4.11 · SLPaper 1Paper 2

Conditional probability

In a minute

A P(A∣B)\mathrm{P}(A \mid B) is the chance of AA once you know BB has happened: divide the overlap by P(B)\mathrm{P}(B). It answers every "given that" question, from tree diagrams to binomial and normal tails. It also gives the clearest test for .

You'll be able to

  • Use the definition P(A∣B)=P(A∩B)P(B)\mathrm{P}(A \mid B) = \frac{\mathrm{P}(A \cap B)}{\mathrm{P}(B)} with probabilities, tables and Venn diagrams.
  • Use P(A∩B)=P(B) P(A∣B)\mathrm{P}(A \cap B) = \mathrm{P}(B)\,\mathrm{P}(A \mid B), and turn a round to find a first-stage probability.
  • Answer "given that" questions about binomial and normal random variables on Paper 2.
  • Use P(A∣B)=P(A)=P(A∣B′)\mathrm{P}(A \mid B) = \mathrm{P}(A) = \mathrm{P}(A \mid B') for , and test whether two events are independent.
Before you start3 quick questions on what this builds on
Check yourself
Events AA and BB have P(A)=0.5\mathrm{P}(A) = 0.5, P(B)=0.35\mathrm{P}(B) = 0.35 and P(A∪B)=0.7\mathrm{P}(A \cup B) = 0.7. Find P(A∩B)\mathrm{P}(A \cap B).
Check yourself
X∼B(5,0.2)X \sim \mathrm{B}(5, 0.2). Find P(X=0)\mathrm{P}(X = 0). Give your answer to 3 s.f.
Check yourself
Y∼N(50,42)Y \sim \mathrm{N}(50, 4^{2}). Use your calculator to find P(Y>54)\mathrm{P}(Y > 54). Give your answer to 3 s.f.

Revise first: 4.6 Probability diagrams, rules and independence, 4.8 Binomial distribution and 4.9 Normal distribution

The formal definition

In 4.6 you met P(A∣B)\mathrm{P}(A \mid B) through tables. Here is why the formula works for any two events.

Once you know BB has happened, every outcome outside BB is ruled out. So BB becomes the whole . The only part of AA still possible is the overlap, A∩BA \cap B.

A rectangle labelled U holds two overlapping circles, A on the left and B on the right. Circle B is shaded and outlined boldly as the new sample space; everything outside B is faded and crossed out as ruled out. The overlap of A and B is labelled as the only part of A still possible. U A B ruled out A ∩ B still possible A′ ∩ B ruled out
Given BB, only circle BB is left. The chance of AA is the overlap's share of it.

Dividing by P(B)\mathrm{P}(B) rescales the new sample space so its total probability is 11. The new chance of AA is the overlap's share of BB.

In the formula booklet
P(A∣B)=P(A∩B)P(B)\mathrm{P}(A \mid B) = \dfrac{\mathrm{P}(A \cap B)}{\mathrm{P}(B)}

You don't need to memorise this; it's in the formula booklet. You do need to spot which event is given. If P(B)=0\mathrm{P}(B) = 0, the formula has no meaning: you can't be told that something impossible happened.

Multiply both sides by P(B)\mathrm{P}(B) for the second form: the chance of BB, times the chance of AA once BB has happened. Learn this one.

Learn this
P(A∩B)=P(B) P(A∣B)\mathrm{P}(A \cap B) = \mathrm{P}(B)\,\mathrm{P}(A \mid B)

The bar is not symmetric. P(B∣A)\mathrm{P}(B \mid A) divides by P(A)\mathrm{P}(A) instead, so it is usually different from P(A∣B)\mathrm{P}(A \mid B).

Worked exampleEvents AA and BB have P(A)=0.45\mathrm{P}(A) = 0.45, P(B)=0.4\mathrm{P}(B) = 0.4 and P(A∩B′)=0.21\mathrm{P}(A \cap B') = 0.21. Find (a) P(A∣B)\mathrm{P}(A \mid B); (b) P(B∣A)\mathrm{P}(B \mid A).

Try it first, then reveal one step at a time.

  1. Both parts need the overlap. AA is split into A∩B′A \cap B' and A∩BA \cap B, so P(A∩B)=0.45−0.21=0.24\mathrm{P}(A \cap B) = 0.45 - 0.21 = 0.24.
Check yourself
Priya's train is cancelled on 8%8\% of days. On a day it is cancelled, the probability that she is late for work is 0.650.65. Find the probability that, on a random day, her train is cancelled and she is late.

On a , every second-stage branch is a conditional probability. So the definition lets you turn a tree round: given how the second stage ended, find the chance of a first-stage branch.

Tree diagram. First stage: easy question E with probability 3/4, hard question E prime with probability 1/4. After E: correct C with 4/5, wrong C prime with 1/5. After E prime: correct C with 1/3, wrong C prime with 2/3. Path products: 3/5, 3/20, 1/12 and 1/6. The two paths ending in C prime are highlighted. E E′ C C′ C C′ 3/4 1/4 4/5 1/5 1/3 2/3 3/4 × 4/5 = 3/5 3/4 × 1/5 = 3/20 1/4 × 1/3 = 1/12 1/4 × 2/3 = 1/6
The highlighted paths both end in C′C'. Together they make P(C′)\mathrm{P}(C').
Worked exampleIn a quiz app, each question is easy with probability 34\frac{3}{4} and hard otherwise. Leo answers an easy question correctly with probability 45\frac{4}{5} and a hard one with probability 13\frac{1}{3}. Let EE be the event "the question is easy" and CC the event "Leo answers correctly". (a) Show that the probability that Leo answers a question wrongly is 1960\frac{19}{60}. (b) Given that Leo answers a question wrongly, find the probability that it was hard.

Try it first, then reveal one step at a time.

  1. Draw the tree. First stage: EE with 34\frac{3}{4} and E′E' with 14\frac{1}{4}. Wrong answers: P(C′∣E)=1−45=15\mathrm{P}(C' \mid E) = 1 - \frac{4}{5} = \frac{1}{5} and P(C′∣E′)=1−13=23\mathrm{P}(C' \mid E') = 1 - \frac{1}{3} = \frac{2}{3}.
Check yourself
Events AA and BB have P(B)=0.25\mathrm{P}(B) = 0.25, P(A∣B)=0.6\mathrm{P}(A \mid B) = 0.6 and P(A∪B)=0.6\mathrm{P}(A \cup B) = 0.6. Find P(A∣B′)\mathrm{P}(A \mid B').

Given that, with a random variable

On Paper 2 the events are often inequalities about a , such as X>7X > 7 or X≤9X \le 9. The definition is the same. The new skill is finding the overlap of two inequalities.

1. Write the definition with the given inequality on the bottom.

2. Find the overlap: the values that satisfy both inequalities. Often one event sits inside the other, so the overlap is just the smaller event.

3. Find each probability with your calculator. Divide the unrounded values, and round only the final answer.

Worked exampleA basketball player scores each free throw with probability 0.70.7, independently of the others. She takes 1010 free throws. Given that she misses at least one throw, find the probability that she scores more than 77. Give your answer to 3 s.f.

Try it first, then reveal one step at a time.

  1. Let XX be the number she scores, so X∼B(10,0.7)X \sim \mathrm{B}(10, 0.7). Missing at least one means X≤9X \le 9.
Worked exampleThe time, TT minutes, that a bus takes for its route is modelled by T∼N(40,62)T \sim \mathrm{N}(40, 6^{2}). A journey is on schedule if it takes between 3232 and 4848 minutes. Given that a journey is on schedule, use your calculator to find the probability that it takes more than 4444 minutes. Give your answer to 3 s.f.

Try it first, then reveal one step at a time.

  1. Write the definition with the given event on the bottom: P(T>44∣32<T<48)=P(44<T<48)P(32<T<48)\mathrm{P}(T > 44 \mid 32 < T < 48) = \frac{\mathrm{P}(44 < T < 48)}{\mathrm{P}(32 < T < 48)}.
Check yourself
A discrete random variable XX has P(X=1)=0.1\mathrm{P}(X = 1) = 0.1, P(X=2)=0.2\mathrm{P}(X = 2) = 0.2, P(X=3)=0.3\mathrm{P}(X = 3) = 0.3 and P(X=4)=0.4\mathrm{P}(X = 4) = 0.4. Find P(X=4∣X>2)\mathrm{P}(X = 4 \mid X > 2).
Check yourself
The heights, HH cm, of sunflowers in a field are modelled by H∼N(165,82)H \sim \mathrm{N}(165, 8^{2}). Given that a sunflower is taller than 170170 cm, find the probability that it is taller than 180180 cm. Give your answer to 3 s.f.

Independence and how to test it

Two events are when knowing about one tells you nothing about the other. So the chance of AA is the same whether BB happened or not.

Learn this
P(A∣B)=P(A)=P(A∣B′)\mathrm{P}(A \mid B) = \mathrm{P}(A) = \mathrm{P}(A \mid B')

The booklet gives the same idea as a product, as in 4.6.

In the formula booklet
P(A∩B)=P(A) P(B)\mathrm{P}(A \cap B) = \mathrm{P}(A)\,\mathrm{P}(B)

So there are three tests. They always agree, so pick the one that fits your numbers:

1. Is P(A∣B)=P(A)\mathrm{P}(A \mid B) = \mathrm{P}(A)?

2. Is P(A∣B)=P(A∣B′)\mathrm{P}(A \mid B) = \mathrm{P}(A \mid B')?

3. Is P(A∩B)=P(A) P(B)\mathrm{P}(A \cap B) = \mathrm{P}(A)\,\mathrm{P}(B)? If the two sides are equal, the events are independent. If not, they are not.

The proof also shows that independence passes to complements. If AA and BB are independent, so are AA and B′B', and so are A′A' and B′B'. For example, P(A′∣B′)=1−P(A∣B′)=1−P(A)=P(A′)\mathrm{P}(A' \mid B') = 1 - \mathrm{P}(A \mid B') = 1 - \mathrm{P}(A) = \mathrm{P}(A').

A survey of 200200 adults records whether each owns a smartwatch (WW) and whether they run regularly (RR).

RR′TotalW362460W′8456140Total12080200\begin{array}{l|cc|c} & R & R' & \text{Total} \\ \hline W & 36 & 24 & 60 \\ W' & 84 & 56 & 140 \\ \hline \text{Total} & 120 & 80 & 200 \end{array}

Worked exampleUse the smartwatch table. One adult is chosen at random. Determine whether RR and WW are independent.

Try it first, then reveal one step at a time.

  1. Use test 2: compare the chance of running in the two rows.
Worked exampleEvents AA and BB are independent, with P(A∣B′)=0.4\mathrm{P}(A \mid B') = 0.4 and P(B)=0.35\mathrm{P}(B) = 0.35. Find P(A∪B)\mathrm{P}(A \cup B).

Try it first, then reveal one step at a time.

  1. AA and BB are independent, so P(A)=P(A∣B′)=0.4\mathrm{P}(A) = \mathrm{P}(A \mid B') = 0.4.
Check yourself
Events AA and BB have P(A)=0.6\mathrm{P}(A) = 0.6, P(B)=0.25\mathrm{P}(B) = 0.25 and P(A∩B)=0.12\mathrm{P}(A \cap B) = 0.12. Are AA and BB independent?
Check yourself
Events AA and BB are independent, with P(A)=0.45\mathrm{P}(A) = 0.45 and P(B)=0.2\mathrm{P}(B) = 0.2. Find P(A′∣B′)\mathrm{P}(A' \mid B').

Common mistakes

  • Dividing by the wrong probability. For "hard, given wrong", divide by P(wrong)\mathrm{P}(\text{wrong}), not P(hard)\mathrm{P}(\text{hard}).
  • Stopping at the overlap, such as P(8≤X≤9)\mathrm{P}(8 \le X \le 9). A "given that" answer always ends with a division by the probability of the condition.
  • Reading a rate within a group as overall. "30%30\% of BB is in AA" means P(A∣B)=0.3\mathrm{P}(A \mid B) = 0.3, so P(A∩B)=0.3 P(B)\mathrm{P}(A \cap B) = 0.3\,\mathrm{P}(B).
  • Treating "more than 77" as X≥7X \ge 7 for a whole-number variable. X>7X > 7 means X≥8X \ge 8.
  • Writing P(A∣B′)=1−P(A∣B)\mathrm{P}(A \mid B') = 1 - \mathrm{P}(A \mid B). Complements change the event before the bar, never the condition.
  • Saying "independent" with no comparison, or multiplying P(A) P(B)\mathrm{P}(A)\,\mathrm{P}(B) when independence was never given.

Exam tips

  • "Find" a conditional probability: write the definition with your own events in it before any numbers. That line is where the method mark is.
  • On Paper 2, carry the full calculator values for the top and the bottom, and round only the final answer to 3 s.f.
  • When a later part uses an earlier probability, such as P(C′)\mathrm{P}(C') on the bottom, a wrong earlier value used correctly can still earn follow-through marks.

Key takeaways

In bookletP(A∣B)=P(A∩B)P(B)\mathrm{P}(A \mid B) = \dfrac{\mathrm{P}(A \cap B)}{\mathrm{P}(B)}
Learn thisP(A∩B)=P(B) P(A∣B)\mathrm{P}(A \cap B) = \mathrm{P}(B)\,\mathrm{P}(A \mid B)
Learn thisThe given event, after the bar, goes on the bottom. P(A∣B)\mathrm{P}(A \mid B) and P(B∣A)\mathrm{P}(B \mid A) are usually different.
Learn thisP(A′∣B)=1−P(A∣B)\mathrm{P}(A' \mid B) = 1 - \mathrm{P}(A \mid B), but P(A∣B′)\mathrm{P}(A \mid B') needs the definition again.
Learn thisWith a random variable, the top is the overlap of the two inequalities: P(X>7∣X≤9)=P(8≤X≤9)P(X≤9)\mathrm{P}(X > 7 \mid X \le 9) = \frac{\mathrm{P}(8 \le X \le 9)}{\mathrm{P}(X \le 9)}.
Learn thisP(A∣B)=P(A)=P(A∣B′)\mathrm{P}(A \mid B) = \mathrm{P}(A) = \mathrm{P}(A \mid B')
In bookletP(A∩B)=P(A) P(B)\mathrm{P}(A \cap B) = \mathrm{P}(A)\,\mathrm{P}(B)
Learn thisTo test independence, compare two values that would be equal, write == or ≠\ne, then conclude.

Practise this

20 past-paper questions on 4.11 in the AA HL bank, each with a mark scheme and a worked explanation.

Practise 4.11 →