Example of discrete random variable

Example of discrete random variable Example 2: Tossing a coin twice. Examples of continuous random variable Examples of continuous random variable
Example 1: Tossing a coin once. I Sample space S = {HH, HT,TH, TT} where H means Head Example 1: Temperature of a given date. Example 2: Stock market index.
I Sample space S = {H, T} where H is head and T is tail and T means tail. I Sample space S = {x œ R : x Ø ≠273.15}. I Sample space S = R + = {x œ R : x Ø 0}.
I Let X be a random variable of total number of heads. I Let X be a random variable of total number of heads. I Let X be a random variable of temperature. I Let X be a random variable of stock market index.
I X = 1 when we have a head, or X (H) = 1. I Then X = 0 means TT, X = 1 means HT or TH, and X = 2 I Support of X is the same as S. I Support of X is the same as S.
means HH.
I X = 0 when we have a tail, or X (T) = 0.
I Support of X is {0,1}.
I Formally, X (TT) = 0, X (TH) = X (HT) = 1, and Notation: R is the set of real numbers and R + is the set Notation: R is the set of real numbers and R + is the set
X (HH) = 2. of non-negative real numbers. of non-negative real numbers.
I Support of X is {0,1,2}.

Chiu Yu KO 6 / 213 Chiu Yu KO 8 / 213 Chiu Yu KO 12 / 213 Chiu Yu KO 13 / 213

Descriptions of discrete random variable Another description of discrete random variable

Example: Toss a fair coin
Example: Toss a fair coin
I Sample space S = {H,T}.
I Sample space S = {H,T}.
I Let X be the random variable of number of head.
Probability mass function (pmf) I Let X be the random variable of number of head. Cumulative distribution function (cdf) I Then X (H ) = 1 and X (T ) = 0.
I Then X (H ) = 1 and X (T ) = 0.
P (x ) = Pr (X = x )
ÿ
F (x ) = Pr (X Æ x ) = P (y ) X 0 1
X 0 1 y Æx P (x ) 1 1
1 1 2 2
P (x ) F (x ) 1
1
Alternative name: 2 2 2
I Probability function (pf) Alternative name:
I P (0) = Pr(X = 0) = Pr(T ) = 12 . I F (0) = Pr(X Æ 0) = Pr(X = 0) = P (0) = 12 .
I Probability distribution function (pdf) I Distribution function (df)
I P (1) = Pr(X = 1) = Pr(H ) = 1
2. I F (1) = Pr(X Æ 1) = Pr(X = 0) + Pr(X = 1) =
P (0) + P (1) = 1.

Chiu Yu KO 16 / 213 Chiu Yu KO 17 / 213 Chiu Yu KO 18 / 213 Chiu Yu KO 19 / 213

Example: Toss a fair coin Example: Toss a fair coin twice

Example: Toss a fair coin twice
X 0 1 I Sample space S = {HH,HT,TH,TT}. X 0 1 2 Example: Rolling a fair die
1 1 1 1 1
P (x ) 2 2 I Let X be the random variable of number of heads. P (x ) 4 2 4
F (x ) 1
1 F (x ) 1 3
1 I Sample space S = {1, 2, 3, 4, 5, 6}.
2 I Hence, X (HH ) = 2, X (HT ) = X (TH ) = 1 and X (TT ) = 0. 4 4
I Let X be the random variable of face value of a fair die.
P (x ) F (x ) P (x ) F (x )
X 0 1 2 I Then X can be integers from 1 to 6 (inclusive).
1 1 1
1 P (x ) 4 2 4 1
F (x ) 1 3
1 X 1 2 3 4 5 6
4 4 1 1 1 1 1 1
3 P (x ) 6 6 6 6 6 6
4 1 2 3 4 5
I Note that P (0) = Pr(X = 0) = Pr(TT ) = 14 , P (1) = F (x ) 6 6 6 6 6 1
1 1 1 1
2 2 Pr(X = 1) = Pr(HT fi TH ) = Pr(HT ) + Pr(TH ) = 12 , and 2 2
I Draw P (x ) and F (x ).
P (2) = Pr(X = 2) = Pr(HH ) = 14 . 1 1
4 4
I Also we have F (0) = P (0), F (1) = P (0) + P (1), and
x x F (2) = P (0) + P (1) + P (2) x x
0 1 0 1 0 1 2 0 1 2
Chiu Yu KO 21 / 213 Chiu Yu KO 22 / 213 Chiu Yu KO 23 / 213 Chiu Yu KO 24 / 213

Expectation Linear property of expectation

Example: Rolling a fair die
Example: Tossing a fair coin
Expectation Let X be the number of head.
Let X be the face value of the die.
Linear property of expectation
X 1 2 3 4 5 6
Weighted average with probability as weights: X 0 1 1 1 1 1 1 1 Let a and b be constants.
P (x ) 6 6 6 6 6 6
1 1
ÿ P (x ) 2 2 Let X be random variables.
E (X ) = xP (x ) Expected value of X is
Expected value of X is
1
3 4
1
3 4
1 1
3 4 3 4 E (aX ) =aE (X )
Denote E (X ) = µX . 3 4
1 1
3 4 E (X ) =1 +2 +3 +4
E (X ) = 0 +1 = 0.5 6 6 6 6 E (X + b ) =E (X ) + b
Expectation is mean of random variable: 2 2 3 4
1
3 4
1
+5 +6 = 3.5 E (aX + b ) =aE (X ) + b
6 6
I keep doing the random experiment
I probability is the relative frequency

Chiu Yu KO 29 / 213 Chiu Yu KO 32 / 213 Chiu Yu KO 33 / 213 Chiu Yu KO 34 / 213

 Example: Stock price
Variance Example: Rolling a fair die Useful formula for variance
Let X be the face value of the die.
Let X be the price of a stock.
X 1 2 3 4 5 6
X 1 2 3 P (x ) 1
6
1
6
1
6
1
6
1
6
1
6
Useful formula for variance
P (x ) 1
4
1
2
1
4 Variance Let a and b be constants.
Recall that E (X ) = 3.5.
Expected value of X is Expected squared difference from mean: Variance of X is Let X be a random variable.
3 4 3 4 3 4
1 1 1 1 2
1
3 4
1
3 4
[x ≠ E (X )]2 P (x )
ÿ
E (X ) = 1 +2 +3 =2 Var (X ) = E [X ≠ E (X )]2 = Var (X ) = (1 ≠ 3.5)
2
+ (2 ≠ 3.5)2 Var (aX ) =a2 Var (X )
4 2 4 6 6
3 4
1
3 4
1 Var (X + b ) =Var (X )
What is the expected value of portfolio that has 100 shares of stock Denote Var (X ) = ‡X2 . + (3 ≠ 3.5)2 + (4 ≠ 3.5)2
X and 300 dollars of cash? 6
3 4
6
3 4 Var (aX + b ) =a2 Var (X )
1 1
+ (5 ≠ 3.5)2 + (6 ≠ 3.5)2
E (100X + 300) = 100E (X ) + 300 = 100(2) + 300 = 500 6 6
¥2.9167

Chiu Yu KO 35 / 213 Chiu Yu KO 37 / 213 Chiu Yu KO 38 / 213 Chiu Yu KO 39 / 213

Example: Stock price Expectation of the sum Variance of the sum

Example
Let X be the price of a stock.
Suppose the mean return and variance for stock X are 0.1 and 0.05.
X 1 2 3 Expectation of the sum Variance of the sum Those for stock Y are 0.05 and 0.01. Consider a portfolio Z that
P (x ) 1 1 1 equally divided between two stocks. That is, Z = 0.5X + 0.5Y .
4 2 4
Expectation of the sum is the sum of expectations: Let X and Y be random variables. Expectation of the portfolio is
Recall that expected value of X is 2. Now variance of X is If X and Y are independent, then we have
ÿÿ E (Z ) =E (0.5X + 0.5Y ) = E (0.5X ) + E (0.5Y )
3 4 3 4 3 4 E (X + Y ) = (x + y )P (x , y ) = E (X ) + E (Y )
1 1 1 1 x y ÿÿ =0.5(0.1) + 0.5(0.05) = 0.075
Var (X ) = (1 ≠ 2)2
4
+ (2 ≠ 2)2
2
+ (3 ≠ 2)2
4
=
2
Var (X + Y ) = (x + y ≠ (E (X + Y ))2 P (x , y )
x y
By linear property of expectation, we have If the two stocks are independent, variance of the portfolio is
What is the variance of portfolio that has 100 shares of stock X and =Var (X ) + Var (Y )
300 dollars of cash? Var (Z ) =Var (0.5X + 0.5Y ) = Var (0.5X ) + Var (0.5Y )
E (aX + bY ) = aE (X ) + bE (Y )
1 =0.52 (0.05) + 0.52 (0.01) = 0.015
Var (100X + 300) = 1002 Var (X ) = 1002
2
= 5000 Variance of the sum is the sum of variances.

Chiu Yu KO 40 / 213 Chiu Yu KO 46 / 213 Chiu Yu KO 47 / 213 Chiu Yu KO 48 / 213

Common Discrete Distributions Uniform Distribution Graph for Uniform distribution Bernoulli Distribution
I Every possible value of X is equally likely I Two outcomes: failure or success
I Support of X : {x1 , . . . , xn } I Probabilities of success and failure: p and 1 ≠ p
P (x ) I Random variable X : number of success
I Uniform 1
Pr (X = xi ) = X = 1, ..., 5 I Support of X : {0, 1}
I Bernoulli n I Then Pr(X = 1) = p and Pr(X = 0) = 1 ≠ p
I Binomial X = 1, ..., 3
I Probability function I Probability function
I Possion
I Hypergeometric X x1 · · · xn X 0 1
P (x ) n1 · · · 1 P (x ) 1 ≠ p p
n x
1 2 3 4 5 I Usage: tossing a fair coin, running an experiment
I Example: number of heads tossing a fair coin, face
value of rolling fair dice with binary outcomes

Chiu Yu KO 51 / 213 Chiu Yu KO 52 / 213 Chiu Yu KO 53 / 213 Chiu Yu KO 54 / 213

Graph for Bernoulli distribution Binomial Distribution Binomial Distribution

Bernoulli Distribution
Denote n a positive integer and 0 < p < 1:
I Probability function: I Support of X : {0, 1, . . . , n}
1≠x P (x ) I n independent Bernoulli trials I Probability function:
P (X = x ; p ) = p x (1 ≠ p ) I each trial has a success probability p
p = 0.5 Q R
p = 0.3 I Random variable X : number of successes n n≠x
I Mean: E (X ) = p P (x ; n, p ) = a bp x (1 ≠ p )
p = 0.8 I Support of X : {0, 1, . . . , n} x
I Variance: Var (X ) = p (1 ≠ p ) I Using combinatorics, we know
I Mean: E (X ) = np
Intuition: n
Q R
n≠x I Variance: Var (X ) = np (1 ≠ p )
I mean: more successes with a higher p x Pr (X = x ; n, p ) = a bp x (1 ≠ p )
x
I variance: lowest when most successes/failures 0 1 Intuition:
(under extreme values of p: p = 0 or p = 1), I Binomial X is the sum of n independent Bernoulli.
highest when most uncertain (p = 1/2) I The mean and variance of X are simply n times of
mean and variance of a Bernoulli random variable.
Chiu Yu KO 55 / 213 Chiu Yu KO 56 / 213 Chiu Yu KO 58 / 213 Chiu Yu KO 60 / 213
Graph for binomial distribution Poisson Distribution
Example

Example
A restaurant owners found that 30% of customers like spicy food. I Partition a period of time with n intervals
P (x ) What is the probability that out of five randomly chosen customers, I Each internal has a Bernoulli trial
A restaurant owners found that 30% of customers like spicy food. no more 2 of them like spicy food?
n = 6, p = 0.5 What is the probability that out of five randomly chosen customers,
I Each trial has success probability p
n = 4, p = 0.3 2 of them like spicy food? Pr(X Æ 2) = Pr(X = 0) + Pr(X = 1) + Pr(X = 2) I Each trial are independent
A B
A B
5 5
A B I Random variable X : total number of successes
5 = (0.3)0 (0.7)5 + (0.3)1 (0.7)4 I Support of X : {0, 1, . . . , n}
Pr(X = 2) = (0.3)2 (0.7)3 ¥ 0.3087 0 1
2 A B
5 I Probability function:
Note that we have n = 5, x = 2 and p = 0.3. + (0.3)2 (0.7)3
x 2 Q R
n n≠x
0 1 2 3 4 5 6 ¥ 0.8369 Pr (X = x ; n, p ) = a bp x (1 ≠ p )
x

Chiu Yu KO 61 / 213 Chiu Yu KO 62 / 213 Chiu Yu KO 63 / 213 Chiu Yu KO 74 / 213

Poisson Distribution Hypergeometric Distribution

Poisson Distribution
Denote ⁄ = np and e ¥ 2.718: I Total number of items: N
I Support of X : {0, 1, . . . , } Example I Number of good items: A Æ N
I When n is very large and p is small, we have I Number of bad items: N ≠ A
I Probability function: The number of customers for a restaurant each hour follows Poisson
I Probability function: distribution with mean 5. What is the probability that exactly 2 I Take n items (without replacement)
Q R e ≠⁄ x
⁄ customers will arrive in next hour? I Random variable X : number of good items
n n≠x e ≠np (np )x P (x |⁄) =
Pr (X = x ; n, p ) = a bp x (1 ≠ p ) ¥ x! e ≠5 52 I Support of X : x œ {0, 1, ..., n}. Note that n Æ A
x x! Pr(X = 2) = ¥ 0.084 I By combinatorics, we have
2!
I Mean: E (X ) = ⁄
when n is large and p is small. Note that ⁄ = 5 and x = 2.
I Variance: Var (X ) = ⁄ (Ax)(N≠A
I Usage: arrival of customers over a period of time n≠x )
Pr (X = x |n, N, A) =
Different from binomial, we have mean = variance (Nn )
I Var (X ) = np (1 ≠ p ) ¥ np because 1 ≠ p ¥ 1. I Usage: quality control, audit
Chiu Yu KO 75 / 213 Chiu Yu KO 76 / 213 Chiu Yu KO 78 / 213 Chiu Yu KO 90 / 213

Hypergeometric Distribution Continuous Random Variable

Denote N, A, n positive integers where A Æ N and
n Æ A: Example Cumulative distribution function (cdf)
I Support of X : {0, 1, . . . , n} In a warehouse, there are 30 computers and 4 of them are defective.
I Probability function: F (x ) = Pr (X Æ x )
What is the probability that two of them are defective out of a
sample of six? Continuous Random Variable
(Ax)(N≠A
n≠x ) Useful property of cdf
P (X = x |n, N, A) = (42)(264 ) ¥ 0.1511.
(Nn ) Pr(X = 2|6, 30, 4) =
(30
6) Denote a and b some constants such that a Æ b.
nA Note that N = 30, A = 4, n = 6 and x = 2.
I Mean: E (X ) = Pr(a < X Æ b ) = F (b ) ≠ F (a )
N
nA N ≠ A N ≠ n Pr(X > b ) = 1 ≠ F (b )
I Variance: Var (X ) =
N N N ≠1
When n = 1, we have a Bernoulli trial.
Chiu Yu KO 91 / 213 Chiu Yu KO 94 / 213 Chiu Yu KO 104 / 213 Chiu Yu KO 109 / 213

Continuous Random Variable Expectation and Variance Common Continuous Distributions

Although we cannot say probability at one value using Example: Area under f(x) is probability
P (x ), we have something similar for a small interval:
f F (a ) Expectation
Probability distribution function (pdf) x
⁄ Œ
a E (X ) = xf (x ) dx I Uniform
If b ≠ a > 0 is small and a Æ x Æ b, we have ≠Œ
I Exponential
f 1 ≠ F (a ) I Normal
Pr(a Æ X < b ) ¥ f (x ) (b ≠ a ) z Variance I Chi-square
a I Student-t
⁄ Œ
Alternatively, for some very small D, we have Var (X ) = [x ≠ E (X )]2 f (x ) dx
F (b ) ≠ F (a )
≠Œ I F
f
Pr(x Æ X < x + D) z
f (x ) ¥ a b
Note that they are the same as the discrete counterparts
D s
after we replace “ ” by “ ”, and forget about “dx ”.
q

Also known as probability density function (pdf)

Chiu Yu KO 110 / 213 Chiu Yu KO 112 / 213 Chiu Yu KO 116 / 213 Chiu Yu KO 118 / 213
 Uniform Distribution
Denote a, b positive number such that a < b: Example Example
Example
I Support of X : [a, b ] The daily rainfall follows a uniform distribution between 1mm and The daily rainfall follows a uniform distribution between 1mm and
The daily rainfall follows a uniform distribution between 1mm and 4mm.
I Cumulative probability function: 4mm. 4mm.
What is the probability that the rainfall is between 1mm and 3mm? What is the probability that the rainfall is between 2mm and 3mm?
What is the expected daily rainfall and its variance?
x ≠a
F (x ; a, b ) = 1+4
Pr(1 Æ X Æ 3) = Pr(X Æ 3) ≠ Pr(X < 1) Pr(2 Æ X Æ 3) = Pr(X Æ 3) ≠ Pr(X < 2)
b≠a E (X ) = = 2.5 = Pr(X Æ 3) ≠ Pr(X Æ 1)
2 = Pr(X Æ 3) ≠ Pr(X Æ 2)
I Mean: E (X ) = a+2 b (4 ≠ 1)2 3 3≠1 1≠1 = F (3) ≠ F (2)
Var (X ) = =F (3) ≠ F (1) = ≠
(b≠a )2 12
=
4 4≠1 4≠1 3≠1 2≠1 2 1
I Variance: Var (X ) = 12 2 = ≠ = ≠ ¥ 0.33.
Note that a = 1 and b = 4. = ≠ 0 ¥ 0.67. 4≠1 4≠1 3 3
3
Intuition:
I Mean: the middle of the interval.
I Variance: proportional to the squared range
Chiu Yu KO 122 / 213 Chiu Yu KO 124 / 213 Chiu Yu KO 125 / 213 Chiu Yu KO 126 / 213

Exponential Distribution
Exponential Distribution
Example
I Waiting time until a success in Possion with mean ⁄ Denote ⁄ a positive number:
Example Suppose the time for next customer to arrive follows exponential
I Memoryless: same probability after waiting I Support of X : (0, Œ)
Suppose the time for next customer to arrive follows exponential distribution with mean 20 minutes.
I Support of X : [0, Œ) I Cumulative probability function: What is the probability that a customer arrive in between next 5
distribution with mean 20 minutes.
I Probability distirubtion function What is the probability that a customer arrive in within next 10 minutes and 15 minutes?
Y F (x ; ⁄) = 1 ≠ e ≠⁄x minutes? Note that mean = 20 implies that ⁄ = 20 1
. Hence, we have
] ⁄e ≠⁄x if x Ø 0 Note that mean = 20 implies that 20 = ⁄1 or ⁄ = 20 1
. Hence, we
f (X = x |⁄) = [ I Mean: E (X ) = ⁄≠1 Pr(5 Æ X Æ 15) = F (15) ≠ F (5)
0 otherwise have 1
(10) 1 1
I Variance: Var (X ) = ⁄≠2 Pr(X Æ 10) = F (10) = 1 ≠ e ≠ 20
¥ 0.3934 = (1 ≠ e ≠ 20 (15) ) ≠ (1 ≠ e ≠ 20 (5) )
I Cumulative distribution function ¥ 0.30634
Intuition:
F (x ) = 1 ≠ e ≠⁄x I When we expect ⁄ of events happen in an hour,
then we expect 1/⁄ is the arrival of the first event
Chiu Yu KO 130 / 213 Chiu Yu KO 131 / 213 Chiu Yu KO 135 / 213 Chiu Yu KO 136 / 213

Normal Distribution Graph for normal distribution Normal Distribution Shorthand notation for normal
Denote µ a number and ‡ a positive number:
I Bell curve I Support of X : (≠Œ, Œ)
I Support of X : (≠Œ, Œ) I Probability desnity function:
I Probability distribution function
µ = 0, ‡ 2 = 4
µ = ≠4, ‡ 2 = 1
Shorthand notation for normal
µ = 0, ‡ 2 = 1 1
2
2 1 1 (x ≠µ)
2
If X follows normal distribution with mean µ, and
f X = x |µ, ‡ =Ô e ≠ 2 ‡2
1 2 1 1 (x ≠µ)
2
2fi‡ 2 variance ‡ 2 , we write
f X = x |µ, ‡ 2 = Ô e ≠ 2 ‡2
2fi‡ 2
I Mean: E (X ) = µ X ≥ N (µ, ‡ 2 ).
I Distribution function: no functional form–required ≠4 0 4 I Variance: Var (X ) = ‡ 2
calculator or table to find F (x )
I Usage: very common–IQ, weight, height Remarks:
I µ: change the location of distribution
I ‡ 2 : change the spread of distribution
Chiu Yu KO 145 / 213 Chiu Yu KO 146 / 213 Chiu Yu KO 148 / 213 Chiu Yu KO 152 / 213

Linear transformation Standardized normal distribution Z-table

1
Linear transformation
µ
A linear transforamtion with a = and b = ≠ : Example: Standard Normal
2
If X ≥ N (µ, ‡ ), then
‡ ‡ Z-table gives F (z ) = Pr(Z < z ) with two versions:
Let Z follow a standard normal distribution. Using the Z-table, we
Standardization I Version 1: z Æ 0. have
aX + b ≥ N (aµ + b, a ‡ ) 2 2
If X ≥ N (µ, ‡ 2 ), then I Version 2: z Ø 0. I Pr (Z < 1.4) = 0.9192
Since f (Z = z ) is symmetric about 0, two verions are I Pr(Z < ≠1.6) = 1 ≠ Pr(Z < 1.6) = 1 ≠ 0.9452 = 0.0548
X ≠µ interchangable because I Pr (≠1.6 < Z < 1.4) = Pr(Z < 1.4) ≠ Pr(Z < ≠1.6) =
Still normal distribution after a linear transformation Z = ≥ N (0, 1)
‡ 0.9192 ≠ 0.0548 = 0.8644
I Adding a constant changes only the mean: parallel Pr(Z < ≠z ) = 1 ≠ Pr(Z < z )
shifting
I Multiplying a constant is changing only mean and Z is called “standard normal”: Z is saying how many
variance: shifting and scaling standard deviations ‡ is X from mean µ?

Chiu Yu KO 153 / 213 Chiu Yu KO 154 / 213 Chiu Yu KO 156 / 213 Chiu Yu KO 157 / 213
Graph for standard normal examples Special values from Z-table Graph for empirical rule Inverse Normal

0.9192
Consider Z ≥ N (0, 1). Using the Z-table inversely, we
z
0 1.4 Rule of thumb can
Empirical rule, or three sigma rule or 68-95-99-rule: I find z such that Pr(Z < z ) for any given
0.0548 68% 0 Æ – Æ 1.
I Pr (≠1 < Z < 1) ¥ 0.68,
z I Pr (≠2 < Z < 2) ¥ 0.95, 95% I By complement rule, we can find find z such that
≠1.6 0 Pr(Z > z ) for any given 0 Æ – Æ 1.
I Pr (≠3 < Z < 3) ¥ 0.99 99%
I Denote z– such that Pr(Z > z– ) = – for
0.8644 z
≠3 ≠2 ≠1 0 1 2 3 0 Æ – Æ 1.
z
≠1.6 0 1.4

Chiu Yu KO 158 / 213 Chiu Yu KO 163 / 213 Chiu Yu KO 164 / 213 Chiu Yu KO 165 / 213

Graph for normal examples

Example: Standard Normal Example 0.1587 Example

Let Z follow a standard normal distribution. Let X follow normal distribution with µ = 20 and ‡ 2 = 16: Let X follow normal distribution with µ = 20 and ‡ 2 = 16:
What is z such that Pr(Z < z ) = 0.95? I Probability that X is less than 16: x I Probability that X is greater than 24:
I Using the Z-table, we have z = 1.645. 16 20
A B A B
What is z such that Pr(Z > z ) = 0.025? X ≠ 20 16 ≠ 20 X ≠ 20 24 ≠ 20
Pr (X < 16) = Pr < Pr (X > 24) = Pr >
I Using complement rule, this is the same as “What is z such 4 4 4 4
that Pr(Z < z ) = 0.975?” = Pr (Z < ≠1) = 0.1587 0.1587 = Pr (Z > 1) = 1 ≠ 0.8413 = 0.1587
I Using the Z-table, we have z = 1.96.
z
≠1 0

Chiu Yu KO 168 / 213 Chiu Yu KO 169 / 213 Chiu Yu KO 170 / 213 Chiu Yu KO 171 / 213

Graph for normal examples Graph for normal examples

Example
0.8413 0.8413
Let X follow normal distribution with µ = 20 and ‡ 2 = 16: 0.1587
0.1587
I Probability that X is between than 16 and 24:
x A
16 ≠ 20 X ≠ 20 24 ≠ 20
B x
20 24 Pr (16 < X < 24) = Pr < < 16 20 24
4 4 4
= Pr (≠1 < Z < 1)
0.1587 = Pr (Z < 1) ≠ Pr (Z < ≠1) 0.6826
=0.8413 ≠ 0.1587 = 0.6826

z z
0 1 ≠1 0 1

Chiu Yu KO 172 / 213 Chiu Yu KO 173 / 213 Chiu Yu KO 174 / 213