Chapter 2

Probability Concepts and Applications

Learning Objectives:

After the completion of the chapter, the students will be able to:
 describe the basic functions of probability analysis;
 describe the statistically dependent and independent events;
 use Bayes’ theorem to establish posterior probabilities;
 describe and provide examples discrete and random variables;
 differentiate discrete and continuous probability distributions; and
 calculate expected values and variances and use the normal table.

What is Probability?

Probability is a numerical statement about the likelihood that an event will occur.
There are several rules, definitions, and concepts associated with probability that are
very important in understanding the use of probability in decision making.

Two Basic Rules of Probability

There are two basic rules regarding the mathematics of probability:

1. The probability, P, of any event or state of nature occurring is greater than or

equal to 0 and less than or equal to 1. That is,
0 ≤ P (event) ≤ 1 (2 -1)

A probability of 0 indicates that an event is never expected to occur. A probability of 1

means that an event is always expected to occur.

2. The sum of the simple probabilities for all possible outcomes of an activity
must equal 1. Regardless of how probabilities are determined, they must
adhere to these two rules.

Types of Probability

There are two different ways to determine probability: the objective approach and the
subjective approach.

The relative frequency approach is an objective probability assessment. The

probability assigned to an event is the relative frequency of that occurrence. In
general,

𝑃 (𝑒𝑣𝑒𝑛𝑡) = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑐𝑐𝑢𝑟𝑒𝑛𝑐𝑒𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑒𝑣𝑒𝑛𝑡

𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑟𝑖𝑎𝑙𝑠 𝑜𝑟 𝑜𝑢𝑡𝑐𝑜𝑚𝑒𝑠
Example: The demand for white latex paint at Paintados Merchandising has always
been 0, 1, 2, 3, or 4 gallons per day. (There are no other possible outcomes and
when one occurs, no other can.) Over the past 200 working days, the owner notes
the frequencies of demand as shown in Table 2.1. If this past distribution is a good
indicator of future sales, we can find the probability of each possible outcome
occurring in the future by converting the data into percentages.

Table 2.1 Relative Frequency Approach to Probability Paint Sales

Demand Quantity Number of Days Probability
0 40 0.20 (= 40/200)
1 80 0.40 (=80/200)
2 50 0.25 (=50/200)
3 20 0.10 (=20/200)
4 10 0.05 (=10/200)
Total 200 1.00 (=200/200)

Thus, the probability that sales are 2 gallons of paint on any given day is P(2 gallons)
= 0.25 = 25%. The probability of any level of sales must be greater than or equal to 0
and less than or equal to 1. Since 0, 1, 2, 3, and 4 gallons exhaust all possible events
or outcomes, the sum of their probability values must equal 1.

Objective probability can also be set using what is called the classical or logical
method. Without performing a series of trials, we can often logically determine what
the probabilities of various events should be. For example, the probability of tossing a
fair coin once and getting a head is

1 Number of ways of getting a head

𝑃(ℎ𝑒𝑎𝑑) =
2 Number of possible outcomes (head or tail)

Similarly, the probability of drawing a spade out of a deck of 52 playing cards can be
logically set as
13
𝑃(ℎ𝑒𝑎𝑑) =
52 Number of ways of getting a
head Number of possible
outcomes
= 25%
=¼ = 0.25
When logic and past history are not available or appropriate, probability values can
be assessed subjectively. The accuracy of subjective probabilities depends on the
experience and judgment of the person making the estimates. A number of
probability values cannot be determined unless the subjective approach is used.
What is the probability that the price of gasoline will be more than PhP4 in the next
few years? What is the probability that our economy will be in a severe depression in
2022? What is the probability that you will be president of a major corporation within
20 years?

There are several methods for making subjective probability assessments. Opinion
polls can be used to help in determining subjective probabilities for possible election
returns and potential political candidates. In some cases, experience and judgment
must be used in making subjective assessments of probability values. A production
manager, for example, might believe that the probability of manufacturing a new
product without a single defect is 0.85.

Mutually Exclusive and Collectively Exhaustive Events

Events are said to be mutually exclusive if only one of the events can occur on any
one trial. They are called collectively exhaustive if the list of outcomes includes
every possible outcome. Many common experiences involve events that have both of
these properties.

In tossing a coin, the possible outcomes are a head or a tail. Since both of them
cannot occur on any one toss, the outcomes head and tail are mutually exclusive.
Since obtaining a head and obtaining a tail represent every possible outcome, they
are also collectively exhaustive.

AB

Figure 2.1 Venn Diagram for Events That Are Mutually Exclusive

Figure 2.1 provides a Venn diagram representation of mutually exclusive events. Let
A be the event that a head is tossed, and let B be the event that a tail is tossed. The
circles representing these events do not overlap, so the events are mutually
exclusive.

The following situation provides an example of events that are not mutually exclusive.
You are asked to draw one card from a standard deck of 52 playing cards. The
following events are defined:

A = event that a 7 is drawn

 B = event that a heart is drawn

The probabilities can be assigned to these using the relative frequency approach.
There are four 7s in the deck and thirteen hearts in the deck. Thus, we have

P(a 7 is drawn) = P(A) = 4/52

P(a heart is drawn) = P(B) = 13/52

These events are not mutually exclusive as the 7 of hearts is common to both event
A and event B. Figure 2.2 provides a Venn diagram representing this situation.
Notice that the two circles intersect, and this intersection is whatever is in common to
both. In this example, the intersection would be the 7 of hearts.

AB

Figure 2.2 Venn Diagram for Events That Are Not Mutually

Exclusive Unions and Intersections of Events

The intersection of two events is the set of all outcomes that are common to both
events. The word and is commonly associated with the intersection, as is the symbol
Ո. There are several notations for the intersection of two events:

Intersection of an event A and event B = A and B

=AՈB

=AՈB

The notation for the probability would be

P (Intersection of event A and event B) = P(A and B)

= P (A Ո B)

= P (A Ո B)

The probability of the intersection is sometimes called a joint probability, which

implies that both events are occurring at the same time or jointly.

The union of two events is the set of all outcomes that are contained in either of
these two events. Thus, any outcome that is in event A is in the union of the two
events, and any outcome that is in event B is also in the union of the two events. The
word or is commonly associated with the union, as is the symbol Ս. Typical notation
for the union of two events would be
 Union of event A and event B = (A or B)

The notation for the probability of the union of events would be

P(Union of event A and event B) = P(A or B)

= P(A Ս B)

In the previous example, the intersection of event A and event B would be

(A and B) = the 7 of hearts is drawn

The notation for the probability would be

P(A and B) = P(7 of hearts is drawn) = 1/52

Also, the union of event A and event B would be

(A or B) = (either a 7 is drawn or a heart is drawn)

and the probability would be

P(A or B) = P(any 7 or any heart is drawn) = 16/52

To see why P(A or B) = 16/52 and not 17/52 (which is P(A) + P(B)), count all of the
cards that are in the union, and you will see there are 16.

Probability Rules for Unions, Intersections, and Conditional Probabilities

The general rule for the probability of the union of two events (sometimes called the
additive rule) is the following:

𝑃(𝐴 or 𝐵) = 𝑃(𝐴) + 𝑃(𝐵) – 𝑃(𝐴 and 𝐵) (2-2)

To illustrate this with the example we have been using, to find the probability of the
union of the two events (a 7 or a heart is drawn), we have

P(A or B) = P(A) + P(B) – P(A and B)

= 4/52 + 13/52 – 1/52

= 16/52

One of the most important probability concepts in decision making is the concept of a
conditional probability. A conditional probability is the probability of an event
occurring given that another event has already happened. The probability of event A
given that event B has occurred is written as P(A | B). When businesses make
decisions, they often use market research of some type to help determine the
likelihood
of success. Given a good result from the market research, the probability of success
would increase.

The probability that A will occur given that event B has occurred can be found by
dividing the probability of the intersection of the two events (A and B) by the
probability of the event that has occurred (B):

𝑃(𝐴𝐵) (2-3)
𝑃 (𝐴|𝐵) =
𝑃(𝐵)

From this, the formula for the intersection of two events can be easily derived and
written as

𝑃(𝐴𝐵) = 𝑃(𝐴 |¬ 𝐵)𝑃(𝐵) (2-4)

In the card example, what is the probability that a 7 is drawn (event A) given that we
know that the card drawn is a heart (event B)? With what we already know, and
given the formula for conditional probability, we have

𝑃(𝐴𝐵)
𝑃 (𝐴|𝐵) =
𝑃(𝐵)

1/52
=
13/5 = 1/13
2

With this card example, it might be possible to determine this probability without
using the formula. Given that a heart was drawn and there are 13 hearts with only
one of these being a 7, we can determine that the probability is 1/13. In business,
however, we sometimes do not have this complete information and the formula is
absolutely essential.

Two events are said to be independent if the occurrence of one has no impact on
the occurrence of the other. Otherwise, the events are dependent.

For example, suppose a card is drawn from a deck of cards and it is then returned to
the deck and a second drawing occurs. The probability of drawing a seven on the
second draw is 4/52 regardless of what was drawn on the first draw because the
deck is exactly the same as it was on the first draw. Now contrast to a similar
situation with two draws from a deck of cards, but the first card is not returned to the
deck. Now there are only 51 cards left in the deck, and there are either three or four
7s in the deck depending on what the first card drawn happens to be.
A more precise definition of statistical independence would be the following: Event A
and event B are independent if

𝑃 (𝐴 | 𝐵) = 𝑃 (𝐴)
Independence is a very important condition in probability as many calculations are
simplified. One of these is the formula for the intersection of two events. If A and
B are independent, then the probability of the intersection is

𝑃 (𝐴 | 𝐵) = 𝑃 (𝐴) 𝑃 (𝐵)

Suppose a fair coin is tossed twice. The events are defined as:

A = event that a head is the result of the first toss

B = event that a head is the result of the second toss

These events are independent because the probability of a head on the second toss
will be the same regardless of the result on the first toss. Because it is a fair coin,
we know there are two equally likely outcomes on each toss (head or tail), so

P (A) = 0.5

and

P (B) = 0.5

Because A and B are independent,

𝑃 (𝐴𝐵) = 𝑃(𝐴)𝑃(𝐵) = 0.5 (0.5) = 0.25

Thus, there is a 0.25 probability that two tosses of a coin will result in two heads.

If events are not independent, then finding probabilities may be a bit more difficult.
However, the results may be very valuable to a decision maker. A market research
study about opening a new store in a particular location may have a positive
outcome, and this would cause a revision of our probability assessment that the new
store would be successful.

Revising Probabilities with Bayes’ Theorem

Bayes’ theorem is used to incorporate additional information as it is made available

and help create revised or posterior probabilities from the original or prior
probabilities. This means that we can take new or recent data and then revise and
improve upon our old probability estimates for an event (see Figure 2.3). Let us
consider the following example.
A cup contains two dice identical in appearance. One, however, is fair (unbiased) and
the other is loaded (biased). The probability of rolling a 3 on the fair die is 1-6, or
0.166. The probability of tossing the same number on the loaded die is 0.60.

We have no idea which die is which, but select one by chance and toss it. The result
is a 3. Given this additional piece of information, can we find the (revised) probability
that the die rolled was fair? Can we determine the probability that it was the loaded
die that was rolled?

Prior Probabilities

Bayes’
Process
New Information

Figure 2.3 Using Bayes’ Process

The answer to these questions is yes, and we do so by using the formula for joint
probability under statistical dependence and Bayes’ theorem. First, we take stock of
the information and probabilities available. We know, for example, that since we
randomly selected the die to roll, the probability of it being fair or loaded is 0.50:

𝑃 (𝑓𝑎𝑖𝑟) = 0.50 𝑃(𝑙𝑜𝑎𝑑𝑒𝑑) = 0.50

We also know that

𝑃 (3|𝑓𝑎𝑖𝑟) = 0.166 𝑃(3|𝑙𝑜𝑎𝑑𝑒𝑑) = 0.60

Next, we compute joint probabilities P(3 and fair) and P(3 and loaded) using the
formula P(AB) = P(AB) * P(B):

𝑃 (3 𝑎𝑛𝑑 𝑓𝑎𝑖𝑟) = 𝑃(3|𝑓𝑎𝑖𝑟) 𝑥 𝑃 ∗ (𝑓𝑎𝑖𝑟)

= ( 0.166)(0.50) = 0.083

𝑃 (3 𝑎𝑛𝑑 𝑙𝑜𝑎𝑑𝑒𝑑) = 𝑃 (3|𝑙𝑜𝑎𝑑𝑒𝑑)𝑥 𝑃(𝑙𝑜𝑎𝑑𝑒𝑑)

= ( 0.60)(0.50) = 0.300

A 3 can occur in combination with the state “fair die” or in combination with the state
“loaded die.” The sum of their probabilities gives the unconditional or marginal
probability of a 3 on the toss, namely, P(3) = 0.083 + 0.300 = 0.383.

toss, namely, P132 = 0.083 + 0.300 = 0.383.

If a 3 does occur, and if we do not know which die it came from, the probability that the
die rolled was the fair one is

𝑃 (𝑓𝑎𝑖𝑟 𝑎𝑛𝑑 3) 0.083

𝑃 (𝑓𝑎𝑖𝑟|3) = = = 0.22
𝑃 (3) 0.383

The probability that the die rolled was loaded is

𝑃 (𝑙𝑜𝑎𝑑𝑒𝑑 𝑎𝑛𝑑 3) 0.300

𝑃 (𝑙𝑜𝑎𝑑𝑒𝑑|3) = = = 0.78
𝑃 (3) 0.383

These two conditional probabilities are called the revised or posterior probabilities for
the next roll of the die.

Before the die was rolled in the preceding example, the best we could say was that
there was a 50–50 chance that it was fair (0.50 probability) and a 50–50 chance that
it was loaded. After one roll of the die, however, we are able to revise our prior
probability estimates. The new posterior estimate is that there is a 0.78 probability
that the die rolled was loaded and only a 0.22 probability that it was not.

Using a table is often helpful in performing the calculations associated with Bayes’
theorem. Table 2.3 provides the general layout for this, and Table 2.4 provides this
specific example.

Table 2.3 Tabular Form of Bayes’ Calculations Given That Event B Has Occurred

State of P(B | State Prior Joint Posterior Probability

Nature of Nature) Probability Probability

A P(B | A) x P(A) = P(B and A) P(B and A) / P(B) = P (A | B)

A’ P (B | A’) x P(A’) = P(P and P(B and A’) / P(B) = P (A’ | B)

A’) P(B)

Table 2.4 Bayes’ Calculations Given That a 3 Is Rolled in Example 7

State of P(B | State Prior Joint Posterior Probability

Nature of Nature) Probability Probability

Fair Die 0.166 x 0.5 = 0.083) 0.083 / 0.383 = 0.22

Loaded 0.600 x 0.5 = 0.300 0.300 / 0.383 = 0.78

Die P(3) = 0.383
General Form of Bayes’ Theorem

Revised probabilities can also be computed in a more direct way using a general form
for Bayes’ theorem:

𝑃 (𝐵|𝐴) 𝑃(𝐴)
𝑃 (𝐴|𝐵) = 𝑃(𝐵|𝐴) 𝑃(𝐴)+𝑃 (𝐵|𝐴 ′ )𝑃(𝐴 ) (2-5)
𝘍

where

A’ = the complement of the event A;

For example: if A is the event “fair die,” then A’ is “loaded die”

We originally saw in Equation 2-3 the conditional probability of event A, given event B,
is

𝑃(𝐴𝐵)
𝑃 (𝐴|𝐵) =
𝑃(𝐵)

Although it may not be obvious to you at first glance, we used this basic equation to
compute the revised probabilities. For example, if we want the probability that the fair
die was rolled given the first toss was a 3, namely, P(fair die | 3 rolled), we can let

event “fair die” replace A in Equation 2-5

event “loaded die” replace A’ in Equation 2-5

event “3 rolled” replace B in Equation 2-5

We can then rewrite Equation 2-5 and solve as follows:

𝑃(𝑓𝑎𝑖𝑟 𝑑𝑖𝑒 |3 𝑟𝑜𝑙𝑙𝑒𝑑)

𝑃 (3 |𝑓𝑎𝑖𝑟)𝑃 (𝑓𝑎𝑖𝑟)
= 𝑃 (3|𝑓𝑎𝑖𝑟)𝑃 (𝑓𝑎𝑖𝑟)+𝑃 (3 |𝑙𝑜𝑎𝑑𝑒𝑑) 𝑃 (𝑙𝑜𝑎𝑑𝑒𝑑)

(0.166)(0.50)
=
(0.166)(0.50) + (0.60)(0.50)

0.083
= 0.383 = 0.22

Further Probability Revisions

Although one revision of prior probabilities can provide useful posterior probability
estimates, additional information can be gained from performing the experiment a
second time. If it is financially worthwhile, a decision maker may even decide to make
several more revisions.

Returning to the previous example, we now attempt to obtain further information

about the posterior probabilities as to whether the die just rolled is fair or loaded. To
do so, let us toss the die a second time. Again, we roll a 3. What are the further
revised probabilities?

To answer this question, we proceed as before, with only one exception. The
probabilities P(fair) = 0.50 and P(loaded) = 0.50 remain the same, but now we must
compute P(3,3 | fair) = (0.166) (0.166) = 0.027 and P(3, 3 | loaded) = (0.6)(0.6) =
0.36. With these joint probabilities of two 3s on successive rolls, given the two types
of dice, we may revise the probabilities:

P (3,3 and fair) = P (3,3 | fair) x P(fair)

= (0.027) (0.5) = 0.013

P (3,3 and loaded) = P (3,3 | loaded) x P (loaded)

= (0.36) (0.5) = 0.18

Thus, the probability of rolling two 3s, a marginal probability, is 0.013 + 0.18 = 0.193,
the sum of the two joint probabilities:

𝑃 (3.3 𝑎𝑛𝑑 𝑓𝑎𝑖𝑟)

𝑃 (𝑓𝑎𝑖𝑟 | 3,3) = 𝑃 (3,3)

0.013
= 0.193 = 0.067

𝑃 (3.3 𝑎𝑛𝑑 𝑙𝑜𝑎𝑑𝑒𝑑)

𝑃 (𝑙𝑜𝑎𝑑𝑒𝑑 | 3,3) = 𝑃 (3,3)

0.18
= 0.193 = 0.933

What has this second roll accomplished? Before we rolled the die the first time, we
knew only that there was a 0.50 probability that it was either fair or loaded. When the
first die was rolled in the previous example, we were able to revise these
probabilities:

probability of the die is fair = 0.22

probability of the die is loaded =

0.78

Now, after the second roll in this example, our refined revisions tell us that

probability the die is fair = 0.067

 probability the die is loaded = 0.933

This type of information can be extremely valuable in business decision making.

Random Variables

We have just discussed various ways of assigning probability values to the outcomes
of an experiment. Let us now use this probability information to compute the
expected outcome, variance, and standard deviation of the experiment. This can help
select the best decision among a number of alternatives.

A random variable assigns a real number to every possible outcome or event in an

experiment. It is normally represented by a letter such as X or Y. When the outcome
itself is numerical or quantitative, the outcome numbers can be the random variable.
For example, consider television sales at an appliance store. The number of
refrigerators sold during a given day can be the random variable. Using X to
represent this random variable, we can express this relationship as follows:

X = number of televisions sold during the day

In general, whenever the experiment has quantifiable outcomes, it is beneficial to

define these quantitative outcomes as the random variable. Examples are given in
Table 2.5.

Table 2.5 Examples of Random Variables

Experiment Outcome Random Variables Range of Random
Variables

Stock 50 Santa Number of Santa Hats X = number of 0, 1, 2, ….., 50

Hats sold Santa hats sold

Inspect 600 items Number of acceptable Y = number of 0, 1, 2, ….., 600

items acceptable items

Send out 5,000 Number of people’s Z = number of 0, 1, 2, ….., 5,000

promotional letters response to letters people’s response
to letters

Build an apartment Percentage of building R = percentage of 0 ≤ R ≤ 100

building completion in 4 building completion
months in 4 months

Test the lifetime of a Length of time the bulb S = time the bulb 0 ≤ S ≤ 80,000
lightbulb (minutes) lasts up to 80,000 burns
mins.
When the outcome itself is not numerical or quantitative, it is necessary to define a
random variable that associates each outcome with a unique real number. Several
examples are given in Table 2.6.

There are two types of random variables: discrete random variables and
continuous random variables. Developing probability distributions and making
computations based on these distributions depend on the type of random variable.

A random variable is a discrete random variable if it can assume only a finite or

limited set of values. Looking at Table 2.5, we can see that the variables associated
with stocking 50 Christmas trees, inspecting 600 items, and sending out 5,000 letters
are all examples of discrete random variables. Each of these random variables can
assume only a finite or limited set of values. The number of Christmas trees sold, for
example, can only be integer numbers from 0 to 50. There are 51 values that the
random variable X can assume in this example.

A continuous random variable is a random variable that has an infinite or an

unlimited set of values. Are there any examples of continuous random variables in
Table 2.5 or 2.6? Looking at Table 2.5, we can see that testing the lifetime of a
lightbulb is an experiment whose results can be described with a continuous random
variable. In this case, the random variable, S, is the time the bulb burns. It can last for
3,206 minutes, 6,500.7 minutes, 251.726 minutes, or any other value between 0 and
80,000 minutes. In most cases, the range of a continuous random variable is stated
as lower value ≤ S ≤ upper value, such as 0 ≤ S ≤ 80,000. The random variable R in
Table 2.5 is also continuous.

Table 2.6 Random Variables for Outcomes That Are Not Numbers

Experiment Outcome Range of Random Random

Variables Variables
Students respond Strongly agree (SA) 5 is SA 1, 2, 3, 4, 5
to a questionnaire Agree (A) 4 is A
Neutral (N) 3 is N
X= 2 is D
Disagree (D)
Strongly Disagree 1 is SD
(SD)
One machine is Defective 0 if defective 0, 1
Y= 1 if not defective
inspected Not defective
Consumers Good 3 if good 1, 2, 3
respond to how Average Z= 2 if average
they like a product Poor 1 if poor
Probability Distributions

Probability Distribution of a Discrete Random Variable

When we have a discrete random variable, there is a probability value assigned to

each event. These values must be between 0 and 1, and they must sum to 1. Let’s
look at an example.

In the ITEC 95 class, there are 100 students who have just completed a math quiz
given on the first day of class. The quiz consists of five very difficult set of problems.
The grade on the quiz is the number of correct answers, so the grades theoretically
could range from 0 to 5. However, no one in this class received a score of 0, so the
grades ranged from 1 to 5. The random variable X is defined to be the grade on this
quiz, and the grades are summarized in Table 2.7. This discrete probability
distribution was developed using the relative frequency approach.

The distribution follows the three rules required of all probability distributions: (1) the
events are mutually exclusive and collectively exhaustive, (2) the individual
probability values are between 0 and 1 inclusive, and (3) the total of the probability
values is 1.

Table 2.7 Probability Distribution for Quiz Scores

Random Variable Number Probability
X (Score)
5 10 0.1 = 10/100

4 20 0.2 = 20/100

3 30 0.3 = 30/100

2 30 0.3 = 30/100

1 10 0.1 = 10/100
Total 100 1.0 = 100/100

The distribution follows the three rules required of all probability distributions: (1) the
events are mutually exclusive and collectively exhaustive, (2) the individual
probability values are between 0 and 1 inclusive, and (3) the total of the probability
values is 1.

Although listing the probability distribution as we did in Table 2.7 is adequate, it can
be difficult to get an idea about characteristics of the distribution. To overcome this
problem, the probability values are often presented in graph form. The graph of the
distribution in Table 2.7 is shown in Figure 2.4.
The graph of this probability distribution gives us a picture of its shape. It helps us
identify the central tendency of the distribution, called the mean or expected value,
and the amount of variability or spread of the distribution, called the variance.

Figure 2.4 Probability Distribution for ITEC 95

Class Expected Value of a Discrete Probability Distribution

Once we have established a probability distribution, the first characteristic that is

usually of interest is the central tendency of the distribution. The expected value, a
measure of central tendency, is computed as the weighted average of the values of
the random variable:

𝐸 (𝑋) = ∑ 𝑋𝑖𝑃(𝑋𝑖)
𝑖=1

= 𝑋1𝑃(𝑋1) + 𝑋2𝑃(𝑋2) + ⋯ + 𝑋𝑛𝑃(𝑋𝑛) (2-6)

where

𝑋1 = 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒′𝑠 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒𝑠

𝑃(𝑋1) = 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑒𝑎𝑐ℎ 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒

𝑛
∑𝑖=1 = (𝑠𝑢𝑚𝑚𝑎𝑡𝑖𝑜𝑛 𝑠𝑖𝑔𝑛 𝑖𝑛𝑑𝑖𝑐𝑎𝑡𝑖𝑛𝑔 𝑤𝑒 𝑎𝑟𝑒 𝑎𝑑𝑑𝑖𝑛𝑔 𝑎𝑙𝑙 𝑛 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒𝑠)

𝐸 (𝑋) = 𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 𝑜𝑟 𝑚𝑒𝑎𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒

The expected value or mean of any discrete probability distribution can be computed
by multiplying each possible value of the random variable, Xi, times the probability,
P(Xi), that outcome will occur and summing the results, ∑. Here is how the expected
value can be computed for the quiz scores:
 𝑛

𝐸 (𝑋) = ∑ 𝑋𝑖𝑃(𝑋𝑖)
𝑖=1

=X1P(X1) + X2P(X2) + X3P(X3) + X4P(X4) + X5P(X5)

= (5) (0.1) + (4) (0.2) + (3) (0.3) + (2) (0.3) + (1) (0.1)

= 2.9

The expected value of 2.9 is the mean score on the quiz.

Variance of a Discrete Probability Distribution

In addition to the central tendency of a probability distribution, most people are

interested in the variability or the spread of the distribution. If the variability is low, it is
much more likely that the outcome of an experiment will be close to the average or
expected value. On the other hand, if the variability of the distribution is high, which
means that the probability is spread out over the various random variable values,
there is less chance that the outcome of an experiment will be close to the expected
value.

The variance of a probability distribution is a number that reveals the overall spread
or dispersion of the distribution. For a discrete probability distribution, it can be
computed using the following equation:

𝜎 = 𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = ∑[𝑋1 − 𝐸 (𝑋)]2 𝑃(𝑋𝑖 )

2
(2-7)
𝑖=1

where:

𝑋1 = 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒′𝑠 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒𝑠

𝐸 (𝑋) = 𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒

[𝑋𝑖 − 𝐸 (𝑋)] = 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑒𝑧𝑐ℎ 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒

𝑃(𝑋𝑖 ) = 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑒𝑎𝑐ℎ 𝑝𝑜𝑠𝑠𝑖𝑏𝑙𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒

To compute the variance, each value of the random variable is subtracted from the
expected value, squared, and multiplied times the probability of occurrence of that
value. The results are then summed to obtain the variance. Here is how this
procedure is done for ITEC 95’s quiz scores:
 5

𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = ∑[𝑋1 − 𝐸 (𝑋)]2 𝑃(𝑋𝑖 )

𝑖=1

𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = (5 − 2.9)2 (0.1) + (4 − 2.9)2(0.2) + (3 − 2.9)2(0.3) + (2 − 2.9)2(0.3) +

(1 − 2.9)2(0.1)

= (2.1)2(0.1) + (1.1)2(0.2) + (0.1)2(0.3) + (−.09)2(0.3) + (−1.9)2(0.1)

= 0.441 + 0.242 + 0.003 + 0.243 + 0.361

= 𝟏. 𝟐𝟗

A related measure of dispersion or spread is the standard deviation. This quantity is

also used in many computations involved with probability distributions. The standard
deviation is just the square root of the variance:

𝜎 = √𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = √𝜎2 (2-8)

where

√ = 𝑠𝑞𝑢𝑎𝑟𝑒 𝑟𝑜𝑜𝑡

𝜎 = 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛

The standard deviation for the random variable X in the example is

𝜎 = √𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒

𝜎 = √1.29 = 1.14

These calculations are easily performed in Excel. Program 2.1A provides the output
for this example. Program 2.1B shows the inputs and formulas in Excel for
calculating the mean, variance, and standard deviation in this example.

Program 2.1A Excel 2016 Output for the ITEC 95 quiz sample
Program 2.1B Formulas in an Excel Spreadsheet for the ITEC 95 Quiz Sample

Probability Distribution of a Continuous Random Variable

There are many examples of continuous random variables. The time it takes to finish
a project, the number of ounces in a barrel of butter, the high temperature during a
given day, the exact length of a given type of lumber, and the weight of a railroad car
of coal are all examples of continuous random variables. Since random variables can
take on an infinite number of values, the fundamental probability rules for continuous
random variables must be modified.

As with discrete probability distributions, the sum of the probability values must equal
1. Because there are an infinite number of values of the random variables, however,
the probability of each value of the random variable must be 0. If the probability
values for the random variable values were greater than 0, the sum would be
infinitely large.

With a continuous probability distribution, there is a continuous mathematical function

that describes the probability distribution. This function is called the probability
density function or simply the probability function. It is usually represented by f(X).
When working with continuous probability distributions, the probability function can
be graphed, and the area underneath the curve represents probability. Thus, to find
any probability, we simply find the area under the curve associated with the range of
interest.

We now look at the sketch of a sample density function in Figure 2.5. This curve
represents the probability density function for the weight of a particular machined
part. The weight could vary from 5.06 to 5.30 grams, with weights around 5.18 grams
being the most likely. The shaded area represents the probability the weight is
between 5.22 and 5.26 grams.
 Figure 2.5 Graph of Sample Density Function

We now look at the sketch of a sample density function in Figure 2.5. This curve
represents the probability density function for the weight of a particular machined
part. The weight could vary from 5.06 to 5.30 grams, with weights around 5.18 grams
being the most likely. The shaded area represents the probability the weight is
between 5.22 and 5.26 grams.

If we wanted to know the probability of a part weighing exactly 5.1300000 grams, for
example, we would have to compute the area of a line of width 0. Of course, this
would be 0. This result may seem strange, but if we insist on enough decimal places
of accuracy, we are bound to find that the weight differs from 5.1300000 grams
exactly, be the difference ever so slight.

This is important because it means that, for any continuous distribution, the
probability does not change if a single point is added to the range of values that is
being considered. In Figure 2.5, this means the following probabilities are all exactly
the same:

P (5.22 < X < 5.26) = P(5.22 < X ≤ 5.26) = P(5.22 ≤ X < 5.26)

= P(5.22 ≤ X ≤ 5.26

The inclusion or exclusion of either or both endpoints (5.22 or 5.26) has no impact on
the probability.

The Binomial Distribution

Many business experiments can be characterized by the Bernoulli process. The

probability of obtaining specific outcomes in a Bernoulli process is described by the
binomial probability distribution. In order to be a Bernoulli process, an experiment
must have the following characteristics:

1. Each trial in a Bernoulli process has only two possible outcomes. These are
typically called a success and a failure, although examples might be yes or
no, heads or tails, pass or fail, defective or good, and so on.
2. The probability stays the same from one trial to the next.
3. The trials are statistically independent.
4. The number of trials is a positive integer.

A common example of this process is tossing a coin.

The binomial distribution is used to find the probability of a specific number of

successes out of n trials of a Bernoulli process. To find this probability, it is necessary
to know the following:

n = the number of trials

p= the probability of a success of any single trial

We let

r = the number of successes

q = 1- p = the probability of a failure

The binomial formula is

𝑛!
𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑟 𝑠𝑢𝑐𝑐𝑒𝑠𝑠𝑒𝑠𝑠 𝑖𝑛 𝑛 𝑡𝑟𝑖𝑎𝑙𝑠 = 𝑟𝑞𝑛−𝑟 (2-9)
𝑟! (𝑛 − 𝑟)! 𝑝

The symbol ! means factorial and n! = n(n-1)(n-2) … (1)

For example: 4! = (4) (3) (2) (1) = 24

Also, 1! = 1 and 0! = 1 by definition

Solving Problems with the Binomial Formula

A common example of a binomial distribution is the tossing of a coin and counting

the number of heads. For example, if we wished to find the probability of 4 heads in 5
tosses of a coin, we would have

𝑛 = 5, 𝑟 = 4, 𝑝 = 0.5 𝑎𝑛𝑑 𝑞 = 1 − 0.5 = 0.5

Thus
𝑛 = 5, 𝑟 = 4, 𝑝 = 0.5 𝑎𝑛𝑑 𝑞 = 1 − 0.5 = 0.5

Thus,
5!
𝑃(4 𝑠𝑢𝑐𝑐𝑒𝑠𝑠𝑒𝑠 𝑖𝑛 5 𝑡𝑟𝑖𝑎𝑙𝑠 = 0.54 0.55−4
4! (5 −
4)!

5 (4)(3)(2)(1)
= (4)(3)(2)(1)(1!) (0.0625)(0.5) = 𝟎. 𝟏𝟓𝟔𝟐𝟓

Thus, the probability of 4 heads in 5 tosses of a coin is 0.15625, or about 16%.

Using Equation 2-9, it is also possible to find the entire probability distribution (all the
possible values for r and the corresponding probabilities) for a binomial experiment.
The probability distribution for the number of heads in 5 tosses of a fair coin is shown
in Table 2.8 and then graphed in Figure 2.6.

Table 2.8 Binomial Probability Distribution for n = 5 and p = 0.50

Number of Heads 𝟓!
𝑷𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕𝒚 = (𝟎. 𝟓)𝒓 (𝟎. 𝟓)𝟓−𝒓
(r) 𝒓! (𝟓 − 𝒓)!

0 5!
0.03125 = (0.5)0 (0.5)5−0
0! (5 − 0)!

1 5!
0.15625 = (0.5)1 (0.5)5−1
1! (5 − 1)!

2 5!
0.31250 = (0.5)2 (0.5)5−2
2! (5 − 2)!

3 5!
0.03125 = (0.5)3 (0.5)5−3
3! (5 − 3)!

4 5!
0.15625 = (0.5)4 (0.5)5−4
4! (5 − 4)!

5 5!
0.03125 = (0.5)5 (0.5)5−5
5! (5 − 5)!
 Figure 2.6 Binomial Probability Distribution for n = 5 and p = 0.50

Solving Problems with Binomial Tables

GGP Electronics is experimenting with the manufacture of a new type of transistor

that is very difficult to mass-produce at an acceptable quality level. Every hour a
supervisor takes a random sample of 5 transistors produced on the assembly line.
The probability that any one transistor is defective is considered to be 0.15. MSA
wants to know the probability of finding 3, 4, or 5 defectives if the true percentage
defective is 15%.

For this problem, n = 5, p = 0.15, and r = 3, 4, or 5. Although we could use the

formula for each of these values, it is easier to use binomial tables for this. The range
of values for n, r and p are shown in Table 2.9. To find these probabilities, we look
through the n = 5 section and find the p = 0.15 column. In the row where r = 3, we
see 0.0244. Thus, P (r=3) = 0.0244. Similarly, P (r=4) = 0.0022, and P(r=5) = 0.0001.
By adding these three probabilities, we have the probability that the number of
defects is 3 more:

P (3 more defects) = P (3) + P(4) + P(5)

= 0.0244 + 0.0022 + 0.0001 = 0.0267

The expected value (or mean) and the variance of a binomial random variable may
be easily found. These are

Expected value (mean) = np (2-10)

Variance = np (1-p) (2-11)

The expected value and variance for the GGP Electronics example are computed as
follows:
Expected value = np = 5 (0.15) = 0.75
Variance = np (1-p) = 5 (0.15) (0.85) = 0.6375
Table 2.9 A Sample Table for the Binomial Distribution

Program 2.2A Excel Output for the Binomial Example

Program 2.2BFunction in an Excel 2016 Spreadsheet for Binomial Probabilities

The Normal Distribution

One of the most popular and useful continuous probability distributions is the normal
distribution. The probability density function of this distribution is given by the rather
complex formula

(2-12)

The normal distribution is specified completely when values for the mean, m, and the
standard deviation, s, are known. Figure 2.7 shows several different normal
distributions with the same standard deviation and different means. As shown,
differing values of µ will shift the average or center of the normal distribution. The
overall shape of the distribution remains the same. On the other hand, when the
standard deviation is varied, the normal curve either flattens out or becomes steeper.
This is shown in Figure 2.8.

Figure 2.7 Normal Distribution with Different Values for µ

Figure 2.8 Normal Distribution with Different Values for 𝝈

As the standard deviation, 𝝈, becomes smaller, the normal distribution becomes
steeper. When the standard deviation becomes larger, the normal distribution has a
tendency to flatten out or become broader.

Area Under the Normal Curve

Because the normal distribution is symmetrical, its midpoint (and highest point) is at
the mean. Values on the X axis are then measured in terms of how many standard
deviations they lie from the mean. As you may recall from our earlier discussion of
probability distributions, the area under the curve (in a continuous distribution)
describes the probability that a random variable has a value in a specified interval.
When dealing with the uniform distribution, it is easy to compute the area between
any points a and b.

Using the Standard Normal Table

When finding probabilities for the normal distribution, it is best to draw the normal
curve and shade the area corresponding to the probability being sought. The normal
distribution table can then be used to find probabilities by following two steps.

STEP 1. Convert the normal distribution to what we call a standard normal

distribution. A standard normal distribution has a mean of 0 and a standard deviation
of 1. All normal tables are set up to handle random variables with µ = 0 and 𝝈 = 1.
Without a standard normal distribution, a different table would be needed for each
pair of m and s values. We call the new standard random variable Z. The value for Z
for any normal distribution is computed from this equation:

𝑋−
𝜇 (2-13)
𝑍=
𝜎

where

𝑋 = value of the random variable we want to measure

𝜇 = mean of the distribution

𝜎 = standard deviation of the distribution

𝑍 = number of standard deviations from X to the mean, 𝜇

For example, if 𝜇 = 100, 𝜎 = 15, and we are interested in finding the probability that the
random variable X (IQ) is less than 130, we want P(X < 130):

𝑋− 𝜇 130 − 100 30
𝑍= = = = 𝟐 𝐬𝐭𝐚𝐧𝐝𝐚𝐫𝐝 𝐝𝐞𝐯𝐢𝐚𝐭𝐢𝐨𝐧𝐬
𝜎 15 15
This means that the point X is 2.0 standard deviations to the right of the mean. This
is shown in Figure 2.9.

Figure 2.9 Normal Distribution Showing the Relationship Between Z Values

and X Values

STEP 2. Look up the probability from a table of normal curve areas. The table of
areas for the standard normal distribution is set up to provide the area under the
curve to the left of any specified value of Z. We will use Table 2.10. The column on
the left lists values of Z, with the second decimal place of Z appearing in the top row.
For example, for a value of Z = 2.00 as just computed, find 2.0 in the left-hand
column and 0.00 in the top row. In the body of the table, we find that the area sought
is 0.97725, or 97.7%. Thus,

𝑃 (𝑋 < 130) = 𝑃(𝑍 < 2.00) = 97.7%

This suggests that if the mean IQ score is 100, with a standard deviation of 15 points,
the probability that a randomly selected person’s IQ is less than 130 is 97.7%. This is
also the probability that the IQ is less than or equal to 130. To find the probability that
the IQ is greater than 130, we simply note that this is the complement of the previous
event and the total area under the curve (the total probability) is 1. Thus,

𝑃(𝑋 > 130) = 1 − 𝑃(𝑋 ≤ 130) = 1 − 𝑃(𝑍 ≤ 2) = 1 − 0.97725 = 0.02275

While Table 2.10 does not give negative Z values, the symmetry of the normal
distribution can be used to find probabilities associated with negative Z values. For
example, P(Z < -2) = P(Z > 2).
Table 2.10 Standardized Normal Distribution Function

Golden Builders Construction Company Example

Golden Builders Construction Company builds primarily three- and four-unit

apartment buildings (called triplexes and quadraplexes) for investors, and it is
believed that the total construction time in days follows a normal distribution. The
mean time to construct a triplex is 100 days, and the standard deviation is 20 days.
Recently, the president of Golden Builders Construction signed a contract to
complete a triplex in 125 days. Failure to complete the triplex in 125 days would
result in severe penalty fees. What is
the probability that Golden Builders Construction will not be in violation of its
construction contract? The normal distribution for the construction of triplexes is
shown in Figure 2.10.

To compute this probability, we need to find the shaded area under the curve. We
begin by computing Z for this problem:

𝑋− 𝜇
𝑍=
𝜎

125 − 100
𝑍=
20

25
𝑍= = 𝟏. 𝟐𝟓
20

Looking in Table 2.10 for a Z value of 1.25, we find an area under the curve of
0.89435. (We do this by looking up 1.2 in the left-hand column of the table and then
moving to the 0.05 column to find the value for Z = 1.25.) Therefore, the probability of
not violating the contract is 0.89435, or about an 89% chance.

Figure 2.10 Normal Distributtion for Golden Builders Construction

Now let us look at the Haynes problem from another perspective. If the firm finishes
this triplex in 75 days or less, it will be awarded a bonus payment of PhP5,000. What
is the probability that Golden Builders will receive the bonus?

Figure 2.11 illustrates the probability we are looking for in the shaded area. The first
step is again to compute the Z value:

𝑋− 𝜇
𝑍=
𝜎

75 − 100
𝑍=
20

−25
𝑍= = −𝟏. 𝟐𝟓
20
 Figure 2.11 Probability that Golden Builders Will Receive the Bonus by
Finishing in 75 Days or Less

This Z value indicates that 75 days is -1.25 standard deviations to the left of the
mean. But the standard normal table is structured to handle only positive Z values.
To solve this problem, we observe that the curve is symmetric. The probability that
Haynes will finish in 75 days or less is equivalent to the probability that it will finish in
more than 125 days. A moment ago (in Figure 2.10) we found the probability that
Haynes will finish in less than 125 days. That value is 0.89435. So the probability it
takes more than 125 days is

𝑃 (𝑋 > 125 = 1.0 − 𝑃(𝑋 ≤ 125)

= 1.0 − 0.89435 = 𝟎. 𝟏𝟎𝟓𝟔𝟓

Thus, the probability of completing the triplex in 75 days or less is 0.10565, or about
11%.

One final example: What is the probability that the triplex will take between 110 and
125 days? We see in Figure 2.12 that

𝑃 (110 < 𝑋 < 125 = 𝑃(𝑋 ≤ 125) − 𝑃 (𝑋 < 100)

That is, the shaded area in the graph can be computed by finding the probability of
completing the building in 125 days or less minus the probability of completing it in
110 days or less.

Figure 2.12 Probability that Golden Builders Will Complete in 110 to 125 Days
 Program 2.3A Excel 2016 Output for the Normal Distribution Example

Program 2.3B Function in an Excel 2016 Spreadsheet for the Normal

Distribution Example

Recall that P(X ≤125 days2 is equal to 0.89435. To find P(X < 110 days), we follow
the two steps developed earlier:
𝑋− 𝜇 10
1. 𝑍= = 110−100
= = 0.5 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛𝑠
𝜎 20 20

2. From Table 2.10, the area for Z = 0.50 is 0.69146. So the probability the triplex
can be completed in less than 110 days is 0.69146. Finally,
𝑃 (110 < 𝑋 < 125 = 0.89435 − 0.69146 = 0.20289
The probability that it will take between 110 and 125 days is about 20%. Programs
2.3A and 2.3B show how Excel can be used for this.

The Empirical Rule

While the probability tables for the normal distribution can provide precise
probabilities, many situations require less precision. The empirical rule was derived
from the normal distribution and is an easy way to remember some basic information
about normal distributions. The empirical rule states that for a normal distribution

approximately 68% of the values will be within ±1 standard deviation of the mean
approximately 95% of the values will be within ±2 standard deviations of the mean

almost all (about 99.7%) of the values will be within ±3 standard deviations of the mean

Figure 2.13 illustrates the empirical rule. The area from point a to point b in the first
drawing represents the probability, approximately 68%, that the random variable will
be within ±1 standard deviation of the mean. The middle drawing illustrates the
probability, approximately 95%, that the random variable will be within ±2 standard
deviations of the mean. The last drawing illustrates the probability, about 99.7%, that
the random variable will be within ±3 standard deviations of the mean.

Figure 2.13 The Empirical Rule

Exponential Distribution

The exponential distribution, also called the negative exponential distribution, is

used in dealing with queuing problems. The exponential distribution often describes
the time required to service a customer. The exponential distribution is a continuous
distribution. Its probability function is given by

𝑓 (𝑋) = 𝜇𝑒−𝜇𝓍 (2-14)

where

𝑋 = 𝑟𝑎𝑛𝑑𝑜𝑚 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒 (𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑡𝑖𝑚𝑒𝑠

𝜇 = 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑢𝑛𝑖𝑡𝑠 𝑡ℎ𝑒 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑓𝑎𝑐𝑖𝑙𝑖𝑡𝑦 𝑐𝑎𝑛 ℎ𝑎𝑛𝑑𝑙𝑒 𝑖𝑛 𝑎 𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 𝑝𝑒𝑟𝑖𝑜𝑑 𝑜𝑓 𝑡𝑖𝑚𝑒

𝑒 = 2.718 (𝑡ℎ𝑒 𝑏𝑎𝑠𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑙𝑜𝑔𝑎𝑟𝑖𝑡ℎ𝑚)

The general shape of the exponential distribution is shown in Figure 2.16. Its expected
value and variance can be shown to be

1
𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 = = 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑡𝑖𝑚𝑒 (2-15)
𝜇

1
𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = (2-16)
2
𝜇

As with any other continuous distribution, probabilities are found by determining the
area under the curve. For the normal distribution, we found the area by using a table
of probabilities. For the exponential distribution, the probabilities can be found using
the exponent key on a calculator with the formula below. The probability that an
exponentially distributed time, X, required to serve a customer is less than or equal to
time t is given by the formula

𝑃 (𝑋 ≤ 𝑡) = 1 − 𝑒−𝜇𝑡 (2-17)

The time period used in describing 𝜇 determines the units for the time t. For example,
if 𝜇 is the average number served per hour, the time t must be given in hours. If 𝜇 is
the average number served per minute, the time t must be given in minutes.

Blithe’s Muffler Example

Blithe’s Muffler Shop installs new mufflers on automobiles and small trucks. The
mechanic can install new mufflers at a rate of about three per hour, and this service
time is exponentially distributed. What is the probability that the time to install a new
muffler will be 12 hour or less? Using Equation 2-17, we have

𝑋 = 𝑒𝑥𝑝𝑜𝑛𝑒𝑛𝑡𝑖𝑎𝑙𝑙𝑦 𝑑𝑖𝑠𝑡𝑟𝑖𝑏𝑢𝑡𝑒𝑑 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑡𝑖𝑚𝑒

𝜇 = 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑡ℎ𝑎𝑡 𝑐𝑎𝑛 𝑏𝑒 𝑠𝑒𝑟𝑣𝑒𝑑 𝑝𝑒𝑟 𝑡𝑖𝑚𝑒 𝑝𝑒𝑟𝑖𝑜𝑑 = 3 𝑝𝑒𝑟 ℎ𝑜𝑢𝑟
1
𝑡 = ℎ𝑜𝑢𝑟 = 0.5 ℎ𝑜𝑢𝑟
2
𝑃 = 𝑋 ≤ 05) = 1 − 𝑒−3(0.5) = 1 − 𝑒−1.5 = 1 − 0.2231 = 0.7769

Figure 2.17 shows the area under the curve from 0 to 0.5 to be 0.7769. Thus, there is
about a 78% chance the time will be no more than 0.5 hour and about a 22% chance
that the time will be longer than this. Similarly, we could find the probability that the
service time is no more 1/3 hour or 2/3 hour, as follows:

1 1
−3( )
𝑃 = (𝑋 ≤ ) = 1 − 𝑒 3 = 1 − 𝑒−1 = 1 − 0.3679 = 0.6321
3
 2 2
−3( )
𝑃 = (𝑋 ≤ ) = 1 − 𝑒 3 = 1 − 𝑒−2 = 1 − 0.1353 = 0.8647
3

Figure 2.17 Probability that the Mechanic Will Install a Muffler in 0.5 Hour

Program 2.5A Excel 2016 Output for the Exponential Distribution

Program 25B Function in an Excel 2016 Spreadsheet for the Exponential

Distribution

The Poisson Distribution

An important discrete probability distribution is the Poisson distribution. We

examine it because of its key role in complementing the exponential distribution in
queuing theory. The distribution describes situations in which customers arrive
independently during a certain time interval, and the number of arrivals depends on
the length of the time interval. Examples are patients arriving at a health clinic,
customers arriving at a bank window, passengers arriving at an airport, and
telephone calls going through a central exchange.

The formula for the Poisson distribution is

(2-18)
where

𝑃(𝑋) = 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑒𝑥𝑎𝑐𝑡𝑙𝑦 𝑋 𝑎𝑟𝑟𝑖𝑣𝑎𝑙𝑠 𝑜𝑟 𝑜𝑐𝑐𝑢𝑟𝑟𝑒𝑛𝑐𝑒𝑠

𝜆 = 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑎𝑟𝑟𝑖𝑣𝑎𝑙𝑠 𝑝𝑒𝑟 𝑢𝑛𝑖𝑡 𝑜𝑓 𝑡𝑖𝑚𝑒 (𝑡ℎ𝑒 𝑚𝑒𝑎𝑛 𝑎𝑟𝑟𝑖𝑣𝑎𝑙 𝑟𝑎𝑡𝑒), 𝑝𝑟𝑜𝑛𝑜𝑢𝑛𝑐𝑒𝑑 “𝑙𝑎𝑚𝑏𝑑𝑎”

𝑒 = 2.718, 𝑡ℎ𝑒 𝑏𝑎𝑠𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑛𝑎𝑡𝑢𝑟𝑎𝑙 𝑙𝑜𝑔𝑎𝑟𝑖𝑡ℎ𝑚

𝑋 = 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑐𝑐𝑢𝑟𝑟𝑒𝑛𝑐𝑒𝑠 (0, 1, 2, … )

The mean and variance of the Poisson distribution are equal and are computed simply
as

Expected value = (2-19)

𝜆
(2-20)
Variance = λ

With the help of the table, the values of 𝑒are easy to find. We can use these in the
formula to find probabilities. For example, if l = 2, from Appendix C we find e-2 =
0.1353.

The Poisson probabilities that X is 0, 1, and 2 when 𝜆 = 2 are as follows:

𝑒 2 (0.1353)1
𝑃(0) = −2 0 = = 0.1353 ≈ 14%
0! 1

𝑒 2 𝑒 2 (0.1353)2
𝑃(1) = −2 1 = −2 = = 0.2706 ≈ 27%
1! 1 1

𝑒 2 𝑒 4 (0.1353)4
𝑃(2) = −2 2 = −2 = = 0.2706 ≈ 27%
2! 2(1) 2

These probabilities, as well as others for 𝜆 = 2 and 𝜆 = 4, are shown in Figure 2.18.
Notice that the chances that 9 or more customers will arrive in a particular time
period are virtually nil. Programs 2.6A and 2.6B illustrate how Excel can be used to
find Poisson probabilities.

It should be noted that the exponential and Poisson distributions are related. If the
number of occurrences per time period follows a Poisson distribution, then the time
between occurrences follows an exponential distribution. For example, if the number
of phone calls arriving at a customer service center followed a Poisson distribution
with a mean of 10 calls per hour, the time between each phone call would be
exponentially distributed with a mean time between calls of 1/10 hour (6 minutes).
 Program 2.6A Excel 2016 Output for Poisson Distribution

Program 2.6B Functions in an Excel 2016 Spreadsheet for the Poisson

Distribution