Teacher notes

5 Quantifying uncertainty: probability

Essential understandings
Probability enables us to quantify the likelihood of events occurring and so evaluate risk. Both
statistics and probability provide important representations which enable us to make predictions,
valid comparisons and informed decisions. These fields have power and limitations and should be
applied with care and critically questioned to differentiate between the theoretical and the
empirical/observed. Probability theory allows us to make informed choices, to evaluate risk, and to
make predictions about seemingly random events.

Content-specific conceptual understandings

This chapter leads to the following content-specific conceptual understandings listed in the subject
guide:

 Approximation in data can approach the truth but may not always achieve it.
 Correlation and regression are powerful tools for identifying patterns and equivalence of
systems.
 Modelling and finding structure in seemingly random events facilitates prediction.
 Different probability distributions provide a representation of the relationship between the
theory and reality, allowing us to make predictions about what might happen.

We have taken these suggested content-specific conceptual understanding statements and, as

recommended in the Teacher Support Manual, developed our own conceptual understandings.
Students are led step-by-step through the investigations to arrive at one or more of these
conceptual understandings:

Conceptual understandings Investigation

Standard Level

Random behavior by nature involves unpredictability in the short Investigation 1

term.

Random behavior by nature involves unpredictability in the short Investigation 2

term.

There is a long-term trend towards the theoretical probability, but

the data from an experiment can diverge from the theoretical at any
point.

The number given by the formula for theoretical probability serves

as a good model for quantifying the likelihood of the event and
allows us to make predictions.

Representing combined probabilities using diagrams can show Investigation 3

structure and pattern that a formula may not.

Mutually exclusive events cannot occur at the same time, which Investigation 4
results in no intersection. Independent events represent unrelated
separate events in which the probability of one event does not affect
the outcome of the other. However, mutually exclusive events
cannot be independent.

When calculating the probability of an event, it may be more Investigation 5

efficient to look at the probability of the complementary event.

© Oxford University Press 2019 Teacher notes 1

 Teacher notes

Syllabus sections covered in this chapter:

 SL4.5*
 SL4.6*
 AHL4.19

Cognitive academic language proficiency

The academic language used in this chapter is listed as "microconcepts" at the start of the chapter.
Moreover, when specific terminology is introduced it is defined clearly and then used in context to
deepen students’ understanding.

Cognitive activators
The stimulus questions, developing inquiry skills, before you start activities, investigations and
modelling activities are cognitive activators. They get students ready to learn and engage with the
subject, providing opportunities for collaborative and peer-to-peer learning, and to develop of
inquiry, investigative and modelling skills.

Digital resources

Prior learning Animated worked GDC skills and Additional

support example support exercises

Page 204: Quantifying Page 219: Example 7 Pages 210, 215,

uncertainty:
Page 219: Example 8 220, 223
probability

Assessment opportunities

End of chapter summary Chapter review Exam-style questions

Page 224 Page 226 Page 227

Opening problem
Two taxi companies operate in Mathcity: Blackcabs and Yellowrides. 85% of the cabs in the city
work for Blackcabs and are coloured black. The rest of the cabs in the city work for Yellowrides and
are coloured yellow. A taxi was involved in a hit and run accident at night. A witness made a
statement to the police that the taxi involved in the accident was yellow. The court carried out a
series of tests of the reliability of the witness, asking her to identify the colour of a random
sequence of taxis. The witness correctly identified each one of the two colours 80% of the time
and failed 20% of the time. What is the probability that the taxi involved in the accident was
yellow?
12
Answer:  0.414
29

© Oxford University Press 2019 2

 Teacher notes

Developing inquiry skills

Write down any similar inquiry questions you might ask if you were asked to predict the reliability
of the witness if 50% of the cars in the city were Yellowrides or if another taxi company Blue Taxis
also operated in Mathcity. What questions might you need to ask in these scenarios?
Answers: How could the diagram for the original problem be adapted? Would the witness
pay closer attention to Yellow cabs when they were less likely (than 50%) to appear?
Would the witness find black and blue harder to distinguish than black and yellow?
The opening problem is frequently answered incorrectly and is an example of the base
rate fallacy which you may like to explore.
This problem has a different context but the same structure:
Designers of medical tests have to quantify “false positives”, in which a person tests
positive for a disease but in fact does not have the disease, and “false negatives”, in which
a person tests negative for a disease but in fact does have the disease.
Clinical trials reveal that in a large population, a disease affects 0.4% of the population. A
medical test has been found to be accurate in that the frequencies of false positives and
false negatives are small. The probability that the test says you have the disease when
you do not in fact have the disease is 0.03. The probability that the test says you have not
got the disease but in fact you do is 0.02.
You take the test and are told that you have the disease. What is the probability that you
actually do have the disease?

5.1 Reflecting on experiences in the world of chance.

First steps in the quantification of probabilities
TOK

Intuition is often not a good way of knowing in Probability. There are lots of other examples in this
field where intuition potentially lets you down. What about in other areas of knowledge? Is it more
or less reliable there?

Investigation 1
Conceptual understanding:
Random behavior by nature involves unpredictability in the short term.
Event D is deliberately set up to be easy to agree on. However, others are far more challenging
to agree on. Students should conclude that assigning probability values derived from personal
knowledge, interests and beliefs is subjective. Students may reflect on Plato’s definition of
knowledge as “Justified true belief”. A particular probability may be justified in part by one’s
feelings, intuitions and experiences.

Investigation 2
Conceptual understandings:
Random behaviour by nature involves unpredictability in the short term.
There is a long-term trend towards the theoretical probability, but the data from an experiment
can diverge from the theoretical at any point.
The number given by the formula for theoretical probability serves as a good model for quantifying
the likelihood of the event and allows us to make predictions.
2 Instructions for the spreadsheet:

© Oxford University Press 2019 3

 Teacher notes

EXPERIMENT: throwing a 12-sided die and recording if the uppermost number is prime (EXCEL)
Type 1 into cell A1, 2 into cell A2, 3 into cell A3. Select cells A1, A2 and A3 and drag down to
A100 to get all the numbers 1..100 in column A. These numbers are the number of trails.
Type “=RANDBETWEEN(1,12)” in cell B1 and drag down to cell b100. These generate random
numbers between 1 and 12, and so they simulate throwing a 12-sided die. Press F9 to throw
the 100 dice again.
Type “=OR(B1=2,B1=3,B1=5,B1=7,B1=11) into cell C1. This gives “TRUE” if the number B1 is
prime, FALSE otherwise. Drag down to cell C100.
Type “=IF(C1,1,0)” in cell D1. This just changes “TRUE” to 1 and “FALSE” to zero. Drag down
to cell D100.
Type “=D1” into cell E1
Type “=E1+D2” into cell E2 and drag down to cell E100. This gives the cumulative number of
times that a prime number was thrown.
Type “=E1/A1” into cell F1 and drag down to cell F100. This gives the experimental probability
of throwing a prime number when the experiment is repeated in 100 trials
Add a graph showing A1..A100 on the x-axis and F1..F100 on the y-axis.
3 and 4Answers will vary.
5 Factual: What is the set of all possible values of theoretical probabilities?
Answer: Theoretical probabilities always lie in the domain [0,1].
6 The graph should fluctuate above and below the line randomly.
7 Factual: What is the relationship between relative frequency and theoretical probability in the
short term?
Answer: In the short term, we can’t know when exactly relative frequency (experimental
probability) may equal the theoretical or for how long it is more or less than the theoretical.
8 Factual: What is the relationship between relative frequency and theoretical probability in the
long term?
Answer: In the long term, relative frequency (experimental probability) approaches the
theoretical. However, it can diverge from the theoretical probability at any time.
9 Conceptual: Does random behaviour involve predictability in the short term or
unpredictability?
Answer: Random behaviour by nature involves unpredictability in the short term.
10 Conceptual: Does random behaviour involve predictability in the long term or unpredictability?
Answer: Both. There is a long-term trend towards the theoretical probability, but the data
from an experiment can diverge from the theoretical at any point.
11 Conceptual: How may we interpret and apply the number quantified by the formula for the
theoretical probability of an event?
Answer: The number given by the formula for theoretical probability serves as a good
model for quantifying the likelihood of the event and allows us to make predictions.

TOK

The game in the St Petersburg paradox follows these rules. You flip a coin; if it lands tails up then
you lose and the game is over. If the coin lands heads up then you win one rouble and the game
continues. The coin is tossed again. If it is tails, then the game ends and you keep the money you
have won. If it is heads, then you win an additional two roubles.

© Oxford University Press 2019 4

 Teacher notes

For each successive head you double your winnings from the previous round, but, at the first tail,
the game is over.
How much would you pay to play this game?

Developing inquiry skills

There are four outcomes in the first opening scenario:
• a taxi is yellow and is identified as yellow

• a taxi is yellow and is identified as black

• a taxi is black and is identified as yellow

• a taxi is black and is identified as black.

Are these equally likely outcomes?

In 1000 trials, how many occurrences of each outcome would you expect?
Answer: No they are not equally likely outcomes. 170, 680, 120 and 30 respectively.

5.2 Representing combined probabilities with diagrams

Investigation 3
Conceptual understanding:
Representing combined probabilities using diagrams can show structure and pattern that a formula
may not.

3
2 P(Bio and Chem)=
8
3 Answers will vary.

3
5 P(total is prime)=
8
6

1 2 3 4 5 6

1 2 3 4 5 6 7

3 4 5 6 7 8 9

4 5 6 7 8 9 10

5 6 7 8 9 10 11

6 7 8 9 10 11 12

8 9 10 11 12 13 14

Students can see from the pattern that for example the outcomes 11 and 5 are equiprobable.
7 Conceptual: What advantages are there in using a diagram in problem solving with combined
probabilities?
Answer: Representing combined probabilities using diagrams can show structure and
pattern that a formula may not.

© Oxford University Press 2019 5

 Teacher notes

TOK

A good question for a debate or blog post would be: “Ethics is an area of knowledge in its own
right. Where do you see an intersection of the areas of knowledge of mathematics and ethics?”
When taking a chance decision in your life, on which skills do you rely and in which order?
Intuition (your gut feeling)? Reason? Emotion? Memory? Faith? Imagination?

Developing inquiry skills

In the first opening scenario, imagine 100 trials. How many outcomes would you expect in each
area shown on this diagram?
Answer:
Cab yellow?
Yes No
Yes 12 68
correct?
Witness

No 3 17

5.3 Representing combined probabilities with diagrams

and formulae
Investigation 4
Conceptual understanding:
Mutually exclusive events cannot occur at the same time, which results in no intersection.
Independent events represent unrelated, separate events in which the probability of one event
does not affect the outcome of the other. However, mutually exclusive events cannot be
independent.

1 2 3 4 5 6 7 8 9 10

Venn diagram P(A) P(A’) P(B) P( A B) P( A B) P( A | B) P(A) P(A) P(A) P( A B)

+ x +P(B) P(B)
P(B) P(B) −
P( A B)

Class of 2019 7 11 1 2 2 4 8 7 2 4
18 18 2 9 3 9 9 36 3 9

Class of 2018 3 16 10 0 13 0 13 30 13 0
19 19 19 19 19 361 19

© Oxford University Press 2019 6

 Teacher notes

Class of 2017 13 6 10 10 13 1 23 130 13 1

19 19 19 19 19 19 361 19

Class of 2016 1 2 5 5 3 1 23 5 3 1
3 3 8 24 4 3 24 24 4 3

1 Each pair of probabilities adds to 1.

2 Because the union of two complementary events is the whole sample space.
3 The probabilities in each pair are equal.
4 The intersection must be subtracted or else it will be counted twice.
5 The probabilities in each pair are equal.
6 Factual: Which Venn diagram shows “mutually exclusive events”?
Answer: The class of 2018
7 Factual: Which Venn diagram shows “independent events”?
Answer: The class of 2016

8 Conceptual: What is the difference between mutually exclusive and independent events? Can
mutually exclusive events be independent? Why or why not?
Answer:
Mutually exclusive events cannot occur at the same time, which results in no intersection.
Independent events represent unrelated, separate events in which the probability of one
event does not affect the outcome of the other. However, mutually exclusive events
cannot be independent.
Reflect:

Are complementary events independent events? Answer: No

Are complementary events mutually exclusive? Answer: Yes

Can mutually exclusive events be independent? Answer: No

Can non-mutually exclusive events be independent? Answer: Yes

Developing inquiry skills

Which of the events in the first opening scenario are independent? Which are mutually exclusive?
Answer: The witness identifying the colour of the cab correctly is independent of the
colour of the cab. The cab is either black or yellow so these are mutually exclusive events,
just as the witness being right or wrong are mutually exclusive events.

© Oxford University Press 2019 7

 Teacher notes

5.4 Complete, concise and consistent representations

TOK

The difference between personal knowledge and shared knowledge offers you the chance to
consider the difference between “what I know” and “what we know”.
Shared knowledge may come from texts, teachers, media etc, personal knowledge is gained
through the experience of the individual, such as ice is cold or rabbits are fluffy.
Somebody who studies computing might view their laptop differently because of their academic
knowledge. Their personal knowledge had been affected by the shared knowledge they had gained
in class. This would be an intersection of the two types of knowledge.

Investigation 5
Conceptual understanding:
When calculating the probability of an event, it may be more efficient to look at the probability of
the complementary event.
1 6
2 3
3 Factual: Hence fill in the table:
Answer:

Number How many probabilities How many probabilities would be

of dice would be found and added found, added and their total
from the tree diagram subtracted from 1 using a
representation complementary events
representation.

2 2 2

3 6 3

4 12 4

5 20 5

k k2  k k

4 Factual: Can you generalise for n1 , n2 ….. nk dice with distinct colours c1 , c2 … ck ?

Answer: It’s more efficient to use the complementary events representation since that
would involve working out k probabilities instead of the k 2  k probabilities that would be
necessary for the tree diagram representation.
5 Conceptual: Why is it useful to calculate probabilities for the complementary event in some
situations?
Answer: When calculating the probability of an event, it may be more efficient to look at
the probability of the complementary event.

TOK

Which option do you think is more likely? Why? We will return to this question at the end of this
section.
Consider the ethics of gambling.

© Oxford University Press 2019 8

 Teacher notes

Developing inquiry skills

Apply what you have learned in this section to represent the first opening problem with a tree
diagram. Hence find the probability that a cab is identified as yellow. Apply the formula for
conditional probability to find the probability that the cab was yellow given that it was identified as
yellow. How does your answer compare to your original subjective judgement?
Answer: cB is the event that the cab was in fact black, cY is the event that the cab was in
fact yellow.
wB is the event that the cab was identified as black by the witness, wY is the event that
the cab was identified as yellow by the witness.

Hence P(wY)  0.85  0.2  0.15  0.8  0.29

P(cY wY) 0.8  0.15

Hence P(cY wY)    0.414
P(wY) 0.29

Random walking!
Approaches to Learning/learner profile: Critical thinking
Exploration Criteria: Mathematical communication (B), Personal engagement (C), Use of
mathematics (E)
IB Topic: Probability, Discrete Distributions
This problem is designed to encourage students to think of simulation as a reasonable and
acceptable approach to probability problems that may be too difficult to approach theoretically as
they develop. This problem has a clearly stated aim and is accessible at first using Mathematical
communication (Criterion B) and methods familiar to students from the chapter. The mathematics
required to prove the result is quite difficult for some, but this should not restrict students from
accessing it and using the tools available to them.
At the beginning of the problem students use basic coin tossing simulations and collect results as a
class to produce more results and a hopefully more accurate answer. At the end of the task
students are asked to consider using computer simulation. Coding for this is not only accessible for
a computer science student or experienced programmer but can be learnt with a little effort and

© Oxford University Press 2019 9

 Teacher notes

Personal engagement (Criterion C). Further personal engagement can be shown by extending the
problem once the code has been mastered.
The proof of the result requires some understanding of probability distributions so this task may be
better covered after that.
The problem

The problem is adapted from a famous problem in a branch of mathematical problems involving
‘random walks’. Study of this branch has contributed to many different areas in physics and
chemistry (Brownian motion and diffusion), biology (genetics, animal movements, population
dynamics), economics (modeling share prices) and computer science (social media suggestions),
amongst others.
Explore the problem

Since the man moves left or right with equal probability, a coin toss can be used to simulate this.
If appropriate, ask:
• Why is a coin toss a suitable simulation?

Students play the game 10 times and find the average number of steps taken. Discuss why this
may not be an accurate result.
Ask:
• What could you do to improve the accuracy of the average?

Discuss the improved result based on a larger sample size.

The average may be getting closer, but you do not know if this is the actual number.
You can only be certain by proving the result theoretically.
Ask:
• What have you noticed so far?

(For example, always an odd number of steps, theoretically could go on forever, etc.)
Calculate probabilities

Students may need help when constructing the tree diagram.

Ask:
• What are the limitations of using a tree diagram in this case?

The tree diagram is very large! It becomes impossible to draw after 6 or 7 tosses.
Remind students how to use a tree diagram to find probabilities.
If needed, to help students find the probability that the man falls into the ditch after a total of
exactly 5 steps, ask:
If the man moves left (L) then left again (L) and then right (R) and then left (L) and then left (L)
then he will be in the ditch. Where is this scenario on your tree diagram?
What is the probability that the man takes this particular sequence of steps? In other words what
is the probability of TTHTT?
Probability is (0.5)5 = 0.03125 = (1/32)
What other sequences of coin tosses will lead to the man falling into the ditch after exactly 5
steps?
TTHTT, THTTT, THHHH, HTTTT, HTHHH, HHTHH
What are the probabilities associated with each of these sequences?
They are all (0.5)5 = 0.03125 = 1/32

© Oxford University Press 2019 10

 Teacher notes

What is the probability that the man falls into the ditch after a total of exactly 5 steps?
6(0.5)5 = 6(0.03125) = 3/16
Minimum number of steps to fall into the ditch is 3.
Maximum is infinite.
Probability that the man falls into the ditch after a total of exactly 3 steps is

1
2(0.5)3 = 2(0.125) =
4
To explain why all the paths have an odd number of steps:
It is an odd number of steps because:
From the centre, after the first step, the man will always be one step away from the centre (two
steps away from the ditch on that side).
From here, after 2 steps, he will either be in the ditch on the same side, back to the same position,
or one step from the centre (two steps away from the ditch) on the other side. This will repeat.
This gives 1 + a multiple of 2 which is odd.
You could use a diagram to demonstrate this, with the centre shown by a black dot, one step from
the centre on either side shown by a red dot, two steps away from the centre on each side shown
by a black dot, and the ditches shown by blue dots.
The probabilities are:

x 1 2 3 4 5 6 7 8 9 10 11 12 ….

P(X = 0 0 1 0 3 0 9 0 27 0 81 0
x) 4 16 64 256 1024

The table alternates between 0 and a value.

3
The next value is calculated by multiplying by .
4
Ask:

3
• Can you explain where the value of comes from in this situation?
4
Again, you could use diagrams to demonstrate this.
Simulation

Students will study expectation in a chapter 13.

For now, if appropriate, you could ask:
• What is an expected value?

You could also give students the formula for calculating the expected number of steps that would
be required:

E( X )   xP(X
1
 x)

E(X) is the expected number of steps.

This will give you the exact theoretical answer to the problem posed.
Help them to understand and use this formula if needed.
© Oxford University Press 2019 11
 Teacher notes

The formula is:

Multiply each probability by its related value and sum the result.
Students may note here that this problem will be complicated because there are an infinite
number of values of x that will result in falling into the ditch.
It is possible to calculate the expected value, but it requires mathematics that will be beyond the
SL and HL syllabus.
As extension, students could perhaps try to do this when they have completed Chapter 13 and
fancy a challenge!
Here is the solution written out:


E( X )   x  P( X
x 1
 x)

1 3 9
 (1  0)  (2  0)  (3  )  (4  0)  (5  )  (6  0)  (7  )  ...
4 16 64
1 3 9 27
 (3  )  (5  )  (7  )  (9  )  ...
4 16 64 264
1 3 9
 (3  5   7   ...)
4 4 64
1 3 3 3
 (3  5   7  ( )2  9  ( )3  ...)
4 4 4 4
This is an infinite Arithmetico-Geometric series. Its sum can be found neatly as follows:

1 3 3 3
E( X )  [3  5  ( )  7  ( )2  9  ( )3  ...]
4 4 4 4
3 1 3 3 2 3 3
 E( X )   [3  ( )  5  ( )  7  ( )3  9  ( )4  ...]
4 4 4 4 4 4

Subtracting,

1 1 3 3 3
E( X )  [3  2  ( )  2  ( )2  2  ( )3  ...]
4 4 4 4 4
3 3 2 3 3
E( X )  3  2  ( )  2  ( )  2  ( )  ...
4 4 4
3 3 3 2 3
E( X )  3  2   [1  ( )  ( )  ( )3  ...]
4 4 4 4

3 1
 32 
4 3
1
4
3
 32 4
4
9

This shows that the theoretical result is 9. Compare this with the result from the earlier simulation
by the class.
At this stage, you could tell the class that the expected number of steps is 9.

© Oxford University Press 2019 12

 Teacher notes

Ask:
• Why do you think simulations are used?

Collecting and recording large numbers of results by hand is very time consuming and can be very
expensive.
Here are some examples of computer coding that could be copied or shown to the students. It
would be possible to replicate these results on most computer coding systems. (If there are
computer science students in the class then they could be encouraged to help the class, although
simple coding should be accessible to all!)
Note: The idea here is that simple computer coding is accessible to all students with a little work,
and then more complicated problems can be solved. This is considerably more efficient than more
manual methods. It is also worth pointing out that this is actually an exceptionally real process in
many, many fields of work such as meteorology, disaster management, economics and finance,
sports predictions, etc. This is one of the real avenues of work for the modern mathematician!
Extension

Suggestions of how students could vary the problem:

• Change the start position.
• Use a bias coin.
• Change the number of steps from the centre to the ditch.
• Change the problem to 2 dimensions.

Students may also be able to devise their own probability question which they could answer using
simulation.

© Oxford University Press 2019 13