Coupon Conspiracy

Heather Parsons
Dickens Nyabuti
Ben Speidel
Megan Silberhorn
Mark Osegard
Nick Thull
Introduction
Today we are going explain two
coupon collecting examples that
incorporate important statistical
elements. We will first begin by using
a well-known example of a coupon
collecting-type of game.
We are going to use the
McDonalds Monopoly Best
Chance Game to connect a real
world example with our Coupon
Collecting Examples.


McDonalds Scam
List of Prizes Allegedly Won
By Fraudulent Means

1996
Dodge Viper
(1) $100,000 Prize
(2) $1,000,000 Prizes
1997 (1) $1,000,000 Prize
1998 $200,000
1999 (3) $1,000,000 Prizes
2000 (3) $1,000,000 Prizes
2001 (3) $1,000,000 Prizes
The Odds of Winning

Instant Win 1 in 18,197
Collect & Win 1 in 2,475,041
Best Buy Gift Cards 1 in 55,000


McDonalds Monopoly
Important Statistical Elements
Expected Value gives the average behavior of
the Random Variable X
Denoted as:

Variance measures spread, variability, and
dispersion of X
Denoted as:

We will use these concepts in more detail in our
fair coupon collecting examples later in our
discussion.
| | { }
i i
x X P x X E = =

| | ( ) ] [ ) (
2
X E X E X Var =
Review
Sample space the set of all possible
outcomes of a random experiment (S)
Event any subset of the sample space
Probability is a set function defined on
the power set of S
So if ,
then P(A) = probability of A
S Ac
Review Continued
Bernoulli Random Variable
Toggles between 0 and 1 values
Where 1 represents a success and 0 a failure
Success probability = p
Probability of Failure = (1-p)
Bernoulli Trials
Experiments having two possible outcomes
Independent sequences of Bernoulli
RVs with the same success probability
E[X] = p Var(X) = p(1-p)
Review Continued
Geometric Random Variables
Performing Bernoulli Trials:
P=success probability
X=# of trials until 1
st
success
Assume
1 0 s < P
| |
( )
( )
2
1
1
p
p
X Var
P
X E

=
=
Expectations of Sums of Random
Variables
For single valued R.V.s:


Suppose ( i.e. g(x,y) = g )

For multiple R.V.s:


R R : g
2

y} Y x, P{X y) g(x, Y)] E[g(X,

x y
= = =

= =
i
i i
} x P{X x ] E[ X
Properties of Sums of Random
Variables

Expectations of Random Variables can be
summed.
Recall:

Corollary:


More generally:

E[Y] E[X] Y] E[X + = +
y} Y x, P{X y) g(x, Y)] E[g(X,
x y
= = =


= =
=
n
1 i
i
n
1 i
i
] E[X ] X E[
Sums of Random Variables

- Sums of Random Variables can be summed
up and kept in its own Random Variable

i.e.


Where Y is a R.V. and is the instance
of the Random Variable X

=
=
0
i
X Y
i
i
X
th
i
Variance & Covariance of
Random Variables
Variance a measure of spread and variability



Facts:
i)
ii)
Covariance
Measure of association between two r.v.s



] E[X]) E[(X (X)
2
var =
) (E[X] ] E[X (X)
2 2
var =
Var(X) a b) (aX
2
var = +
)(Y-E[Y])] E[(X-E[X] (X,Y) = cov
( ) ( ) ( ) ( ) X,Y Y X Y X cov 2 var var var + + = +
Properties of Covariance

Where X, Y, Z are random variables, C constant:






(X) (X,X) var cov =
(X,Z) (X,Y) Z) (X,Y cov cov cov + = +
(X,Y) c * (cX,Y) cov cov =
(Y,X) (X,Y) cov cov =
[X]E[Y] E[XY] E (X,Y) = cov
McDonalds

CouponConspiracy I
Suppose there are M different types of coupons, each
equally likely.

Let
X = number of coupons one needs to collect in order
to get the entire set of coupons.
Problem:
Find the expected value and variance of X
i.e E [X]
Var (X)
Solution:
Idea: Break X up into a sum of simpler random variables
Let
X
i
= number of coupons needed after i distinct types have been
collected until a new type has been obtained.
Note:

=
=
1
1
m
i
i
X
X
the X
i
are independent so
| | | |
( ) ( )

=
=
=
1
1
1
1
m
i
i
m
i
i
X
X
Var X Var
E X E
Observe:
If we already have i distinct types of coupons, then using Geometric
Random Variables,

P(next is new)
m
i m ) (
=
Regarding each coupon selection as a trial
X
i
= number of trials until the next success
and
P(next is new)

m
i m ) (
=
We know that the

| |
( )
) ( ) (
2 2
2
2
1
1
i m i m
m
p
X
X
mi
m
i p
Var
i m
m
p
E
i
i

=
|
|
|
.
|

\
|
|
.
|

\
|
=

= =
so
| | | |
( ) ( )
( )

=
|
|
|
.
|

\
|
= =
|
.
|

\
|

= =
1
0
2
1
0
1
0
1
0
m
i
m
i
i
m
i
m
i
i
i m
X
X
mi
Var X Var
i m
m
E X E
so
| |


=

=

=
|
.
|

\
|

=
1
0
1
0
1
m
i
m
i
i m
m
i m
m
X E
(

=
(

+ + + + =
(

+ + +

+ =

=
m
i
i
m
m
m
m m m
m
1
1
1
..........
3
1
2
1
1
above, expression the reversing
1
1
2
1
..........
2
1
1
1 1
Eulers Constant
| | m m
i
m X E
m
i
so
m
i
m
i
m
i
m
i
m
log
1
log
1
log
1
1
1
1
lim
~ |
.
|

\
|
=
~
|
.
|

\
|
=

=
=
=

( )
( )
( )

=
|
|
|
.
|

\
|
=
|
|
|
.
|

\
|
=
1
1
2
1
0
2

,
m
i
m
i
i m
i m
mi
mi
X Var
So
( )
slide next in the trick a use will we his, simplify t to

1
1
2

=
|
|
|
.
|

\
|
=
m
i
i m
i
m
( )
( )
( )
( )
( )
( )
( )
( ) i m i m
i m
i m i m
i m m
i m m
m m i i

=

=
+
=
2 2
2
2 2


sum. the of numerator
the from m g subtractin and to m adding : Trick
( ) ( ) ( ) i m i m i m
m i

=
1
2 2
( ) ( )
( ) ( )

=
|
|
|
.
|

\
|

|
|
|
.
|

\
|
|
|
|
.
|

\
|

|
|
|
.
|

\
|
1
1
1
1
2
2
1
1
1
1
2
1 1
1
m
i
m
i
m
i
m
i
i m i m
m
i m i m
m
m
m
m
Applying the trick from the previous slide to,
( )

=
|
|
|
.
|

\
|
1
1
2
m
i
i m
i
m we get,
pulling out the m in the first sum,
Expanding the sums from the previous slide
( ) ( ) ( ) ( ) ( ) ( )
(
(

+ + +
(
(

+ + + =
1 2 1 1 2 1
1
.....
1 1 1
.....
1 1
2 2 2
2
m m
m
m m
m
( ) ( ) ( ) ( ) ( ) ( )
(
(

+ + +
(
(

+ + +

=
1 2 1 1 2 1
1
......
1 1 1
......
1 1
above, sum the of order the reversing
2 2 2
2
m
m
m
m


=

=
=
1
1
1
1
2
2
1 1
m
i
m
i
i
m
i
m
By the Basel series we will
simplify this in the next slide.
Explaining the Basel Series
Basel Series
( )
m m
m m X Var
2
2
2
2
2
2 2 2
) log(
6
6
to converges Series Basel the m, large a for
6
......
1 1 1
3 2 1
~
|
|
.
|

\
|
~
= + + +
t
t
t
In Conclusion
So clearly the variance is:



and the expected value is:



These approximations are dependent on the fact that
m is a large number approaching infinity. Where m
is the number of different types of coupons.
( )
m
X Var
2
~
| | m m X E log ~
McDonalds

CouponConspiracy II
The 2
nd
Coupon Conspiracy
Given:
Sample of n coupons
m possible types
X := number of distinct types of coupons

Find:
E[X] (expected value of x)
Var(X) (variance of x)

Redefining X
Decompose X into a sum of Bernoulli indicators

1 if a type is present

0 otherwise
1
i
i
Let X
i m

s s
1
Note:
m
i
i
X X
=
=

1 2
(knowing one coupon type occured lowers
the opportunity for the other coupons types to occur)
Remark: , , , are dependent
m
X X X
| | | |
1
Note:
m
i
i
E X E X
=
=

Redefining Var(X)
( ) ( )
( )
1 1 1
Re : ,
,
m m m
i i i
i i i
call Var X Cov X X
Var X Var X Cov X X
= = =
=
| | | |
= =
| |
\ . \ .

( )
1 1
1 1
,
,
m m
i j
i j
m m
i j
i j
Cov X X
Cov X X
= =
= =
| |
=
|
\ .
=

(by covariance bilineararity)

Summary thus far
( )
( )
1 1
,
m m
i j
i j
Var X Cov X X
= =
=

| | | |
1
m
i
i
E X E X
=
=

Success Probability
Recall: is Bernoulli, with success Probabilitly P
i
X
{ 1} 1 { 0}
i i
P P X P X = = = =
m-1
[ ] 1
m
i
n
Thus E X
| |
=
|
\ .
th
Note: { 0} {type does not occur on j trial}
i
P X P i = =
1 1
1 { 0}
m-1
1
m
n n
i
j j
P P X
n
= =
= =
| |
=
|
\ .
[ [
Recap
m-1
[ ] 1
m
i
n
E X
| |
=
|
\ .
m-1 m-1
[ ] 1
m m
i
n n
Var X
(
| | | |
(
=
| |
( \ . \ .

1
1
, [ ] [ ]
m-1
1
m
m-1
1
m
m
i
i
m
i
So E X E X
n
n
m
=
=
=
(
| |
(
=
|
( \ .

(
| |
(
=
|
( \ .

Var(X)
( )
To calculate ( ), we need to find ,
i j
Var X Cov X X
( )
, [ ] [ ] [ ]
i j i j i j
Cov X X E X X E X E X =
In particular when i j =
( ) ( )
k
i j i j
A event that type is present
[ ] { 1}
A A 1 A A
i j i j
Let k
E X X P X X
c
P P
=
= =
= =
( )
i j i j
By De Morgan, A A A A
c
c c
=
( )
1 A A
i j
c c
=
( ) ( ) ( ) ( ) { }
Using the Inclusion/Exclusion rule:
P A B P A P B P A B = +
[ ]
i j
E X X
( ) ( ) ( )
1 A A A A
i j i j
c c c c
P P P
(
= +
(

[ ] cont.
i j
E X X
( ) ( ) ( )
[ ] 1 A A A A
i j i j
1 1 2
1
1 2
1 2
i j
c c c c
E X X P P P
n n n
m m m
m m m
n n
m m
m m
(
= +
(

(

| | | | | |
(
= +
| | |
( \ . \ . \ .

(

| | | |
(
=
| |
( \ . \ .

( ) ( )
( )
i j
i j
1 1
Re : A , A
2
A A
c c
c c
n n
m m
call P P
m m
n
m
P
m

| | | |
= =
| |
\ . \ .

| |
=
|
\ .
( )
i j
Cov X X
1
Re : [ ] [ ] 1
1 2
[ ] 1 2
i j
i j
n
m
call E X E X
m
n n
m m
E X X
m m
| |

| |
|
= =
|
|
\ .
\ .
(

| | | |
(
=
| |
( \ . \ .

2
1 1 2
( , ) = 1 2 1
2
1 2 1 1
1 2 1 2
2
2 1
i j
n n
n
n n n
m m m
Cov X X
m m m
n n
m m m m
m m m m
n
m m
m m
| | (

| | | | | |
|
(

| | |
|
( \ . \ . \ .
\ .

| | | | | | | |
= + +
| | | |
\ . \ . \ . \ .

| | | |
=
| |
\ . \ .
( )
Re : Var ( , )
i i j
call X Cov X X =
1 1
( )= ( , )
m m
i j
i j
Var X Cov X X
= =

Back to the Var(X)

1 1
1 2 2 2
1 3 2 3 3 3
1 4 2 4 3 4 4 4
1 2 3 4
( , ),
( , ), ( , ),
( , ), ( , ), ( , ),
( , ), ( , ), ( , ), ( , ),
( , ), ( , ), ( , ), ( , ), ( , )
m m m m m m
Cov X X
Cov X X Cov X X
Cov X X Cov X X Cov X X
Cov X X Cov X X Cov X X Cov X X
Cov X X Cov X X Cov X X Cov X X Cov X X
(
(
(
(
(
(
(
(
(

1 1
( , ) 2 ( , )
m m m
i i i j
i i j i
Cov X X Cov X X
= = <
= +

Var(X) in terms of Covariance
( )
1 1
( , ) 2 1 ( , )
m m
i i i j
i i
Cov X X m m Cov X X
= =
= +

1 1
( ) ( , ) 2 ( , )
m m m
i i i j
i i j i
Var X Cov X X Cov X X
= = <
= +

Simplifying the summation, We get:
( )
2
2
1 1 2 1
1 2 2
,
2
1 2 1
2
( ) ( 1)
n
n
n n n
m m m m
m m m
m m m m
Simplified
n n
m m m
Var X m m m m
m m m
| | | |

| | | | | | | |
| |
= +
| | | |
| |
\ . \ . \ . \ .
\ . \ .

| | | | | |
= +
| | |
\ . \ . \ .
Var(X) equals
1 1
Recall: ( ) = 1
n
i
n
m m
Var X m
m m
| |

| | | |
|

| |
|
\ . \ .
\ .
( )
1 1
( ) ( , ) 2 1 ( , )
m m
i i i j
i i
Var X Cov X X m m Cov X X
= =
= +

2
1 2 1
2
( ) ( 1)
n
n n
m m m
Var X m m m m
m m m

| | | | | |
= +
| | |
\ . \ . \ .
In Case you Forgot:
m-1
[ ] 1
m
n
E X m
(
| |
(
=
|
( \ .

[ ] log E X m m ~
2
( ) Var X m ~
Independent Case:
Dependent Case:
Analytic vs. Simulation
Approaches
For a simulation we will consider a fair set
of the McDonalds Coupons from Problem
I.
We will analytically find E[X] and Var(X)
and compare these two values with the
results from our computer simulation.
Analytic Approach
Given M = 10, compute
( )
m m
m m X Var
2
2
2
) log(
6
~
|
|
.
|

\
|
~
t
| | m m
i
m X E
m
i
log
1
1
~ |
.
|

\
|
=

=
References
Sheldon Ross
Author of Probability Models for Computer
Science, Academic Press 2002
2-3 pages from C++ books
McDonalds

Dr. Deckelman
Any Questions???