Continuous Probability Distributions

1. Summary Statistics for RVs

2. EV of a sum of RVs

3. Continuous PD

4. Summary Statistics for continuous RVs

5. Normal Distribution

6. Summary

Lecture Note: Probability Distributions Contents 1

Summary Statistics for RVs
1. Stock returns vary with market conditions
2. Ex of the prob function of two stock returns
Condition Prob stock A stock B
Excellent 0.25 20% 10%
Good 0.4 10% 5%
Poor 0.25 −4% 4%
Crash 0.1 −20% 0%
3. Not easy to compare the RVs or PDs directly
4. Can characterize RVs by summary statistics, such
as expected value and SD, based on their PD
5. Summary statistics computed from the population
dist are called population statistics
Lecture Note: Probability Distributions Summary Statistics for RVs 2
Expected Value (EV)
1. Expected value (EV), or mean, of a RV is a measure
of the central tendency for the RV
2. EV is a weighted mean of all values of the RV
where each data is weighted by its probs
k
X
E(X) = µ = xi P(X = xi )
i=1
3. Ex 1. X ∼ Ber(p)
i. P(X = 1) = p
ii. P(X = 0) = 1 − p
iii. µ = 0 × (1 − p) + 1 × p = p
iv. EV of a Bernulli RV equals the prob of the event
or success
Lecture Note: Probability Distributions Summary Statistics for RVs 3
Expected Value
4. EV does not necessarily correspond to a value that
the RV can actually take
5. Ex 2. X : Roll of a die
1
i. P(X = k) = , k = 1, 2, . . . , 6
6
ii. X follows a uniform dist
1 1 1 1 1 1
iii. µ = 1 × +2× +3× +4× +5× +6×
6 6 6 6 6 6
7
=
2
iv. EV of rolling a die is 7/2

Lecture Note: Probability Distributions Summary Statistics for RVs 4

Expected Returns
1. Prob functions of returns on Stocks A and B
Condition Prob Stock A Stock B
Excellent 0.25 20% 10%
Good 0.4 10% 5%
Poor 0.25 −4% 4%
Crash 0.1 −20% 0%
2. EV of stock returns is called the expected return
3. Stock A’s expected return
E(rA ) = 0.25 × 20 + 0.4 × 10 + 0.25 × (−4) + 0.1 × (−20)
= 6%
4. Can expect 6% return from Stock A on average
5. Expected return on Stock B is 5.5%
Lecture Note: Probability Distributions Summary Statistics for RVs 5
Expected Returns
6. Can we conclude Stock A is better than Stock B?
7. Stock A appears to be riskier than Stock B
8. Expected returns alone do not provide enough
information to make proper investment decisions
9. Need additional measures to evalueate risk
10. What is the risk of stock investment?
11. Suppose you know a stock price will fall by 20%
tomorrow with prob 1 (0.5), is it a risk?
12. Risk associated with stock investment reflects the
uncertainty of its returns
13. Uncertainty can be quantified by the variability or
dispersion of RVs
Lecture Note: Probability Distributions Summary Statistics for RVs 6
Variance
1. Variance (Var) measures the variability in the values
of a RV around its mean
2. Variance of a RV is the EV of the squared
deviations from the EV of the RV
Var(X) = σ2
= E(X − µ)2
Xk
= (xi − µ)2 P(X = xi )
i=1
3. Ex.1 X ∼ Ber(p)
σ2 = (0 − p)2 × (1 − p) + (1 − p)2 × p
= p(1 − p)
Lecture Note: Probability Distributions Summary Statistics for RVs 7
Variance
3. Ex.2 Roll of a die
6 !2
X7 1
σ 2
= i− ×
i=1
2 6
!2 !2 !2
7 1 7 1 7 1
= 1− + 2− + 3−
2 6 2 6 2 6
!2 !2 !2
7 1 7 1 7 1
+ 4− + 5− + 6−
2 6 2 6 2 6
35
=
12

Lecture Note: Probability Distributions Summary Statistics for RVs 8

Variance
4. Ex. Var of Stock A’s return (µA = 6%)
i. Probability functions of two stock returns
Condition Probability stock A stock B
Excellent 0.25 20% 10%
Good 0.4 10% 5%
Poor 0.25 −4% 4%
Crash 0.1 −20% 0%
ii. σ2A = 0.25 × (20 − 6)2 + 0.4 × (10 − 6)2
+0.25 × (−4 − 6)2 + 0.1 × (−20 − 6)2
= 148%2
5. Unit of Var for stock returns (%) is %2
6. Var is difficult to interpret because its unit is squared
Lecture Note: Probability Distributions Summary Statistics for RVs 9
Standard deviation (SD)
1. Standard deviation (SD) is the positive square root
of the variance
√
2. σ = σ2
3. SD has the same units as the data, making it is
easier to interpret than the Var
√
4. Ex. SD of Stock A: σA = 148 = 12.2%
5. Average variation of returns on Stock A is 12.2%
6. SD serves as a measure of uncertainty or risk
7. SD of stock returns is called volatility
8. Stock B’s volatility = σB = SD(rB ) = 3.0%
9. Stock B may be preferable due to its smaller risk
and comparable expected returns
Lecture Note: Probability Distributions Summary Statistics for RVs 10
EV of a sum of RVs
1. What are the EV and Var of Bin(n, p)?
n!
2. P(X = k) = pn (1 − p)n−k
(n − k)!k!
n
X n!
3. µ = E(X) = k· pk (1 − p)n−k
k=0
(n − k)!k!
n
X n!
4. Var(X) = (k − µ)2 · pk (1 − p)n−k
k=0
(n − k)!k!
5. These values can be easily computed using the
properties of the EV and Var of a sum of RVs

Lecture Note: Probability Distributions EV of a sum of RVs 11

EV of a sum of RVs
6. If X and Y are a RV, and a and b are a constant
i. E(aX + b) = aE(X) + b
ii. E(X + Y) = E(X) + E(Y)
7. Do not need to consider the PD of the sum of RVs
to calculate the EV of the sum of RVs
8. If X and Y are independent (no relationship)
i. E(XY) = E(X)E(Y)
ii. Var(X + Y) = Var(X) + Var(Y)
9. Variance of a sum of RVs is not generally equal to
the sum of individual variances

Lecture Note: Probability Distributions EV of a sum of RVs 12

Ex. Two coin tosses
1. Suppose you toss a fair coin twice
2. X : RV taking 1 (0) if head (tail) on the 1st coin toss
3. Y : RV taking 1 (0) if head (tail) on the 2nd coin toss
4. X and Y follow Ber(1/2)
5. E(X) = E(Y) = 1/2
6. E(X + Y) = E(X) + E(Y) = 1/2 + 1/2 = 1
7. Distribution of X + Y
X+Y 0 1 2
P 1/4 1/2 1/4
8. E(X + Y) = 1 · 1/2 + 2 · 1/4 = 1 = E(X) + E(Y)
Lecture Note: Probability Distributions EV of a sum of RVs 13
Ex. Two coin tosses
9. Var(X) = Var(Y) = 1/4
10. Var(X + Y) = Var(X) + Var(Y) = 1/4 + 1/4 = 1/2
11. Distribution of X + Y
X+Y 0 1 2
P 1/4 1/2 1/4

1 1 1
12. Var(X + Y) = (0 − 1)2 · + (1 − 1)2 · + (2 − 1)2 ·
4 2 4
1
=
2
= Var(X) + Var(Y)

Lecture Note: Probability Distributions EV of a sum of RVs 14

Ex. n coin tosses
1. Suppose you toss a coin with prob for a head equal
to p, repeated n times
2. Xi : RV taking 1 (0) if head (tail) on the ith coin toss
3. Xi ∼ Ber(p)
4. X : total number of heads in n coin tosses
5. X = X1 + X2 + · · · + Xn
6. X ∼ Bin(n, p)
n
X n!
7. µ = E(X) = k· pk (1 − p)n−k
k=0
(n − k)!k!
n
X n!
8. Var(X) = (k − µ)2 · pk (1 − p)n−k
k=0
(n − k)!k!
Lecture Note: Probability Distributions EV of a sum of RVs 15
Ex. n coin tosses
9. E(X) and Var(X) can be computed easily using the
results of a sum of RVs
10. E(Xi ) = p
 n  n
X  X
11. E(X) = E   Xi  =
 E(Xi ) = np
i=1 i=1
12. Var(Xi ) = p(1 − p)
 n  n
X  X
13. Var(X) = Var   Xi  =
 Var(Xi ) = np(1 − p)
i=1 i=1

Lecture Note: Probability Distributions EV of a sum of RVs 16

Continuous PD
1. Continuous RV can take any values within an
interval or a collection of intervals
i. Heights of grade 3 students
ii. Daily maximum temperature
iii. Daily stock returns of a stock
2. What is the prob that the height of a grade 3 student
is exactly 125cm?
3. For continous RVs, we cannot consider the prob of
taking an exact value
4. Impossible to assign probs to all possible values
5. Assume P(X = x) = 0 for all x for continuous RVs

Lecture Note: Probability Distributions Continuous PD 17

Continuous PD
6. Can assess the prob of a continuous RV taking a
value within a specific interval
7. What is the prob that the height of a grade 3 student
is btw 120cm and 130cm?
8. Continuous PD is represnted by a probability
density function, denoted by f (x)

Lecture Note: Probability Distributions Continuous PD 18

Probability density function
1. Density function f (x) must satisfy
i. f (x) ≥ 0, ∀x
R∞
ii. −∞
f (x) = 1
2. Ex. Uniform distribution btw a and b, U(a, b)
1

if a ≤ x ≤ b


f (x) = 

b−a


0 otherwise


1/(b-a)
0

-
a b
Lecture Note: Probability Distributions Continuous PD 19
Probability density function
3. f (x) is related to a likelihood of X taking x
4. Area under the curve of f (x) btw a1 and a2 gives the
probability that X lies within that interval
R a2
5. a1
f (x)dx = P(a1 < X < a2 ) = P(a1 ≤ X ≤ a2 )

Lecture Note: Probability Distributions Continuous PD 20

Probability density distribution
6. Histogram of a large number of observations drawn
from a continuous distribution tends to resemble its
probability density function of(x)
Histogram f Time
0
N
0

>, 0
."!:
...--
C: 0

,0

I.()
0
0

0
0
0

0 2 4 6 8 10 12
Lecture Note: Probability Distributions Continuous PD
Time 21
Summary Statistics for continuous RVs
1. EV and Var are defined similarly with discrete RVs
using the density function and integration
Z ∞
E(X) = µ =
x f (x)dx
Z ∞ −∞

Var(X) = E(X − µ) = 2
(x − µ)2 f (x)dx
−∞
SD(X) =
p
Var(X)
2. Ex. Uniform distribution U(a, b)
a+b
Z b
x
µ= dx =
a b−a 2
(x − µ)2
Z b
(b − a)2
Var(X) = dx =
a b−a 12
Lecture Note: Probability Distributions Summary Statistics for continuous RVs 22
Ex. Uniform distribution
1. Suppose X follows U(1, 6)
2. Probability (density function of X
0.2 if 1 ≤ x ≤ 6
f (x) =
0 otherwise
0.2
0

-
1 6

3. P(X < 1.5) = (1.5 − 1) × 0.2 = 0.1

4. P(X > 3) = (6 − 3) × 0.2 = 0.6
5. E(X) = (6 + 1)/2 = 3.5
6. Var(X) = (6 − 1)2 /12 = 25/12
Lecture Note: Probability Distributions Summary Statistics for continuous RVs 23
Normal distribution
1. Normal distribution is one the most important PDs
for describing a continuous RV btw −∞ and ∞
2. Developed by Carl Friedrich
Gauss to describe measurement
errors and natural variations
3. Appear in a variety of applications
i. Heights and weights
ii. Economic growth rates
iii. Inflation rates
iv. Test scores
4. Play a fundamental role in IS

Lecture Note: Probability Distributions Normal Distribution 24

Normal distribution
5. Normal density is a symmetric, bell-shaped curve

6. Density function f (x) of

" the normal
# dist
(x − µ)2 1
f (x) = √ exp −
2πσ 2σ2
7. Need to determine µ (EV) and σ (SD) to specify a
normal distribution
Lecture Note: Probability Distributions Normal Distribution 25
Normal distribution
8. µ and σ are the parameters of the normal dist
Statistics for Business and Economics (14e, Metric Version)

9. If X follows a normal dist with the parameters µ and

Normal Probability Distribution
σ, we write X ∼ N(µ, σ2 ) (4 of 7)
Theµmean
10. cancan
bebeany
any numerical value:and
real value negative, zero, or positive.
determines the center
(location) of the curve

11. µ also corresponds to the median and mode of the

dist
© 2020 Cengage. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part, except for use as permitted in a license distributed with a certain product or service or otherwise on a
password‐protected website or school‐approved learning management system for classroom use.

Lecture Note: Probability Distributions Normal Distribution 26
ormal Probability
Normal Distribution (5 of 7)
distribution
The12. σ determines
standard the spreadthe
deviation determines (width) ofthe
width of thecurve:
curvelarger values
in 13.
wider, flatterσ
Larger curves.
produces a wider and flatter dist
14. Smaller σ results in a wider and flatter curve

engage. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part, except for use as permitted in a license distributed with a certain product or service or otherwise on a
protected website or school‐approved learning management system for classroom use.

Lecture Note: Probability Distributions Normal Distribution 27

ribution (7 of 7)
Normal distribution
15. Approximately 95.4% of the values of a normal RV
are within ±2 SDs of its mean

n.

n.

Lecture Note: Probability Distributions Normal Distribution 28

Normal distribution
16. If X and Y are normally distributed, X + Y also follows
a normal dist
17. If X ∼ N(µ, σ2 ), the standardized variable
X−µ
Z= ∼ N(0, 1)
σ
18. N(0, 1) is called the standard normal dist
19. N(0, 1) serve as the limiting dist for the Central Limit
Theorem (CLT)
20. Percentiles of N(0, 1) often used in IS
i. 99.5 percentile: 2.576
ii. 97.5 percentile: 1.960
iii. 95 percentile: 1.645
Lecture Note: Probability Distributions Normal Distribution 29
Normal distribution
21. For X ∼ N(0, 1), P(X < −1.96) = 0.025 = P(X > 1.96)

2.5% prob 2.5% prob

-1.96 1.96

22. P(X < x), P(X > x), and P(|X| > x) are called
lower-tail, upper-tail, and two-tail probs

Lecture Note: Probability Distributions Normal Distribution 30

Discrete vs continuous distributions
1. RVs and PDs are classified as either discrete or
continuous
2. Key differences between the two types of PDs
Discrete dist Continuous dist
Possible values finite or infinite se- any value within an
quence interval
P(X = xi ) > 0 possible P(X = x) = 0 for all x
X Z b
P(a ≤ X ≤ b) p(xi ) f (x) dx
a≤xi ≤b a
k
X Z ∞
EV xi P(X = xi ) xf (x)dx
i=1 −∞
Examples Uniform, Binomial Uniform, Normal
Lecture Note: Probability Distributions Summary 31