Sec 4.

Poisson
Distribution
 𝑿~𝑷𝒐𝒊 𝝁
X is a 𝑿~𝑷𝒐𝒊 𝝀
Poisson rv.
If X counts Lambda
the number of
successes in 𝑒 −𝜇 𝜇 𝑥
some length 𝑃 𝑋=𝑥 =
𝑥!
of time/space
then… 𝑥 = 0, 1, 2, …

The Binomial and Poisson random

variables both count the number of
successes but under different i.e. 𝜇 𝐸 𝑋 = 𝑉𝑎𝑟(𝑋)
circumstances.
Example At a particular restaurant, an average of 4 plates are
broken each week.

𝑋 = 𝑛𝑜. 𝑜𝑓 𝑝𝑙𝑎𝑡𝑒𝑠 𝑏𝑟𝑜𝑘𝑒𝑛 𝑒𝑎𝑐ℎ 𝑤𝑒𝑒𝑘

𝑋~𝑃𝑜𝑖 4
𝑒 −𝜇 𝜇 𝑥
𝑃 𝑋=𝑥 =
𝑥!

(a) P( 2 plates are broken next week )

𝑒 −4 42
= 𝑃 𝑋=2 = = 0.1465
2!
𝑋 = 𝑛𝑜. 𝑜𝑓 𝑝𝑙𝑎𝑡𝑒𝑠 𝑏𝑟𝑜𝑘𝑒𝑛 𝑒𝑎𝑐ℎ 𝑤𝑒𝑒𝑘 𝑒 −𝜇 𝜇 𝑥
𝑃 𝑋=𝑥 =
𝑋~𝑃𝑜𝑖 4 𝑥!

(b) P( at most 1 plate is broken next week )

= 𝑃 𝑋≤1

= 𝑃 𝑋=0 + 𝑃 𝑋=1

𝑒 −4 40 𝑒 −4 41
= +
0! 1!

= 0.0916
 𝑋~𝑃𝑜𝑖 4
(c) P( more than 3 plates are broken next week )
𝑒 −𝜇 𝜇 𝑥
𝑃 𝑋=𝑥 =
= 𝑃 𝑋>3 𝑥!

= 1 − 𝑃(𝑋 ≤ 3)

𝑒 −4 40 𝑒 −4 41 𝑒 −4 42 𝑒 −4 43
=1 − − − −
0! 1! 2! 3!

= 0.3712
x x x x      
0 1 2 3 4 5 6 7 8 9
 Stays a
No change Poisson r.v.
The mean is of type of
linked to a distribution
specific length of
time/space.
Double length
A change of Old: µ= 10
length of New: µ=20
time/spaces Proportional
results in… change in µ
Divide length by 10
Old: µ = 30
New: µ = 3
Example continued. 𝑋 = 𝑛𝑜. 𝑜𝑓 𝑝𝑙𝑎𝑡𝑒𝑠 𝑏𝑟𝑜𝑘𝑒𝑛
𝑒𝑎𝑐ℎ 𝑤𝑒𝑒𝑘
(d) What is the probability that 10 plates are
𝑋~𝑃𝑜𝑖 4
broken in the next month (4 weeks)?
𝑒 −𝜇 𝜇 𝑥
𝑃 𝑋=𝑥 =
𝑥!
Define: 𝑌 = 𝑛𝑜. 𝑜𝑓 𝑝𝑙𝑎𝑡𝑒𝑠 𝑏𝑟𝑜𝑘𝑒𝑛 𝑒𝑎𝑐ℎ 𝑚𝑜𝑛𝑡ℎ
𝑌~𝑃𝑜𝑖 16 𝑃 𝑌 = 10

𝑒 −16 1610
=
10!
1 week 4 plates
×4 ×4 = 0.0341
4 weeks 16 plates
 At a particular hospital, babies are born at a rate of 20 babies
Example per hour.
a) What is the probability that 11 babies are born in 30 mins?

Define: 𝑋 = 𝑛𝑜. 𝑜𝑓 𝑏𝑎𝑏𝑖𝑒𝑠 𝑏𝑜𝑟𝑛 𝑝𝑒𝑟 ℎ𝑜𝑢𝑟 𝑋~𝑃𝑜𝑖 20

Define: 𝑌 = 𝑛𝑜. 𝑜𝑓 𝑏𝑎𝑏𝑖𝑒𝑠 𝑏𝑜𝑟𝑛 𝑖𝑛 30 𝑚𝑖𝑛𝑠 𝑌~𝑃𝑜𝑖 10

60 mins 20 babies
÷2 ÷2
30 mins 10 babies 𝑒 −10 1011
𝑃 𝑌 = 11 =
11!
Example Cont. b) What is the probability that 8 babies are born in 15 mins?

Define: 𝑭 = 𝒏𝒐. 𝒐𝒇 𝒃𝒂𝒃𝒊𝒆𝒔 𝒃𝒐𝒓𝒏 𝒊𝒏 𝟏𝟓 𝒎𝒊𝒏𝒔 𝑭~𝑷𝒐𝒊 5

𝑿 = 𝑛𝑜. 𝑜𝑓 𝑏𝑎𝑏𝑖𝑒𝑠 𝑏𝑜𝑟𝑛 𝑝𝑒𝑟 ℎ𝑜𝑢𝑟 𝒀 = 𝑛𝑜. 𝑜𝑓 𝑏𝑎𝑏𝑖𝑒𝑠 𝑏𝑜𝑟𝑛 𝑖𝑛 30 𝑚𝑖𝑛𝑠

𝑋~𝑃𝑜𝑖 20 𝑌~𝑃𝑜𝑖 10
÷4 Can find the ÷2
60 mins 15 mins mean of F 30 mins 15 mins
using either
20 babies X or Y 20 babies
5 babies 5 babies
÷4 ÷2

𝑒 −5 58
𝑃 𝐹=8 =
8!
 Cumulative Poisson
Distribution Tables

𝑭 𝒙 =𝑷 𝑿≤𝒙

Must arrange all

situations so that
they deal with “≤”
 1st column: Top row:
Values that 𝑋 can take on. Some values of 𝜇 (𝜆)

Cumulative probabilities
Example Suppose 𝑋~𝑃𝑜𝑖 3.5 . Find 𝑃 𝑋 ≤ 6 .

Step 1: Find 𝜆 = 3.5 in the top row.

Step 2: Find 6 in the first column.

Step 3:
Line up the row
and column to
get

𝑃 𝑋≤6
= 0.9347 0.9347
Example Suppose 𝑋~𝑃𝑜𝑖 4.3 . Find 𝑃 2 ≤ 𝑋 ≤ 6 .

𝑃 2≤𝑋≤6
0.0719
=𝑃 𝑋≤6 − 𝑃 𝑋≤1

= 0.8558 − 0.0719

= 0.7839 0.8558
 Full In a particular port city there are, on average, 8 oil tankers in the
Example harbour per day.
What is the probability that there are 6 tankers in the harbour
tomorrow?

𝑋 = 𝑛𝑜. 𝑜𝑓 𝑡𝑎𝑛𝑘𝑒𝑟𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑜𝑟𝑡 𝑖𝑛 𝑎 𝑑𝑎𝑦

𝑋~𝑃𝑜𝑖 8
𝑒 −8 86
𝑒 −𝜇 𝜇𝑥 𝑃 𝑋=6 =
6!
𝑃 𝑋=𝑥 =
𝑥!
= 0.1221
 What is the probability that there are fewer than 3 tankers in the harbour
tomorrow?

𝑒 −𝜇 𝜇 𝑥
𝑃 𝑋=𝑥 =
𝑥!

0 1 2 3 4 5
𝑃 𝑋<3 =𝑃 𝑋≤2

= 𝑃 0 + 𝑃 1 + 𝑃(2)

𝑒 −8 80 𝑒 −8 81 𝑒 −8 82
= + +
0! 1! 2!

= 0.0138
 What is the probability that there will be 40 tankers in the port next week?

Define: 𝑋 = 𝑛𝑜. 𝑜𝑓 𝑡𝑎𝑛𝑘𝑒𝑟𝑠 𝑖𝑛 𝑡ℎ𝑒

𝑝𝑜𝑟𝑡 𝑖𝑛 𝑎 𝑑𝑎𝑦
𝑊 = 𝑛𝑜. 𝑜𝑓 𝑡𝑎𝑛𝑘𝑒𝑟𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑜𝑟𝑡 𝑖𝑛 𝑎 𝑤𝑒𝑒𝑘
𝑋~𝑃𝑜𝑖 8
𝑊~𝑃𝑜𝑖 56

𝑒 −56 5640
1 day 8 tankers 𝑃(𝑊 = 40) =
40!
×7 ×7
= 0.00496
7 days 56 tankers
 What is the probability that there will be 3 tankers in the port in half a day?

Define: 𝑋 = 𝑛𝑜. 𝑜𝑓 𝑡𝑎𝑛𝑘𝑒𝑟𝑠 𝑖𝑛 𝑡ℎ𝑒

𝑝𝑜𝑟𝑡 𝑖𝑛 𝑎 𝑑𝑎𝑦
𝐻 = 𝑛𝑜. 𝑜𝑓 𝑡𝑎𝑛𝑘𝑒𝑟𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑜𝑟𝑡 𝑖𝑛 ℎ𝑎𝑙𝑓 𝑎 𝑑𝑎𝑦 𝑋~𝑃𝑜𝑖 8

𝐻~𝑃𝑜𝑖 4

𝑒 −4 43
1 day 8 tankers 𝑃(𝐻 = 3) =
3!
÷2 ÷2
1 = 0.1954
day 4 tankers
2
What is the probability that there will be at least 2, but less than 7, tankers
in the port in half a day?

𝐻 = 𝑛𝑜. 𝑜𝑓 𝑡𝑎𝑛𝑘𝑒𝑟𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑝𝑜𝑟𝑡 𝑖𝑛 ℎ𝑎𝑙𝑓 𝑎 𝑑𝑎𝑦

𝑃 2≤𝐻<7 𝐻~𝑃𝑜𝑖 4

=𝑃 𝐻≤6 − 𝑃 𝐻≤1

x x ✔ ✔ ✔ ✔ ✔ x x

0 1 2 3 4 5 6 7 8
𝑃(2 ≤ 𝐻 < 7)

= 𝑃 𝐻 ≤ 6 −𝑃 𝐻 ≤ 1

(Part of the table has been removed)

= 0.8893 − 0.0916
0.0916
= 0.7977

0.8893
 Try it yourself

On average, 5 customers arrive at Nando’s takeaway every hour. The

arrivals are random and independent.

a) What is the probability that 2 to 5 customers arrive in half an hour?

b) If the server at the counter typically spends 3 minutes taking each

customer’s order, how long would you expect the server to spend taking
orders from customers in 4 hours?
 Poisson
approximation
of a Binomial rv 𝑛 ≥ 100 with 𝑛𝑝 ≤ 10
𝑛 ≥ 20 with 𝑝 ≤ 0.05
When n is large
and p is small

mean = 𝜇 = 𝑛𝑝
A Poisson rv can
be a good variance = 𝜎 2 = 𝑛𝑝
approximation of a
binomial rv
Example
The probability that a commercial airplane is involved in a fatal
accident is 0.000 0004. If a random sample of 9 000 000 commercial
airplanes is selected, what is the probability that at most 4 will be
involved in fatal accidents?

𝒑 𝒏
𝑋 = 𝑛𝑜. 𝑜𝑓 𝑎𝑖𝑟𝑝𝑙𝑎𝑛𝑒𝑠 𝑖𝑛𝑣𝑜𝑙𝑣𝑒𝑑 𝑃(𝑋 ≤ 4)
𝑖𝑛 𝑓𝑎𝑡𝑎𝑙 𝑎𝑐𝑐𝑖𝑑𝑒𝑛𝑡𝑠

𝑋~𝐵𝑖𝑛( 𝟗 𝟎𝟎𝟎 𝟎𝟎𝟎 ; 𝟎. 𝟎𝟎𝟎 𝟎𝟎𝟎 𝟒 )

 0.000 0004 = 𝑝 ≤ 0.05 𝜇=𝑛𝑝

9 000 000 = 𝑛 ≥ 20 = 9 000 000 × 0.000 000 4

= 3.6

So 𝑋 can be approximated by 𝑌~𝑃𝑜𝑖 𝟑. 𝟔

𝑃(𝑋 ≤ 4)

=𝑃 𝑌≤4

= 0.7064 0.7064
 POISSON DISTRIBUTION

= POISSON.DIST(𝑥; 𝜇; TRUE/FALSE)
FALSE:
𝒙 If it is not cumulative
The specific probability i.e.
value of the 𝑃 𝑋=𝑥
random 𝝁 TRUE:
variable 𝑋 The mean. If it is cumulative
probability i.e.
𝑃 𝑋≤𝑥
Example Power outages at UKZN can cause servers to crash. Based on the
last few years, it has been found that there are, on average, 9
server crashes per year.
What is the probability that UKZN will have 11 servers crash this coming year?
=POISSON.DIST(𝑥; 𝜇; TRUE/FALSE)
Define: 𝑌 = 𝑛𝑜. 𝑜𝑓 𝑠𝑒𝑟𝑣𝑒𝑟𝑠 𝑡ℎ𝑎𝑡 𝑐𝑟𝑎𝑠ℎ 𝑖𝑛 𝑎 𝑦𝑒𝑎𝑟
𝑌~𝑃𝑜𝑖 9
𝑃 𝑌 = 11

Use
“FALSE”
 What is the probability that UKZN will have more than 3 servers crash in the
next month?
Define: 𝑌 = 𝑛𝑜. 𝑜𝑓 𝑠𝑒𝑟𝑣𝑒𝑟𝑠 𝑡ℎ𝑎𝑡
𝑐𝑟𝑎𝑠ℎ 𝑖𝑛 𝑎 𝑦𝑒𝑎𝑟
𝑀 = 𝑛𝑜. 𝑜𝑓 𝑠𝑒𝑟𝑣𝑒𝑟𝑠 𝑡ℎ𝑎𝑡 𝑀~𝑃𝑜𝑖( 0.75 ) 𝑌~𝑃𝑜𝑖 9
𝑐𝑟𝑎𝑠ℎ 𝑖𝑛 1 𝑚𝑜𝑛𝑡ℎ

12 months 9 crashes
÷12 ÷12
1 month 0.75 crashes

𝑃 𝑀>3
= 1 − 𝑃(𝑀 ≤ 𝟑)
Use “TRUE”
What is the probability that UKZN will have fewer than 3 servers crash in the
next 4 months?

Define: 𝑌 = 𝑛𝑜. 𝑜𝑓 𝑠𝑒𝑟𝑣𝑒𝑟𝑠 𝑡ℎ𝑎𝑡

𝑐𝑟𝑎𝑠ℎ 𝑖𝑛 𝑎 𝑦𝑒𝑎𝑟
𝐹 = 𝑛𝑜. 𝑜𝑓 𝑠𝑒𝑟𝑣𝑒𝑟𝑠 𝑡ℎ𝑎𝑡
𝐹~𝑃𝑜𝑖( 𝟑 ) 𝑌~𝑃𝑜𝑖 9
𝑐𝑟𝑎𝑠ℎ 𝑖𝑛 4 𝑚𝑜𝑛𝑡ℎ𝑠

12 months 9 crashes
÷3 ÷3
4 months 3 crashes

𝑃 𝐹<3
= 𝑃 𝐹≤2
Use “TRUE”