in both.

58. Put the solving steps in order

Ranking

Here is the routine that the worked examples have followed. Put its steps in the order you would carry them out on a new problem.

Put in order

Order the steps for solving a worded two-event probability problem.

  1. Name the events and write every given number in probability notation, marking conditionals
  2. Decide from the wording whether the question asks for AND, OR, a conditional or a complement
  3. Write the general rule for that question
  4. Simplify only if independence or exclusivity is stated or checked
  5. Substitute and compute, one move per line
  6. Check the answer against its bounds and state it in the words of the problem

Why: Writing the givens in notation first is what exposes the conditionals, and the conditionals decide which form of each rule you can use. The general rule goes down before any simplification, so a shortcut is only ever a deliberate deletion. The final check catches impossible answers, and the sentence in context is what the question actually asked for.

The step people skip is the first one. Writing 75 percent of those who support as the probability of B given C, rather than as a plain 0.75, is what stops you multiplying by the wrong factor later.

The last step is the one that turns a number into an answer. A probability of 0.715 means nothing to Carlos's coach; the sentence Carlos makes at least one of the two goals about 72 percent of the time does.

The step people rush is the fourth. It is always safe to leave a problem in its general form; it is never safe to simplify without a reason.

59. Putting it together

Section

Part 6

60. Pattern: two rules, two permissions

Pattern

Everything in this section fits in one small table. Two questions, two general rules that are always true, and two conditions that each let you delete one term.

questiongeneral rule (always true)simplifies whento
both: A AND BP(B) P(A given B)independent: P(A given B) = P(A)P(A) P(B)
at least one: A OR BP(A) + P(B) − P(A AND B)mutually exclusive: P(A AND B) = 0P(A) + P(B)
neither1 − P(A OR B)(always)
  1. Hear the question in the wording: both, either, neither, given.
  2. Write the givens in notation, marking which numbers are conditionals and on what.
  3. Write the general rule. It is never wrong.
  4. Delete a term only with a reason: independence stated or tested, exclusivity forced by the situation.
  5. Check the bounds: an AND is at most each event; an OR is at least each event and at most 1.

The checks in the last line are free and catch most errors. An AND bigger than one of its events, or an OR smaller than one of its events, means a rule was used the wrong way round. The pizza exercise showed they can even catch bad data.

61. Check: the multiplication rule

Check

One question, one rule. Decide which conditional you are given before you multiply.

Check your understanding

P(A) = 0.40, P(B) = 0.25 and P(B given A) = 0.30. What is P(A AND B)?

  • A. 0.12 (correct)
  • B. 0.10
  • C. 0.075
  • D. 0.65

Answer: A

Why: The given conditional is B given A, so A is the plain factor: 0.40 times 0.30 is 0.12. The probability of B is not needed for this AND at all.

Why B tempts people
That is 0.40 times 0.25, the plain product, which is only valid if the events are independent; nothing says they are, and in fact 0.30 is not 0.25.
Why C tempts people
That multiplies the conditional by the probability of B, but the conditional has A to the right of its bar, so A is the factor that goes with it.
Why D tempts people
That adds the two plain probabilities, which is the start of an OR calculation, not an AND.

The first thing to find is the bar. The conditional here is B given A, so the event to the right of the bar is A, and the plain probability of A is the factor that goes with it.

If you chose the plain product, go back to Carlos. The plain product is the general rule with the conditional replaced, and that replacement needs independence. Here the conditional, 0.30, differs from the plain probability of B, 0.25, so the events are dependent and the product undercounts.

62. Check: the addition rule

Check

Now an OR. Ask whether the events can happen together before deciding what to subtract.

Check your understanding

P(A) = 0.45, P(B) = 0.35 and P(A AND B) = 0.15. What is P(A OR B)?

  • A. 0.80
  • B. 0.65 (correct)
  • C. 0.50
  • D. 0.1575

Answer: B

Why: The events overlap, so use the general addition rule: 0.45 plus 0.35 is 0.80, and taking the overlap away once gives 0.65.

Why A tempts people
That is the plain sum, which counts the overlap of 0.15 twice. It is only correct for mutually exclusive events, and these have a nonzero intersection.
Why C tempts people
That subtracts the overlap twice. It was counted twice in the sum, so one subtraction leaves it counted exactly once, which is what you want.
Why D tempts people
That is 0.45 times 0.35, a product, which would be an AND probability under independence, not an OR probability.

The overlap is given as 0.15, not zero, so the events can happen together and the subtraction must stay. Everything after that is one addition and one subtraction.

A quick bounds check would have ruled two of these out without any arithmetic. An OR must be at least as big as each event, so anything below 0.45 is impossible, and that eliminates the product answer at once.

63. Check: independent or exclusive?

Check

Two separate tests, two separate verdicts. Run each one on its own.

Check your understanding

P(A) = 0.3, P(B) = 0.5 and P(A AND B) = 0.15. Which statement is true?

  • A. A and B are independent and not mutually exclusive (correct)
  • B. A and B are mutually exclusive and independent
  • C. A and B are mutually exclusive but not independent
  • D. A and B are neither independent nor mutually exclusive

Answer: A

Why: The product 0.3 times 0.5 is 0.15, exactly the joint probability, so the events are independent. The joint probability is 0.15, not zero, so they are not mutually exclusive.

Why B tempts people
Mutually exclusive would need the joint to be zero, and it is 0.15. Besides, events with positive probabilities can never be both exclusive and independent.
Why C tempts people
The joint 0.15 is not zero, so they are not mutually exclusive, and the product test shows they are independent.
Why D tempts people
The product test passes exactly: 0.3 times 0.5 equals the given 0.15, so the events are independent.

Run the tests one at a time and write each verdict down before looking at the options. Product test: multiply 0.3 by 0.5 and compare with the joint. Exclusivity test: is the joint zero? Two lines, two verdicts, and only then match them to a statement.

Notice that the two properties could not both hold here, whatever the numbers. Exclusivity makes the joint zero; independence makes it the product of two positive numbers. That is the intuition slide from Part 4 put to work.

64. Explain the two shortcuts

Explain it to yourself

The product rule and the plain sum are the two formulas students remember from this section. Explain where each comes from.

Discussion prompt

In two or three sentences, explain why the plain product and the plain sum are special cases of the general rules. Name the term each one deletes and the condition that allows the deletion.

A model answer uses the words replace and delete. Independence replaces a conditional with a plain probability; exclusivity deletes an intersection. Both are single substitutions into a rule that was already true.

If your explanation mentioned only the formulas, add the reason. The point of the section is that you never need to memorise four rules: two general ones, and two deletions you are sometimes allowed to make, cover everything.

A good test of your explanation: could someone use it to decide, for a new problem, whether the shortcut is allowed? If it names the condition to check, it passes.

65. Exit ticket

Exit ticket

On a given morning, the chance of rain is 0.3 and the chance your bus is late is 0.2. On rainy mornings the bus is late 40 percent of the time.

Discussion prompt

Find the probability of rain AND a late bus, and of rain OR a late bus. Then say whether the events are independent and whether they are mutually exclusive, with the number that decides each.

Write the three givens in notation first. The 40 percent is measured only on rainy mornings, so it is the probability of a late bus given rain, and rain is the plain factor that multiplies it.

Every part of the lesson is in this one problem: the conditional to the right of the bar picks the factor, the OR needs the AND, and the two properties are tested separately. If any part felt uncertain, the matching Part is the one to revisit.

66. Recap

Recap

Two rules, always true, and two conditions that shorten them. The general forms are the ones to write first on every problem; the shortcuts are what is left after you delete a term you have earned the right to delete.

ruleformulashortcut and its condition
multiplication (AND)P(A AND B) = P(B) P(A given B)P(A) P(B), only if independent
addition (OR)P(A OR B) = P(A) + P(B) − P(A AND B)P(A) + P(B), only if mutually exclusive
complement (not)P(A′) = 1 − P(A)the addition rule for A and A′

\[ P(A\text{ AND }B) \le P(A), \qquad P(A) \le P(A\text{ OR }B) \le 1 \]

Independence is about a conditional staying the same; exclusivity is about an intersection being empty. They are different questions, and exclusive events with positive probabilities are always dependent. Next, Section 3.4 puts counts into contingency tables, where every probability in this lesson can be read off a single grid.

OpenStax Introductory Business Statistics 2e, §3.3 Two Basic Rules of Probability §3.3 — the same two rules with business examples

Sources

  1. OpenStax Introductory Statistics 2e, §3.3 Two Basic Rules of Probability — Illowsky & Dean, OpenStax / Rice University, CC BY 4.0, pp. 178-183
  2. OpenStax Introductory Business Statistics 2e, §3.3 Two Basic Rules of Probability — Illowsky & Dean, OpenStax / Rice University, CC BY 4.0

Want this taught 1-on-1? Alexander tutors Statistics — $55/session, free consultation.

Book on Wyzant · Text (657) 465-8108