2149080

Probability and Random Variables

Lecture 3 -
Conditional Probability
Slides by John Wiley & Sons, Inc. Applied Statistics and Probability for Engineers, by
Montgomery and Runger.

Dr. R. Melisa DENİZ

ruchan.deniz@medipol.edu.tr

Fall 2024
October, 11-18
Content

3 Conditional
Probability

v Conditional Probability
v Multiplication & Total Probability Rules
v Chain Rule
v Bayes’ Theorem
v Independence

1
Learning Objectives for Lecture 3
After careful study of this lecture, you should be
able to do the following:
1. Interpret and calculate conditional probabilities of
events.
2. Use Counting Techniques with Conditional
Probability.
3. Determine the independence of events and use
independence to calculate probabilities.
4. Use Bayes’ theorem to calculate conditional
probabilities.
5. Extend conditional probability case to multiple
events.
6. Realize the importance of computer simulations
for probability.
2
Random Experiment

vAn experiment is a procedure that is

• carried out under controlled
conditions, and
• executed to discover an unknown
result
vAn experiment that can result in
different outcomes, even though it is
repeated in the same manner every
time, is called a random experiment.
3
Sample Spaces

vTo model and analyze a random

experiment, we must understand the set of
possible outcomes from the experiment.
vThe set of all possible outcomes of a
random experiment is called the sample
space, 𝑆.
• A sample space is often defined based on the
objectives of the analysis.
• 𝑆 is discrete if it consists of a finite or countable
infinite set of outcomes.
• 𝑆 is continuous if it contains an interval
(either finite or infinite) of real numbers.
4
Example 2.1 | Camera Flash
Continuous & Discrete Random Variables: Randomly
select a camera and record the recycle time of a flash (𝑥).
𝑆 = 𝑅! = {𝑥 | 𝑥 > 0}, i.e., the positive real numbers.
v Suppose it is known that all recycle times are between
1.5 and 5 seconds.
𝑆 = {𝑥 | 1.5 < 𝑥 < 5} is continuous.
v It is known that the recycle time has only three values
(low, medium or high).
𝑆 = {𝑙𝑜𝑤, 𝑚𝑒𝑑𝑖𝑢𝑚, ℎ𝑖𝑔ℎ} is finite discrete.
v Whether the camera conforms to minimum recycle time
specifications or not. 𝑆 = {𝑦𝑒𝑠, 𝑛𝑜} is discrete.
v Cameras are tested until the flash recycle time fails to
meet the specs.
S = {n, yn, yyn, yyyn, yyyyn, …} is countably infinite.
5
Example 2.3 | Message Delays
Tree Diagrams: Sample spaces can also be described
graphically with tree diagrams.
Messages are classified as on-time (o) or late (l) within
the time specified by the system design.

Figure 2-5 Tree diagram for three messages

S = {ooo, ool, olo, oll, loo, lol, llo, lll } 6
Events

An event is a subset of the sample space of a

random experiment.

Event combinations
v The union of two events consists of all outcomes that are
contained in either of the two events, denoted as 𝐸1 ∪ 𝐸2.
v The intersection of two events is the event that consists of
all outcomes that are contained in both of the two events,
denoted as 𝐸1 ∩ 𝐸2.
v The complement of an event in a sample space is the set
of outcomes in the sample space that are not in the event.
We denote the complement of the event 𝐸 as 𝐸! 𝑜𝑟𝐸" .

7
Example 2.6 | Events

Sample Spaces & Events: Suppose that the recycle times of

two cameras are recorded. Consider only whether or not the
cameras conform to the manufacturing specifications. We
abbreviate yes and no as y and n.
Consider the sample space 𝑆 = {𝑦𝑦, 𝑦𝑛, 𝑛𝑦, 𝑛𝑛}.
v Suppose that the subset of outcomes for which at least one
camera conforms is denoted as 𝐸1. Then, 𝐸1 = {𝑦𝑦, 𝑦𝑛, 𝑛𝑦}.
v Suppose that the subset of outcomes for which both
cameras do not conform, denoted as 𝐸2, contains only the
single outcome, 𝐸2 = {𝑛𝑛}.
v Other examples of events are 𝐸3 = Ø (the null set) and 𝐸4 = 𝑆
(the sample space).
v Suppose that the subset of outcomes for which at least one
camera doesn’t conform is denoted as 𝐸5.
Then 𝐸5 = {𝑦𝑛, 𝑛𝑦, 𝑛𝑛}, 𝐸# ∪ 𝐸$ = 𝑆, 𝐸# ∩ 𝐸$ = 𝑦𝑛, 𝑛𝑦 , 𝐸#! = 𝑛𝑛 8
Example 2.7 | Camera Recycle Time

Union, Intersection & Complement of Events:

Camera recycle times (x) might use the sample
space S = R+.
Let E1 = {x | 10 ≤ x < 12} and E2 = {x | 11 < x <
15}. Then,

9
Venn Diagrams
Venn Diagrams: We can use Venn Diagrams to
represent a sample space of the random experiment
and events in a sample space.

Figure 2-8 Venn Diagrams of events 10

Mutually Exclusive Events

Two events denoted as A and B such that

A∩ 𝐵 = Ø are said to be mutually exclusive.

Figure 2-9 Mutually exclusive events

11
Mutually Exclusive Events - Laws
v The definition of the complement of an event
implies that 𝐸 ! ! = 𝐸

v The distributive law for set operations implies:

𝐴 ∪ 𝐵 ∩ 𝐶 = 𝐴 ∩ 𝐶 ∪ (𝐵 ∩ 𝐶)
𝐴∩𝐵 ∪𝐶 = 𝐴∪𝐶 ∩ 𝐵∪𝐶

v De Morgan’s laws imply that

𝐴 ∪ 𝐵 ! = 𝐴! ∩ 𝐵′
𝐴 ∩ 𝐵 ! = 𝐴! ∪ 𝐵′

v Also, remember that

𝐴 ∩ 𝐵 = 𝐵 ∩ 𝐴 and 𝐴 ∪ 𝐵 = 𝐵 ∪ 𝐴
12
Counting Techniques
vDetermining the outcomes in the sample
space (or an event) can sometimes become
more difficult.
vIn these cases, counting techniques where
the total number of outcomes in the sample
space and various events are used to
analyze the random experiments.
vThese methods are referred to as counting
techniques:
• Multiplication Rule
• Permutations
• Combinations 13
Counting Techniques: Multiplication
Rule
Assume an operation can be described as a
sequence of k steps, and
v The number of ways to complete step 1 is n1, and
v The number of ways to complete step 2 is n2 for
each way to complete step 1, and
v The number of ways to complete step 3 is n3 for
each way to complete step 2, and so fourth

The total number of ways to complete the

operation is 𝑛1×𝑛! × ⋯× 𝑛𝑘.

14
Example 2.9 | Web Site Design

Multiplication Rule: The design for a

Website is to consist of four colors, three
fonts, and three positions for an image.
From the multiplication rule, the total
number of outcomes:
4 x 4 x 3 = 36 different designs are possible.

Practical Interpretation: The use of the

multiplication rule and other counting
techniques enables one to easily determine
the number of outcomes in a sample space or
event and this, in turn, allows probabilities
of events to be determined.
15
Counting Techniques: Permutations

A permutation of the elements is an ordered

sequence of the elements.

Consider a set of elements, such as S = {a, b, c}.

Here abc, acb, bac, bca, cab, cba are all the
permutations of the elements of S.

The number of permutations of n different

elements is n! where the result follows from the
Multiplication Rule:

n!= n×(n−1)×(n−2)×...×2×1
Linear Permutation
16
Counting Techniques: Permutations
Permutations of Subsets
The number of permutations of subsets of r elements
selected from a set of n different elements is
!!
# = !"! ! #$"! ! %$&&&"! ! " + #$ =
"
!

"! ! " $!

Example 2.10 | Printed Circuit Board

A printed circuit board has eight different locations in
which a component can be placed. If four different
components are to be placed on the board, how many
different designs are possible?
"( " ! ' ! ! ! & ! )(
! =)
"
= = " ! ' ! ! ! & = %$!"#
+" " )*( )(
different designs are possible 17
Counting Techniques: Permutations

Permutations of Similar Objects

The number of permutations of n = n1 + n2 + … + nr objects
of which 𝑛1 are of one type, 𝑛2 are of a second type, … , and 𝑛𝑟
are of an 𝑟 "# type is
"#
"! #"" #$%%%$"! #

Example 2.9 | Hospital Schedule

A hospital operating room needs to schedule three knee
surgeries (k) and two hip surgeries (h) in a day. The
number of possible sequences of three knee and hip
surgeries is 5! 5⋅ 4 ⋅ 3!
= = 10
2!⋅ 3! 2 ⋅1⋅ 3!
The 10 sequences are easily summarized as:
{𝑘𝑘𝑘ℎℎ, 𝑘𝑘ℎ𝑘ℎ, 𝑘𝑘ℎℎ𝑘, 𝑘ℎ𝑘𝑘ℎ, 𝑘ℎ𝑘ℎ𝑘, 𝑘ℎℎ𝑘𝑘, ℎ𝑘𝑘𝑘ℎ, ℎ𝑘𝑘ℎ𝑘, ℎ𝑘ℎ𝑘𝑘, ℎℎ𝑘𝑘𝑘}
18
Counting Techniques: Combinations

The number of combinations, subsets of r elements that

can be selected from a set of n elements, is denoted as

4 4 4!
𝐶3 = 3 =3! 463 !

v In combinations, order is not important.

v Every subset of r elements can be indicated by
listing the elements in the set and marking each
element with a “*” if it is to be included in the
subset.
v Therefore, each permutation of r*s and (n – r)
blanks indicates a different subset.
v For example, if the set is S = {a, b, c, d}, the subset
{a, c} can be indicated as:
19
Example 2.14 | Sampling without
Replacement
In random experiments in which items are
selected from a batch, an item may or may not be
replaced before the next one is selected. This is
referred to as sampling with or without
replacement, respectively.

A bin of 50 parts contains 3 defectives and 47

non-defective parts. A sample of 6 parts is
selected from the 50 without replacement. How
many samples of size 6 contain 2 defective parts?

v 1st step is to choose 2 defective parts from the 3

defective parts:
3!
= 3 different ways
2!⋅1! 20
Example 2.14 | Sampling without
Replacement
v 2nd step is to select the remaining 4 non-defective
parts from the 47 acceptable parts in the bin:
47! 47 ⋅ 46 ⋅ 45⋅ 44 ⋅ 43!
= = 178, 365 different ways
4!⋅ 43! 4 ⋅ 3⋅ 2 ⋅1⋅ 43!

From the multiplication rule, the number of subsets

of size 6 that contain exactly 2 defective parts is:
3⋅178, 365 = 535, 095

As an additional computation, the total number of

different subsets of size 6 is:
50 50!
( )= = 15,890, 700
6 6!44! 21
Interpretations and Axioms of
Probability

Probability is used to quantify the

likelihood, or chance, that an outcome of a
random experiment will occur.
vThe likelihood of an outcome is
quantified by assigning a number from
the interval [0,1] to the outcome (or a
percentage from 0 to 100%).
v0 indicates an outcome will not occur.
v1 indicates that an outcome will occur
with certainty.
22
Interpretations and Axioms of
Probability
Subjective probability, or degree of belief
v Different individuals will no doubt assign different
probabilities to the same outcomes
v Example: “The chance of rain today is 30%”

Relative frequency probability

v Another interpretation of probability is based on the
conceptual model of repeated replications of the random
experiment. It is interpreted as the limiting value of the
proportion of times the outcome occurs in n repetitions of
the random experiment as n increases beyond all bounds.
v Example: If we assign probability 0.2 to the outcome that
there is a corrupted pulse in a digital signal, we might
interpret this assignment as implying that, if we analyze
many pulses, approximately 20% of them will be
corrupted. 23
Interpretations and Axioms of
Probability
Equally Likely Outcomes: Whenever a sample space
consists of n possible outcomes that are equally
likely, the probability of each outcome is 1/n.

v When tossing a coin, each outcome “head” or “tail”

is equally likely with probability of 1/2.
v For example, suppose that we select 1 laser diode
randomly from a batch of 100.
• Randomly implies that it is reasonable to assume
that each diode in the batch has an equal chance of
being selected.
• The probability model for this experiment assigns
probability of 1/100 = 0.01 to each of the 100
outcomes, because each outcome in the sample
space is equally likely. 24
Interpretations and Axioms of
Probability

vFor a discrete sample space, the

probability of an event can be defined
by the reasoning used in the
preceding example.
vProbability of an event: For a discrete
sample space, the probability of an
event E, denoted as P(E), equals the
sum of the probabilities of the
outcomes n in E:
𝑃 𝐸 = 𝑃 𝐸! + 𝑃 𝐸" +….+ 𝑃 𝐸#
25
Example 2.16 | Probabilities of Events

v A random experiment can result in one of the

outcomes {a, b, c, d} with probabilities 0.1, 0.3, 0.5,
0.1, respectively.
v Let A denote the event {a, b}, B the event {b, c, d},
and C the event {d}:
• 𝑃(𝐴) = 0.1 + 0.3 = 0.4
• 𝑃(𝐵) = 0.3 + 0.5 + 0.1 = 0.9
• 𝑃(𝐶) = 0.1
• 𝑃 𝐴¢ = 1 − 𝑃 𝐴 = 1 − 0.4 = 0.6 𝑎𝑛𝑑 𝑃 𝐵¢ = 1 − 0.9 =
0.1 𝑎𝑛𝑑 𝑃 𝐶¢ = 1 − 0.1 = 0.9
• Because 𝐴 ∩ 𝐵 = {𝑏}, 𝑡ℎ𝑒𝑛 𝑃(𝐴 ∩ 𝐵) = 0.3
• Because 𝐴 ∪ 𝐵 = {𝑎, 𝑏, 𝑐, 𝑑}, 𝑡ℎ𝑒𝑛 𝑃(𝐴 ∪ 𝐵) = 1.0
• Because 𝐴 ∩ 𝐶 = {𝑛𝑢𝑙𝑙}, 𝑡ℎ𝑒𝑛 𝑃(𝐴 ∩ 𝐶 ) = 0
• Because 𝐴∪C = {𝑎,𝑏,𝑑}, then 𝑃(𝐴∪C) = 0.5 26
Axioms of Probability

Probability is a number that is assigned to

each member of a collection of events from a
random experiment that satisfies the
following properties:
1. P(S) = 1
2. 0 ≤ P(E) ≤ 1
3. For any two events E1 and E2 with 𝐸1 ∩ 𝐸2 = Ø,
0
𝑃(𝐸1 ∪ 𝐸$ ) = 𝑃(𝐸1) + 𝑃(𝐸2) - P(𝐸1 ∩ 𝐸2 )
The axioms imply that:
• 𝑃(Ø) = 0 and 𝑃(𝐸′ ) = 1 – 𝑃(𝐸)
• If E1 is contained in E2, then 𝑃(𝐸1) ≤ 𝑃(𝐸2). 27
Unions of Events and Addition Rules

vJoint events are generated by applying

basic set operations to individual events,
specifically:
• Unions of events, 𝐴 ∪ 𝐵
• Intersections of events, 𝐴 ∩ 𝐵
• Complements of events, 𝐴¢ or 𝐴7

vProbabilities of joint events can often be

determined from the probabilities of the
individual events it comprises.

28
Example 2.19 | Semiconductor Wafers
Joint Events: A wafer is randomly selected from
a batch of 940 wafers in a semiconductor
manufacturing process.
vLet H denote the event that the wafer contains
high levels of contamination
• Then P(H) = 358/940.
vLet C denote the event of the wafer is in center of
a sputtering tool
• Then P(C) = 626/940.
)*;T$3*#<*=<6**8
!*#$T>3#T$3*# 6*$T8
!"#$"% C'E"
)*H ,-. L0 ,01
23E4 --1 1.L 5,0
6*$T8 L1L 5-. 9.:
29
Example 2.19 | Semiconductor Wafers

𝑃(𝐻 ∩ 𝐶) is the probability that the wafer is from the center

of the sputtering tool and contains high levels of
contamination:
𝑃(𝐻 ∩ 𝐶) = 112/940
The event (𝐻 ∪ 𝐶) is the event that a wafer is from the
center of the sputtering tool or contains high levels of
contamination (or both):
𝑃 𝐻 ∪ 𝐶 = 𝑃 𝐻 + 𝑃 𝐶 − 𝑃(𝐻 ∩ 𝐶)
= (358 + 626 - 112)/940 = 872/940
Practical Interpretation: To )*;T$3*#<*=<6**8
!*#$T>3#T$3*# 6*$T8
better understand the sources !"#$"% C'E"
of contamination, yield from )*H ,-. L0 ,01
different locations on wafers 23E4 --1 1.L 5,0
are routinely aggregated. 6*$T8 L1L 5-. 9.:
30
Addition Rule - Two Events

Probability of a Union

v For any events A and B:

𝑃 𝐴∪𝐵 = 𝑃 𝐴 + 𝑃 𝐵 −𝑃 𝐴∩𝐵

vFor mutually exclusive events A and B:

0
𝑃 𝐴∪𝐵 = 𝑃 𝐴 + 𝑃 𝐵 −𝑃 𝐴∩𝐵
“a null set”

31
Addition Rule – Multiple Events

Probability of a Union

v For any events A, B and C:

𝑃 𝐴∪𝐵∪𝐶 = 𝑃 𝐴 + 𝑃 𝐵 + 𝑃 𝐶 −𝑃 𝐴∩𝐵
− 𝑃 𝐴 ∩ 𝐶 −𝑃 𝐵 ∩ 𝐶 + 𝑃 𝐴 ∩ 𝐵 ∩ 𝐶

v For mutually exclusive events:

A collection of events 𝐸% , 𝐸$ , … , 𝐸& are said to be

mutually exclusive if for all pairs 𝐸' ∩ 𝐸( = Ø.

For a collection of mutually exclusive events,

𝑃 𝐸% ∪ 𝐸$ ∪ ⋯ ∪ 𝐸& = 𝑃 𝐸% + 𝑃 𝐸$ + ⋯ + 𝑃 𝐸&
32
Conditional Probability
v Probabilities should be re-evaluated as additional
information becomes available. Once we get new
information, probabilities will change.
• Conditional Probability ≜ Change of Sample Space
v A useful way to incorporate additional information into a
probability model is to assume that the outcome that will be
generated is a member of a given event. The probability that
one event happens given that another event is already
known to have happened is called conditional probability of
B given A.
v 𝑃(𝐵|A): The probability of event 𝐵 occurring, given that
event 𝐴 has already occurred.
v Example: The probability of a Turkish person knowing a
programming language where the person is a Medipol
33
engineering student.
Conditional Probability Rule
v The conditional probability of event A given event B
occured, denoted as 𝑃(A|B) is:

!(#∩%)
𝑃 A|B = for 𝑃 𝐵 >0
!(%)

v From a relative frequency perspective

of n equally likely outcomes:
• 𝑃(𝐴) = (number of outcomes in 𝐴)/n
• 𝑃(𝐵) = (number of outcomes in 𝐵)/n
• 𝑃(𝐴 ∩ 𝐵) = (number of outcomes in A ∩ B )/n
! #∩% number of outcomes in A ∩ B /n
• 𝑃 A|B = =
! % number of outcomes in B /)
In conditional probability, we change the sample space
to the observed events.
34
Example 2.22 | Surface Flaws &
Defectives –Venn Diagram
Conditional Probability: Table 2- 𝑆 = 400
342 𝐹
3 shows that 400 parts are 𝐷
classified by surface flaws and as
functionally defective. Observe 18 10 30
that:
𝑃(𝐷|𝐹) = 10/40 = 0.25
𝑃(𝐷|𝐹 # ) = 18/360 = 0.05
Practical Interpretation
/0?1"*5@6**A0<%)*B10))F#F"C The probability of being defective
is five times greater for parts with
:;<#0$"*=10>)
surface flaws. This calculation
!"#"$%F'" (")*+! , -.*+!" , /.%01 illustrates how probabilities are
(")*+# , 23 24 54 adjusted for additional information.
The result also suggests that there
-.*+#" , 63 675 685 may be a link between surface
/.%01 73 693 733 flaws and functionally defective
parts, which should be
investigated. 35
Example 2.22 | Surface Flaws and
Defectives – Venn Diagram
Refer to Table 2-3 again. /0?1"*5@6**A0<%)*B10))F#F"C
:;<#0$"*=10>)
There are 4 probabilities !"#"$%F'" (")*+! , -.*+!" , /.%01
conditioned on flaws: (")*+# , 23 24 54
In some books, 𝑃(𝐷 ∩ 𝐹) -.*+#" , 63 675 685
is expressed as 𝑃(𝐷, 𝐹). /.%01 73 693 733

10
𝑃(𝐷 ∩ 𝐹) 400 10
𝑃 D|𝐹 = = =
𝑃(𝐹) 40 40
400
𝑃(𝐷) = 28/400 𝑃(𝐹) = 40/400
𝑃 D|𝐹 = 10/40 𝑃 F|𝐷 = 10/28
𝑃 D’|𝐹 = 30/40 P F’|𝐷 = 18/28
𝑃 D|𝐹′ = 18/360 𝑃 F|𝐷′ = 30/372
𝑃 D’|𝐹′ = 342/360 𝑃 F’|𝐷′ = 342/372 36
Example 2.23 | Surface Flaws &
Defectives –Tree Diagram
Sampling without Replacement: The tree diagram
illustrates sampling two parts without replacement:
v At the 1st stage (surface flaw), every original part of
the 400 is equally likely.
v At the 2nd stage (defective), the probability is
conditional upon the part drawn in the prior (1st)
stage. /0?1"*5@6**A0<%)*B10))F#F"C
:;<#0$"*=10>)
!"#"$%F'" (")*+! , -.*+!" , /.%01
(")*+# , 23 24 54
-.*+#" , 63 675 685
/.%01 73 693 733

Figure 2-14 Tree diagram for parts classification

37
Example| Heights & Weights of
Students (Mutually Exclusive Events)
Table Joint probabilities for heights and weights of college students

v What is the sum of all probabilities P(Hi) à It is 1.

v What is the probability that a student has a height between
H2 and H4, and that his weight is between W2 and W4,
inclusive.
• Since the sample space is discrete, we sum all probabilities in
the region: 0.12 + 0.06 + 0.02 + 0.06 + 0.014 +…+ 0.10 = 0.64
v Given that a student is heavier than W3, what is the
probability that his height is H5?
• We sum probabilities in for H5, when weight is W4 or W5 and
normalize for the given probability: P(H5|>W4 or W5)=0.12/0.34 38
Axioms of Conditional Probability
v Axiom 1
𝑃 A|B =
,(-∩.)
≥ 0 for 𝑃(𝐴 ∩ 𝐵) ≥ 0 and P(B) > 0. 𝑆
,(.)
𝐴 𝐵
v Axiom 2
𝑃(𝑆 ∩ 𝐵) 𝑃(𝐵)
𝑃 S|B = = =1
𝑃(𝐵) 𝑃(𝐵)

v Axiom 3: If A and C are mutually exclusive

events, then 𝐴 𝐶
𝑃[(𝐴 ∪ 𝐶) ∩ 𝐵] 𝑃[ 𝐴 ∩ 𝐵 ∪ (𝐶 ∩ 𝐵)]
𝑃 𝐴∪𝐶 𝐵 = =
𝑃[𝐵] 𝑃[𝐵]
,[ -∩. ] ,[(1∩.)]
= + =𝑃 𝐴 𝐵 +𝑃 𝐶 𝐵
,[.] ,[.]
𝐵
What is P(A|A) and P(B|B)? à 1 39
Random Samples & Conditional
Probability
v Random means each item is equally likely to be
chosen.
• If more than one item is sampled, random means
that every sampling outcome is equally likely.
Example: In the random experiment of dice rolling,
the probability of each outcome is 1/6 every time
you roll it.
v 2 items are taken from the batch 𝑆 = {𝑎, 𝑏, 𝑐}
without replacement.
• Ordered sample space: 𝑆 = {𝑎𝑏, 𝑎𝑐, 𝑏𝑐, 𝑏𝑎, 𝑐𝑎, 𝑐𝑏}
P(each outcome in S)=1/6
• Unordered sample space: 𝑆 = {𝑎𝑏, 𝑎𝑐, 𝑏𝑐}
• This is done by enumeration – too hard for big
sample spaces L
• Use conditional probability to avoid enumeration. 40
Sampling Without Enumeration
v A batch of 50 parts contain 10 made by
Tool 1 (red) and 40 made by Tool 2 (blue).
We take a sample of n=2 parts from the batch.
v What is the probability that the 2nd part came
from Tool 2, given that the 1st part came from Tool 1?
• P(1st part came from Tool 1) = 10/50
• P(2nd part came from Tool 2 | 1st part came from Tool 1 ) = 40/49
v “To select randomly” = At each step of the sample, the items
remaining in the batch are equally likely to be selected.
v Let E be, outcomes with the first selected part from tool 1 and
the second part from tool 2. Then,
𝑃 𝐸 = 𝑃 𝐸1 ∩ 𝐸2 = 𝑃 𝐸' 𝐸( . 𝑃(𝐸()
P(Part from Tool 1, then from Tool 2 sequence) = (40/49).(10/50)
Sometimes a partition of the question into successive picks is an
41
easier method to solve the problem.
Example 2-24| Sampling Without
Replacement

A production lot of 50 parts contain 3

defectives. Two parts are selected at
random. What is the probability that
the 2nd is defective, given that the first
part is defective?
vA: The event that the 1st part selected is
defective.
vB: The event that the 2nd part selected is
defective.
vProbability desired is P (B|A) = 2/49.
42
Cafe Talk: Another Birthday Problem

v Cevdet has two children. The older child is a boy.

What is the probability that both kids are boys?
• The sample space S = {BB, BG} à P[BB] = ½.

v Ali has two children. At least one of them is a boy.

What is the probability that both kids are boys?
• Since we don’t know whether the boy is older or
younger than the other sibling, our sample space
here is larger: S = {BB, BG, GB} à P[BB] = 1/3.
• Using the conditional probability, we obtain the
same result, however the sample space changes to S
= {BB, BG, GB, GG} à
P BB = P The other sibling boy One sibling boy) =
)(*+,-+./ 1 2.3 4+,-+./ 5 267 ,894) 1/< 1
)(:.7 4+,-+./ +4 ,89)
= =/< = = 43
The “Monty Hall Problem”

v Monty Hall is (was?) the host of a

TV game show called “Let’s Make
a Deal”. On the show there are 3
doors, behind one of which is a
car, while the rest have goats.
v Monty Hall, the host, asks you to
pick a door.
v Let’s say you pick door 3.
v Monty Hall opens (e.g.) door 1
and shows there is a goat behind
the door 1. You are now given the
choice of either sticking with your
original choice of door 3 or
switching to door 2.
Should you switch? 44
The “Monty Hall Problem”: Solution

You Should
SWITCH !!!!
Try for the case where there are 100 doors, where Monty
Hall opens 98 doors and leaves you with only two choices. 45
Multiplication Rule
Multiplication Rule (Two Events)
v Definition of conditional probability:
)(+#∩+$)
𝑃(𝑋1 | 𝑋2) = )(+$)
v The conditional probability definition can be
rewritten to generalize it as the multiplication rule
which gives an alternative, more intuitive
formulation:
𝑃 𝑋1 ∩ 𝑋2 = 𝑃 𝑋1 𝑋2 . 𝑃 𝑋2 = 𝑃 𝑋2 𝑋1 . 𝑃(𝑋1)

This expression is derived by

exchanging the roles of X1 &
X2.

Multiplication Rule (More Than Two Events):

𝑃 𝑋% ∩ ⋯ ∩ 𝑋𝑛
= 𝑃 𝑋1 ∩ ⋯ ∩ 𝑋𝑡 . 𝑃(𝑋./% ∩ ⋯ ∩ 𝑋𝑛 | 𝑋1, … , 𝑋𝑡) 46
Example 2-26| Machining Stages
Multiplication Rule: The probability that, a part made in
the 1st stage of a machining operation passes inspection,
is 0.90. The probability that, after it passes the first
inspection that it passes inspection again after the 2nd
stage, is 0.95. What is the probability that a part meets
specifications?

1st Stage 2nd Stage

Let A & B denote the events that the 1st & 2nd stages meet
specifications.
From the Multiplication Rule we have:

𝑃 𝐴 ∩ 𝐵 = 𝑃 B|𝐴 . 𝑃(𝐴) = (0.95). (0.90) = 0.85

Why didn’t we use the expression 𝑃 A|B . 𝑃(𝐵)?
Also, can we determine 𝑃[𝐵] ? 47
Chain Rule

Multiplication rule general form:

𝑃(𝑋1 ∩ ⋯ ∩ 𝑋𝑛) =
= 𝑃(𝑋1 ∩ ⋯ ∩ 𝑋𝑡) 𝑃(𝑋234 ∩ ⋯ ∩ 𝑋𝑛 | 𝑋1, … , 𝑋𝑡)

Chain rule is derived by successive application of the

multiplication rule:
𝑃(𝑋1 ∩ ⋯ ∩ 𝑋564 ∩ 𝑋𝑛) =
= 𝑃 𝑋1 ∩ ⋯ ∩ 𝑋564 . 𝑃(𝑋𝑛 | 𝑋1, … , 𝑋564)

= 𝑃 𝑋1 ∩ ⋯ ∩ 𝑋567 . 𝑃(𝑋564| 𝑋1, … , 𝑋567) . 𝑃(𝑋𝑛 | 𝑋1, … , 𝑋564)

………
= 𝑃 𝑋1 . 𝑃(𝑋2 | 𝑋1) … 𝑃(𝑋564 𝑋1, … , 𝑋567 . 𝑃(𝑋𝑛 | 𝑋1, … , 𝑋564)
5
= Π894 𝑃 𝑋𝑖 𝑋1, … , 𝑋864) 48
Example 2-25| Sampling for Defects

Chain Rule: 3 parts are taken/sampled

randomly. What is the probability that the
first two are defective, while the third is
not?
hi kl jii
v𝑃 𝑑𝑑𝑛 = . . = 0.0032
jhi jkl jkj
• What is 𝑃 𝑑𝑛𝑑 and 𝑃 𝑑𝑛𝑛 ?

vWhich model did we use?

• Sampling with replacement?
• Or Sampling without replacement? 49
Total Probability Rule – Two Events
Two Mutually Exclusive Subsets
𝑆

Figure 2-15 Partitioning an event into two mutually exclusive subsets.

v A & A’ are mutually exclusive.

v (𝐴 ∩ 𝐵) and (𝐴! ∩ 𝐵) are mutually exclusive.
v Since A ∪ A’ = S, A and A’ are exhaustive.
v B = (𝐵 ∩ 𝐴) ∪ (𝐵 ∩ 𝐴! ) à 𝑃(𝐵) = 𝑃(𝐵 ∩ 𝐴) + 𝑃(𝐵 ∩ 𝐴! )

Total Probability Rule (Two Events)

𝑃(𝐵) = 𝑃(𝐵 ∩ 𝐴) + 𝑃(𝐵 ∩ 𝐴! )

= 𝑃(𝐵|𝐴). 𝑃(𝐴) + 𝑃 𝐵 𝐴! . 𝑃(𝐴! ) 50

Example 2-27 | Semiconductor
Contamination
Total Probability Rule with Two Events: Information about
product failure based on chip manufacturing process
contamination is shown in the table. What is the probability that
a chip fails?

!"#$%$F'F() /.0.'+#* !"#$%$F'F()

#*+H%F'-". 1#2(%$F2%(F#2 #*+/.0.'
34533 6F78 349
3433: ;#(+6F78 34<

v F: The chip fails.

H and H’ are
H: The chip has high contamination. mutually exclusive
events.
H’: The chip doesn’t have high contamination.
v 𝑃(𝐹|𝐻) = 0.100 and 𝑃 𝐻 = 0.2 → 𝑃(𝐹 ∩ 𝐻) = 𝑃(𝐹|𝐻). 𝑃(𝐻) = 0.02.
𝑃(𝐹|𝐻! ) = 0.005 and 𝑃 𝐻! = 0.8 → 𝑃(𝐹 ∩ 𝐻′) = 𝑃(𝐹|𝐻′). 𝑃(𝐻′) = 0.004.
From the total probability rule:
𝑃(𝐹) = 𝑃(𝐹 ∩ 𝐻) + 𝑃(𝐹 ∩ 𝐻! ) = 0.02 + 0.004 = 0.024 51
Total Probability Rule – Multiple Events
Multiple Mutually Exclusive Subsets
𝑆

Figure 2-16 Partitioning an event into several mutually exclusive subsets.

Total Probability Rule (More Than Two Events)

Let E1, E2, …, Ek be k mutually exclusive & exhaustive
subsets:
𝑃 = 𝑃 𝐵 ∩ 𝐸4 + 𝑃 𝐵 ∩ 𝐸7 + ⋯ + 𝑃 𝐵 ∩ 𝐸:
= 𝑃 𝐵|𝐸4 . 𝑃𝐸4 + 𝑃 𝐵|𝐸7 . 𝑃 𝐸7 + ⋯ + 𝑃 𝐵|𝐸: . 𝑃(𝐸: )
𝑃 𝐵 = ? 𝑃 𝐵 𝐸) ⋅ 𝑃 𝐸)
52
)
Example 2-28 | Refined Contamination
Data
Total Probability Rule (Multiple Events):
Continuing the discussion of contamination during chip
manufacture, the probabilities are given on the right side.
What is the probability that a chip fails?

!"#$%$F'F() /.0.'+#* !"#$%$F'F()

#*+H%F'-". 1#2(%$F2%(F#2 #*+/.0.'
34533 6F78 349
34353 :.;F-< 34=
34335 /#L 34?

v F: The event that the product fails.

H: The event that the chip has high cont. H, M and L are
mutually
M: The event that the chip has low cont. exclusive
L: The event that the chip has medium cont. events.
53
 Example 2-28 | Refined Contamination
Data
!"#$%$F'F() /.0.'+#* !"#$%$F'F()
#*+H%F'-". 1#2(%$F2%(F#2 #*+/.0.'
34533 6F78 349 Figure 2-17 Tree diagram for three
34353 :.;F-< 34= events
34335 /#L 34?

𝑃(𝐹 ∩H) 𝑃(𝐹′ ∩ H) 𝑃(𝐹 ∩ M) 𝑃(𝐹′ ∩ M) 𝑃(𝐹 ∩ L) 𝑃(𝐹′ ∩ L)

Since we desire to
From the total probability rule: find the probability
of failing parts, we
𝑃 𝐹 = 𝑃 𝐹 𝐻 .𝑃 𝐻 + 𝑃 𝐹 𝑀 .𝑃 𝑀 + 𝑃 𝐹 𝐿 .𝑃 𝐿 only consider
= 𝑃(𝐹 ∩ H) + 𝑃(𝐹 ∩ M) + 𝑃(𝐹 ∩ L) 𝑃(𝐹 ∩ H), 𝑃(𝐹 ∩ M)
and 𝑃(𝐹 ∩ L).
= 0.02 + 0.003 + 0.0005 =0.0235
What is the probability that a chip does not fail? 54
Independence - Two Events
v Two events are statistically
independent if any one of the
following equivalent statements B
are true:
1. P(B|A) = P(B)
A
2. P(A|B) = P(A)
3. P(A∩ B) = P(A). P(B)
A B
vThis means that occurrence of one
event has no impact on the occurrence
of the other event.
• If P B A = P B , then the outcome of the
experiment is in event A does not affect the
probability that the outcome is in event B.
• Can A and B happen at the same time? Can
you give some examples? 55
Example 2-29| Sampling with
Replacement
Sampling with Replacement: A production lot of 50
parts contains 3 defectives. Two parts are selected at
random, but the first is replaced before selecting the 2nd.
v A: The event that the 1st part selected is defective.
𝑃(𝐴) = 3/50

v B: The event that the 2nd part selected is defective.

𝑃(𝐵) = 3/50

v What is the probability that the 2nd is defective, given

that the first part is defective?
P(B|A)= 3/50

v What is probability that both are defective?

As P(B|A) = 𝑃(𝐵) = 3/50, events A and B are independent.
Due to independence between A and B we have:
𝑃 𝐴 ∩ 𝐵 = 𝑃 𝐴 . 𝑃(𝐵) = (3/50).(3/50) = 9/2500. 56
Example 2-30| Flaw & Functions
Independence Test (Two Events): The data in the tables
below shows whether the events are independent or not:

+, +.
P(D|F)= 0.25 ≠ 𝑃(𝐷) = -.. = 0.07 P(D|F)=0.05= 𝑃(𝐷) = -.. = 0.05

Events D and F are Events D and F are

dependent. independent.
57
Statistical Independence vs.
Mutual Exclusiveness
v If A and B are mutually exclusive then if B occurs,
then A cannot occur.
• P(A ∩ B) = 0
• P(A|B)= 0
• P(A U B) = P(A) + P(B)
v If A and B are statistically independent then if B
occurs, then A can occur
• P(A|B) = P(A)
• P(A ∩ B) = P(A).P(B)
v Then, it can be shown that
• 𝑃 𝐴! ∩ 𝐵! = 𝑃 𝐴! . 𝑃(𝐵! )
v Can two events A and B be both mutual exclusive
and independent?
Two events can be mutual exclusive and independent if it is
58
guaranteed that one of them cannot occur.
Example 2.31| Conditioned vs.
Unconditioned
Independence Test (Two Events): A production lot
of 50 parts contains 3 defectives. Two parts are
selected at random, without replacement.
v A: The event that the 1st part selected is defective:
𝑃(𝐴) = 3/50
v B: The event that the 2nd part selected is defective:
𝑃(𝐵) = 𝑃 𝐵 𝐴 . 𝑃(𝐴) + 𝑃 𝐵 𝐴# . 𝑃(𝐴# )
𝑃(𝐵)= (2/49).(3/50) + (3/49).(47/50)
𝑃(𝐵) = 3/50 is unconditional, same as 𝑃 𝐴 .
v Since 𝑃(𝐵|𝐴) ≠𝑃(𝐵), then A and B are dependant!
v Since P(A ∩ B) =𝑃 𝐵 𝐴 . 𝑃(𝐴) = (2/49).(3/50) ≠ 0, then
A and B are not mutually exclusive! 59
Independence - Multiple Events
The events E1, E2, … , Ek are independent if
and only if, for any subset of these events the
following two conditions hold:

1. 𝑃 𝐸'% ∩ 𝐸'$ ∩ 𝐸'& = 𝑃 𝐸'% . 𝑃 𝐸'$ … 𝑃(𝐸'& )

2. Joint independent implies pairwise
independence, but not the other way around:

𝑃 𝐸% ∩ 𝐸$ = 𝑃 𝐸% ∗ 𝑃[𝐸$ ]
𝑃 𝐸% ∩ 𝐸0 = 𝑃 𝐸% ∗ 𝑃 𝐸0
⋮
𝑃 𝐸$ ∩ 𝐸0 = 𝑃 𝐸$ ∗ 𝑃[𝐸0 ]
𝑃 𝐸$ ∩ 𝐸1 = 𝑃 𝐸$ ∗ 𝑃[𝐸1 ]
⋮ 60
Example 2-33| Semiconductor Wafers

Independence Test (Multiple Events): The

probability that a wafer contains a large particle of
contamination is 0.01. The wafer events are
independent. That is, the probability that a wafer
contains a large particle does not depend on the
characteristics of any of the other wafers.
If 15 wafers are analyzed, what is the probability that
no large particles are found?
v Ei: The event that the ith wafer contains no large
particles
v P(Ei) = 1- 0.01 = 0.99.
v Due to independence of wafers with k=15 events:
P(E1∩E2 ∩ … ∩ E15) = P(E1).P(E2)…P(E15)
= (0.99)15 = 0.86. 61
Example 2-32/34| Series/Parallel
Circuit
Independence Test (Two Events): This circuit operates only
if there is a path of functional devices from left to right. The
probability that each device functions is shown below. Each
device fails independently. What is the probability that the
circuit operates?
𝑇

𝐴 𝐵
𝐵
𝑃 𝐴 𝑎𝑛𝑑 𝐵 = 𝑃 𝐴 ∩ 𝐵 Let T & B denote the events that
= 𝑃 𝐴 . 𝑃 𝐵 = 0.8 . 0.9 the top and bottom devices operate.
There is a path if at least one of
A & 𝐵 𝑎𝑟𝑒 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 them operates 𝑃(𝑇 𝑜𝑟 𝐵) = 𝑃(𝑇 ∪
Since there is a path if 𝐵) = 1 − 𝑃 𝑇 ∪ 𝐵 ! = 1 − 𝑃(𝑇 ! ∩ 𝐵! ) =
both operate, P(Circuit 1 − (𝑃 𝑇 ! . 𝑃(𝐵! ) = 1 − 0.05+ = 0.9975.
operates)= 0.72.
T & 𝐵 𝑎𝑟𝑒 𝑖𝑛𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑡 62
Example 2-35|Advanced Circuit
Independence Test (Multiple Events): A series of parallel
circuits: This circuit operates only if there is a path of
functional devices from left to right. The probability that
each device functions is shown. Each device fails
independently. What is the probability that the circuit
operates?

𝑅
Events L, M
𝑀 and R are
𝐿 independent.
Partition the graph into 3 columns with L, M and R.
P(L) = 1- 0.13, P(M) =1- 0.52, and P(R)= 0.99.
P(The circuit operates) = 𝑃 𝐿 ∩ 𝑀 ∩ 𝑅 =P(L).P(M).P(R)
= (1 – 0.13).(1-0.052).(0.99) = 0.9875 63
Independent Events via Venn Diagram

v It is given that the events 𝐴 and 𝐵 are independent.

v 𝑃 𝐴 = 0.4 and 𝑃(𝐵) = 0.5. 𝑆=1
v What is P(A|B)? 𝐴
0.3
𝐵
0.2 0.2 0.3

P(S) = 1
P(A and B) = P(A∩B) = P(A).P(B) = 0.2
P(A and B’ ) = P(A∩B’ ) = 0.2
P(A’ and B) = P(A’∩B) = 0.3
P(A or B)’ = P(A∪B)’ = 1- (0.2+0.2+0.3) = 0.3
) 2∩3 4.$
P(A conditioned on B) = P(A|B) = )(3)
= 4.6 64
Conditional Independence
v A and B are conditionally independent given that C occurs:
𝑃 𝐴∩𝐵 𝐶 =𝑃 𝐴 𝐶 ⋅𝑃 𝐵 𝐶
§ Example: A is the height of a child, and B is the number of words
that the child knows. It seems that when A is high, B is high, too.
The height and the number of words known by the kid are NOT
independent, but they are conditionally independent if you
provide the kid’s age (C).
§ However, independent events can become dependent after
conditioning as well.
v Which of the following is conditionally independent? à (a)
In Figure (b), when B happens, A and C cannot happen at the same
time which means if C occurs, then A cannot occur.
𝐴 𝐶
𝐴 𝐵
S

𝐶 𝐵
65
(a) (b)
Bayes’ Theorem

v According to Bayes idea, we observe conditional

probabilities through prior information.
v It is a systematic way of including new evidence.
v It allows us to find the probability of a cause (reason for
failure of printers: hardware, software or other
problems) given its effect (failure).
v The conditional probabilities commonly provide the
probability of an event (such as failure) given a
condition (such as high or low contamination).àP(F|H)
v But after a random experiment generates an outcome,
we are naturally interested in the probability that a
condition was present (high contamination) given an
outcome (a semiconductor failure). In other words, we
desire to find the probability of failure due to high
contamination via Bayes’ Rule. àP(H|F). 66
Bayes’ Theorem
Bayes’ Theorem

Probability of outcome B, Priori probability of the

when the condition A already condition
occured

𝑃(𝐵|𝐴). 𝑃(𝐴) for 𝑃(𝐵) > 0

𝑃(𝐴|𝐵) =
𝑃(𝐵)
Posterior probability of 𝐴 or the Priori probability that
condition, the evidence (outcome)
given the outcome is true.

v Note the reversal of the condition! 𝑃 𝐵 𝐴 à 𝑃 𝐴 𝐵

v So, we can either check the outcome given the
condition, or given the outcome, we ask about the
condition. 67
Example 2-36| Bayes’ Theorem
Bayes’ Theorem (Two Events): What is the failure rate of chips
due to exposure to high contamination in Ex. 2-27?

𝑃 𝐹 𝐻 . 𝑃(𝐻) (0.10). (0.2)

𝑃(𝐻|𝐹) = = = 0.83
𝑃(𝐹) 0.024
Note that we first need to find
𝑃 𝐹 using the total probability rule as P F
depends on the P(H) and P(H’ ):
High cont. failure Not high cont. failure

𝑃 𝐹 = 𝑃 𝐹 𝐻 . 𝑃 𝐻 + 𝑃 𝐹 𝐻! . 𝑃(H’)
= 0.10 . (0.2) + (0.0050. (0.8) = 0.024 68
Example 2-37| Medical Diagnostic

Bayes’ Theorem (Two Events): During a medical screening of

the population, the probability that the test correctly
identifies someone with the disease as positive is 0.99, and
probability that the test correctly identifies someone without
the disease as negative is 0.95. The incidence of the disease
in the general population is 0.0001. You take the test and
the result is positive. What is the probability that you have
the disease?

D and D’ are
D: The event that you have the disease mutually
S: The event that the test signals positive exclusive events.

P(𝑆 + |𝐷+ ) = 0.95 so P(𝑆|𝐷+ ) = 1 − P(𝑆 + |𝐷+ ) = 0.05

P(D) = 0.0001 so P(𝐷+ ) = 1 − 0.0001 = 0.9999
P S D = 0.99, but we desire to find P(D|S).
69
Example 2-37| Medical Diagnostic
By Bayes’ Theorem:
, 𝑆𝐷 .,(;)
𝑃(𝐷|𝑆) =
, 𝑆𝐷 ., ; 3, 𝑆 𝐷# .,(;0 )

Total Probability Rule

0.99 . (0.0001) 1
= = = 0.002
0.99 . (0.0001) + (0.05). (1 − 0.0001) 506

Before the test, your chance of

having the disease was 0.0001. After
the positive result, your chance is
0.002. So, your risk of having the
disease has increased 20 times =
0.002/0.00010, but is still tiny.
Do you trust this result?
Why such a result?
70
Bayes’ Theorem - Multiple Events
Collectively Exhaustive Events
When the union of the events
forms the sample space itself, S
then such events are called B
exhaustive events and denoted as
E1∪E2∪ ⋯ ∪ Ek =S.
v It is guaranteed that one of the
events occurs.
v We also say that these events
partition the sample space.
If E1, E2, … Ek are k mutually exclusive and exhaustive
events and 𝐵 is any event, then according to the total
probability theorem for 𝑃 𝐵 > 0:
𝑃 𝐵|𝐸, . 𝑃(𝐸, )
𝑃(𝐸1|𝐵) =
𝑃 𝐵|𝐸, . 𝑃(𝐸, ) + 𝑃 𝐵|𝐸- . 𝑃(𝐸- ) + ⋯ + 𝑃 𝐵|𝐸. . 𝑃(𝐸. )
Note that the denominator expresses the total probability while the
71
numerator is always one term of the denominator.
Example 2-38| Bayesian Network

Bayes’ Theorem (Multiple Events): Bayesian

networks are used on Web sites of high-tech
manufacturers to allow customers to quickly
diagnose problems with products. A printer
manufacturer obtained the following probabilities
from its database. Printer failures are of 3 types:
hardware P(H) = 0.1, software P(S)=0.6, and other
P(O)=0.3. Also, P(F|H) = 0.9, P(F|S) = 0.2 and
P(F|O) = 0.5.

Find the maximum of P(H|F), P(S|F), P(O|F) to

direct the diagnostic effort. In other words, what is
the most probable cause of product problems?
72
Example 2-38| Bayesian Network
! ( " ) = ! ( " ! # ) ! ( # ) + ! ( " ! $ ) ! ( $ ) + ! ( " ! % ) ! (% )
= "#$ ( "#%) + "#& ( "#' ) + "#( ( "#)) = "#)'
! ( " ! # ) ! ! ( # ) "#$ ! "#%
!(# ! " ) = = = "#&("
!(" ) "#)'
! ( " ! $ ) ! ! ( $ ) "#& ! "#'
!($ ! " ) = = = "#)))
!(" ) "#)'
! ( " ! % ) ! ! ( % ) "#( ! "#)
! (% ! " ) = = = "#*%+
!(" ) "#)'

Note that the conditionals based on

Failure add to 1. Since the Other
category is the most likely cause of the
failure, diagnostic effort should be so
initially directed.
73
Important Terms & Concepts of Lecture 3

v Addition rule v Probability

v Axioms of probability v Random experiment
v Bayes’ theorem v Random variable
v Combination • Discrete
v Conditional probability • Continuous
v Equally likely outcomes v Sample space
v Event • Discrete
v Independence • Continuous
v Multiplication rule v Total probability rule
v Mutually exclusive v Tree diagram
events v Venn diagram
v Total outcomes v With replacement
v Without replacement
74
EXTRA SLIDES
An Example of Conditional
Probability
vIn a thin film manufacturing process,
the proportion of parts that are not
acceptable is 2%. However, the
process is sensitive to contamination
that can increase the rate of parts
rejection.
vIf we know that the plant is having
filtration problems that increase film
contamination, we will presume that
the rejection rate has increased.
76
Another Example of Conditional
Probability

Figure 2-13 Conditional probability of rejection for parts with surface flaws
and for parts without surface flaws. The probability of a defective part is
not evenly distributed. Flawed parts are five times more likely to be
defective than non-flawed parts, i.e., P(D|F) / P(D|F’).

77
The “Monty Hall Problem”: Solution

v We want to calculate:
P(1|MH opens 2) = probability the prize is behind
door 1 and MH opens door 2.
P(3|MH opens 2) = probability the prize is behind
door 3 and MH opens door 2
v P(A|MH opens 2) = probability the prize is behind
door 1 and MH opens door 2
P(1|MH opens 2) = P(MH opens 2|1)*P(1)/P(MH
opens 2)
P(3|MH opens 2) = probability the prize is behind
door 3 and MH opens door 3
P(3|MH opens 2) = P(MH opens 2|3)*P(3)/P(MH
opens 2)
78
The “Monty Hall Problem”: Solution
v Apriori, the prize can be behind doors 1, 2, 3 with equal
probability: P(1)=P(2)=P(3)=1/3
We need to calculate the probability that MH opens door 2:
P(MH opens 2)=P(MH opens 2|1)*P(1)+P(MH opens 2|2)*P(2) + P(MH
opens 2|3)*P(3)
v We can calculate the likelihoods assuming you’ve chosen door 1:
The probability that MH opens door 2 if the prize is behind door 2:
P(MH opens 2|2) = 0 (game would be over, no fun!)
The probability that MH opens door 2 if the prize is behind door 1:
P(MH opens 2|1) = 1/2 (could have opened 2 or 1)
The probability that MH opens door 2 if the prize is behind door 1:
P(MH opens 2|3) = 1 (no other choice door 1 is already picked, 3
gives game away)
P(MH opens 2)=1/3*1/2 +1/3*0+1/3*1=1/6+1/3=1/2
You Should
P(A|MH opens B) = (1/3*1/2)/(1/2)=1/3
SWITCH !!!!
P(C|MH opens B) = (1/3*1)/(1/2)=2/3
79