Notion of Equity

Notion of Equity

367 CHAPTER 30 N O TIO N 3 0.1 . OF E Q U IT Y IN TR OD UCT ION Th e n o tio n of - e q u ity portan ce . The sh ip m e n ts fro m vern...

631KB Sizes 4 Downloads 95 Views

367

CHAPTER 30 N O TIO N

3 0.1 .

OF

E Q U IT Y

IN TR OD UCT ION

Th e

n o tio n

of

-

e q u ity

portan ce .

The

sh ip m e n ts

fro m

verned

lo n g -te rm

lo ts .

by

D EFINITIO N

tra d e

of

a same

A c c o rd in g

to

producer

agreem ents th e s e

is

d e te rm in e d

in

In

com m ercial

s a m p lin g ,

te rm s

com ponent w hich of

th e

th e

co m m e rcia l

to

a same

p ro v id in g

agreem ents,

of

dry

we

ta ken

assume

w e ig h in g ,

o n ly

th is

in

ta re

re s id u a l

u n d o u b te d ly pe o f

is

in

co m m o d itie s

th e

w e ig h ts

can

in to

a

account

in

w here

a lw a ys

is

of

e x tre m e

c o n s is ts

of

s e rie s

consum er.

Such

fo r

hundreds

te n s ,

purchase

and

d e fin e

s a m p lin g

n e a rly

c ritic a l

th e

s h ip m e n ts

p ric e

"c ritic a l

i t

of

or each

co nte nts c o nte nt"

purchase

are

even

u s u a lly

lo t

p ric e

th e

of

or

of

s h ip m e n t on

sam ples.

p ro p o rtio n

th e

go­

th o u s a n d s

e v a lu a te d as

im ­ of

m a te ria l

of

a

o b je c t

co ntract.

We s h a l l gross

a ris e s m in e ra l

very

th e

fo llo w in g

w e ig h in g ,

errors

b e in g

o p tim is tic

book.

It

w ou ld

s e c tio n s

m o is tu re th e

s a m p lin g

sa m p lin g

h ypotheses

anyway

th a t

but

re q u ire

re le v a n t and

e rrors th e ir

m easurem ents

a n a ly s is ,

in

th e

a ss a yin g

are

sense.

These

w id e r

d is c u s s io n

a m u lti-v o lu m e

such

fa lls

te x tb o o k

o u ts id e s till

to

as e xact, are

th e

sco­

be w r i t ­

te n . We s h a l l in s ta n c e , chase ly

riu m -

-

iro n

in

such m a in

T h is

as,

e tc

p e n a liz e d

fo r

in

p e n a liz e d e tc

c ritic a l

com ponent,

a s in g le

te rm

on

ju s tifie d In

in s ta n c e component

a b a rite

v a lu a b le

by

th e

a g iv e n

fa c t

of

th a t

co m m od ity,

th e

m ain

th e

fo rm u la

a ll

te rm s

several

of

v a lu a b le

component

d e te rm in in g th e

th e

fo rm u la

com ponents

are

fo r

pur­

are

u s u a l­

u s u a lly

c ri­

: :

iro n

in

co nce n tra te ,

co m ponent(s)

:

an

iro n

e tc

ore,

z in c

in

a z in c

c o n c e n tra te ,

ba­

...

g o ld ,

s ilv e r,

in

a

lead

or

copper

con­

... im p u ritie s

c o n c e n tra te ,

ore,

is

a s in g le w ords

same w a y .

v a lu a b le

secondary

on

o th e r

th e

s u lp h a te

ce ntrate ,

rite

in

p ric e .

b u ilt

tic a l

re aso n

or

e tc

o ff-s iz e ...

:

a rs e n ic

or

z in c

in

a

lea d

co nc e n tra te ,

o v e rs iz e

or

u n d e rs iz e

s ilic a

in

a ba­

... m a te ria l

:

to

a g iv e n

mesh

in

an

368

We s h a l l

:

c a ll

p^n

:

th e

tru e

p^n

:

th e

s e ttle m e n t

te n t

m arket

a ^ n as

p ric e p ric e

of

a g iv e n

of

as

lo t

w ith

c a lc u la te d

a random e s t i m a t o r

of

a ^ .

c ritic a l

when

Th e

u s in g

p ric e

content

th e

p^n

a^n

sam ple

is

c ritic a l

a random

con­

e s tim a to r

of PLn:

VEn

th e

"re la tiv e

%

“ d

VEn

is

the

cu m u la tiv e

^

s e ttle m e n t

,E

:

e rror"

“ Sn -

.

"

:

p. y l tj

a random

(V

fo r

v a lu e ),

re la tiv e

d iffe re n c e

between

p Ln

'Ln

v a ria b le . tru e

m arket

p ric e

of

a s e rie s

of

N lo ts

L

= 1,

(n

n

2,

...

N)

ρίΣ =Σ ρίη n

ΡςΣ :

the

cu m u la tiv e

s e ttle m e n t

:

th e

cu m u la tiv e

of

th e

s e ttle m e n t vc

error

d e fin e d

PS Z - P LJ:

?

is

th e

v a lu e s , is

w e ig h te d

we

equal

s h a ll

to

th e

me an

a d m it

:

E q u ity

:

3 0.2.

:

as

1 =—

E

:

by

J

th e

s h ip m e n ts

p ra c tic a l

u s u a lly

purposes

have

th is

a<*n ,

v a lu a b le

at

Fo rm u la s

pressed

p ric e

le a s t

by

lin k in g

or

fro m

th e

ores p^n

p^n up

one

most

and

is

w ith in

com ponents,

c om m odity

me an

:

n is

s a id

to

d e fin itio n w hich

be of

"e q u ita b le " VE

n

:

in v o lv e s

when

th e

me an

of

th e

: m(Psn)

= p Ln

FUNCTION OF THE

CONTENT

m e ta l-b e a rin g

s e ttle m e n t

c o m p a ra b le

w e ig h te d

ve

proce d u re zero

m e an

n

N

As a ll

PR OPER TI ES OF THE S ETTLEMENT P R IC E ASSUMED TO BE A L I N E A R

W ith th e

is

n

fo r

m(VEn ) = 0

CRITICAL

te n t

VE

VEn ·

a rith m e tic a l

A s a m p lin g

v a ria b le

of

th a t

Σ

random

:

N lo ts

ξ ρ ίη

ve

D e fin itio n

of

PL n VEn

Σ ν Ε Σ

s e rie s

" Σ pSn

p Sz VE

p ric e

n e a rly

th e is

th e

a lw a y s

lim its

p o s itiv e . p ric e

co ntract

general

c o n ce n tra te s

to

of

fo rm u la ,

lin e a r

a g iv e n

W ith

p^n to th e

a

as w e l l

th e

as w i t h fu n c tio n

range

p e n a liz e d c ritic a l

fix e d

in d u s tria l

m in e ra ls ,

of

th e

c ritic a l

by

th e

co n tra ct.

com ponents,

p^n

co n te n t

may v a r y

next

but

as

a general

v a lid

fo r

a s in g le

a ^ ru le ,

component

is

th e y

:

con­ W ith

n e g a tiv e .

can

fro m be

one ex­

:

369

Psn

= ML n [ ( a Sn

ML n

:

dry w weight of tthe lot ei* 9 ^ lie

pQ

:

p ric e

of

'

one

p e n a liz e d by

W

to n

0

Pln

and

s m e ltin g

of

m e ta l,

a c ritic a l

a n > t>n ,

* 1 "

and

pure

m e ta l

m in e ra l

:

'

w ith

P2n

:

*-n '

co nte nt

p 2n

^

b n ) ( p On ’

or

a^

m in e ra l

(p o s itiv e )

(n e g a tiv e )

in

or

p e n a lty

a m a te ria l

per

to n

of

c h a ra c te riz e d

,

p ara m e te rs

p ro c e s s in g

or

m a te ria l

a c c o u n tin g

charges,

e tc

fo r

...

s m e ltin g

Some

of

or

p ro c e s s in g

lo s s e s ,

these

p a ra m e te rs

can

be

zero. W h ether th e

m e ta l

producer th is

is

by

th e

beyond

We c a n

PL n

th e se

th e

P in )

as

th is

procedure

is

th e

th e y

same

Pln >

.

,

( a Sn -

aLn}

s e c tio n

th e

e x a c tly



p ro p o s itio n are

w ork

u s u a lly in

a lto g e th e r im po se d

fa vo u r

of

:

upon

th e

in th e

buyer

but

fo rm u la

:

p 2n] I

,

(lin e a r

m(VEn )

= 0

th e

e q u ita b le very

re s u lt

N becomes

of

h y p o th e s is ),

reduces

e q u ita b le

fo r

he

has

what

bn ,

its e lf

p Qn a n d

p ln

to

:

is

u n b ia se d .

re m ain

s a m p lin g

is

of

as

o n ly it

i f

it

u s u a lly

does· w i t h

s ta tis tic s

th a t

a

VE

When

th e

lo n g -ra n g e te n d s

tow ards

: :

VΕ Σ

re c e iv e d ,

in te re s ts

and

la rg e ,

in fin ite

is

The

i f

m a th e m a tic a l

sa m p lin g

s h ip p e d . i f

is

becomes

a known

as

both

in e q u ita b le

te n d s

th e

s e lle r

p a rtie s :

tow ards has

have

VE^, t e n d s

m(VEn ) been

been

= 0.

p a id

e q u a lly

to w a rd s

m(VEn )

The

buyer

e x a c tly

fo r

has what

re sp e cte d .

φ

0.

T wo

cases

m ay

: m(VEn )

is

v a lu e .

Th e

in te re s ts

tru e th e

to

'

c o n d itio n

s h ip m e n ts

it

Then,

i f

has

-

in

e q u ity

ru le ,

th e y

book.

b n ) ( p 0n

ML n

a n o th e r

= a Ln

m(VEn ) . 1)

assumed

th e

s a m p lin g

co n tra ct,

th e

th is

is

in d u s trie s ,

a general

n p Ln

number N o f

tru e

* 1 ’

bn)
The

-

W

-

m( a Sn>

a ris e

"e q u ita b le "

a c c o rd in g

:

When,

2)

'

p^n

As of

we d e d u c e

co nsta nt,

p a id

scope

c a lc u la te

n =

are

m e ta llu rg ic a l

w hich

vr

he

and

consum er.

= ML n [ ( a L n

fro m

fo rm u la s

m in in g

p o s itiv e

is

v a lu e .

The

in te re s ts

sa m p lin g

of

m(VEn )

:

th e

s a m p lin g th e

s e ttle m e n t

in e q u ity

buyer

n e g a tiv e

of

th e

:

are th e

in

s e ttle m e n t

are

w orks

is

fa v o u r

s ig n ific a n tly

s y s te m a tic a lly

h ig h e r

th a n

th e

of

In

long

run,

th e

s e lle r.

th e

je o p a rd iz e d .

p ric e in

is

fa vo u r

s ig n ific a n tly

in e q u ity

s e lle r

w orks

p ric e

s y s te m a tic a lly of

th e

buyer.

je o p a rd iz e d .

lo w e r In

th e

th a n

th e

lon g

run,

370

As

fa r

as

sa m p lin g

a~

is

are

u n b ia s e d .

3 0.3.

its e lf

is

concerned

e q u ita b le ,

the

and

ina sm u ch

s e ttle m e n t

is

as

th e

e q u ita b le

fo rm u la

i f

and

lin k in g

o n ly

i f

PROPER TI ES OF THE SETTLEMENT P R I C E ASSUMED TO BE A N O N - L I N E A R THE C R I T I C A L

Fo rm u la s tra c t.

up

th e se

m e te rs.

An e x a m p l e

c e rta in

iro n

ore

p^n

to

sam ples

FUNCTION OF

CONTENT

lin k in g

O u tsid e

up

th e

p~

w ill

is

to

lim its ,

a^

are

th e

illu s tra te

p a id

lin e a r

fo rm u la th is

as

fo llo w s

o n ly

rem ains p o in t.

fo r

one

w ith in lin e a r

A c c o rd in g

to n

of

d ry

lim its

d e fin e d

b u t w ith to

a c e rta in

ore

:

per

1

in

d iffe re n t

th e

con-

para­

c o n tra ct,

a

-

N ormal

:

65

% Fe

> a^n >

62

% Fe

:

21

-

Low g r a d e

1

:

62

% Fe

> a^n >

61

%

Fe

:

p e n a lty

of

42

cents

per

1

% belo w

62

%

-

Low g r a d e

2

:

61

% Fe

> a^n >

60

% Fe

:

p e n a lty

of

70

cents

per

1

% below

61

%

T h is

13.2-

range

fo rm u la ,

p Sn

us

w hich

$ per

may seem c o m p l e x ,

is

cents

illu s tra te d

in

%

F ig .

Fe

30.1.

to n

-----------

13.1-

1 13.0-

A

I I1

12.9-

I I1

/ 1 12.8-

/ /

12.7-

/

I

1 1

1

\I

1

1

1 19 1C. £0 -

12.5-

.. -·-

■'''

J

/

G

______________

/

12.3-

/

12.2/ / 12.1-

11 9 -

1 1 Β

1 I

|

1

1

/

12.4- /

12.0-

/ S

/ 60

/

//

/

/

/ 1 1 1 I 1

1 // / 1 J 1/ 1 1

!

1

1

1

1

1

1

!

1 I 1 1 | 1 1 I 1

fX

λ

!

\\

V \i l

i1

1 1

'

/i

\

1 1 \

/

1

\

J w

\\

1 \

1 1 1

1 e>1

t32

a. Sn

---------- 1---------- 1---------- 1------% Fe 63

Flg. 30.1. Example of non-tineaA neZoutionshtp between the settlement pnlce of an Inon one and I t s c/UXlcal content Fe 1 .

371

L e t's the

suppose

broken

te n t

a^n

sae. mal

lin e ,

is ,

fo r

A s su m in g and

th a t

th e

apex

A fo r

th e

th e

sake

th e

both

c o n fid e n c e

re s p o n d in g

s e ttle m e n t

Due

change

to

ves, th e is

th e

th e

lo w e r

o rd in a te th e

m (pS n)

of

co u ld

and

a p p lie d

be

s h o u ld

o u r e x a m p le , be

th e

a lit tle

be

to th e

best

we

can

even

th e n

th e

w ill

ever

be

3 0.4 .

TE

η

:

:

in to

c a ll

as

is

by

fa c to r is

sa m p lin g

is

of

of

sam ple th e

p ra c tic a lly broken

th e

of

con­

a b s c is ­ nor­

lin e th e

co r­

o rd in a te s .

d is s y m m e tric a l The

th e

curve is

tru e

p ric e

h a l­ p^n

is

s e ttle m e n t

p ric e

w h ic h ,

to

due

its

:

< 0

of

is

In

b ia s

1.0 0 2

(p o in t

is

th a t

The

a ll

of

th e

a^n

is

d e p a rtu re

e xam ples

fig .

fa c to r

of

when

in

3 0 .1 .,

th is

c o n tra ct,

la rg e r

order near

of 62

th is

th a n

%.

1

%

0 .2 It

in

s h o u ld

B ).

concerned, and

in

an e q u i t a b l e

a c o rre c tiv e

th e

As

S in c e

lik e

buyer.

be

in e q u ita b le .

p ^ .

downwards

a correct

in e q u ita b le .

pS n ,

consequence

u s in g

%

61

of

one.

of

fre q u e n cy

s h o u ld

apexes

th e

th e

tw o

th e

As

near

of

The

th e

co nte nt

is

on

a x is

upper

m e an

th e

of

d is trib u tio n

th e made

th e

s e ttle m e n t

of

of

a x is

curve

arrow s

p ^ .

concave

a g a in s t

is

th e

G of

d e fin itio n

fa v o u r

a c h ie v e is

th e

The

fre q u e n c y

th e

one

can

we

th e re fo re

th in g s

seem v e r y

are,

do n o t h in g u nb ia sed

th e

about

s a m p lin g

chances

th a t

it

:

and

th is

p o in t

rem ote.

AND RANDOM ERROR I N

COMMERCIAL

S AM PL IN G

:

Tc

_ a Sn

'

error

c o m m itte d

on

lo t

L

n

.

By

d e fin itio n

:

aLn

a,L n

be

lo ts

of

th e

regarded

as

v a ria n c e c u m u la tiv e

T|-

asI

Σ As

than

th e

IMPORTANCE OF B I A S

s a m p lin g

th e

g ra v ity

a lo n g

a lo n g

than

but

m(VEn )

of

in

The

curve

la rg e r

c o n s id e ra tio n

to ta l

tio n a ry ς

to

th e

When

Τ Ε

fa r

is

ta ke n

" can

a^n

s e ttle m e n t

RELATIVE

We s h a l l

when

do

of

lin e

sam ple

th e

d o lla rs ),

s m a lle r

p ro v id e d

as

A,

w ith

d is trib u tio n

fre q u e n c y Fe.

re p re se n te d

u n b ia s e d ,

c o rre c tiv e

la rg e r

is

o b v io u s ly

w orks

th e

U n fo rtu n a te ly , th e

broken

a lw a y s

%

± 2 σ ( a^n ).

in v o lv e s is

th e

a ^ n = 62

p o in t

(1 3.02

a consequence

e q u ity

to

at

ce ntre

s a m p lin g

is

p o s s e s s io n

la c k

th e

w h ic h

th e

e q u ity

A

corresponds fre q u e n c y

u n b ia s e d , of

p Sn

n e c e s s a rily

< PL n

Though

our

of

is

be

in te rv a l

b e in g

p o in t

o rd in a te

d is s ym m e try ,

fro m

h a lf

of

to

s lo p e

a Ln The

c o n v e n ie n c e ,re p re s e n te d

sid e s

p ric e

of

co n te n t

in s ta n c e .

of

s a m p lin g

s y m m e tric a l,

d e lim it

tru e

" a

th e

s a me m a t e r i a l a random

are

v a ria b le

sa m ple d w ith

under

ro u tin e

a s ta tio n a ry

m e an

c o n d itio n s , m(TE)

and

TEn

a s ta ­

σ 2 (ΤΕ). s a m p lin g

e rro r

co m m itte d

on

th e

set

of

N lo ts

Ln

(n

= 1,

2,

aLZ

LZ

w e ig h ts

ML n

of

th e

lo ts

are

very

o fte n

c o m p a ra b le ,

we m ay w r i t e

:

..N )

372 1 m (TE_)

= m( T E )

and

σ 2(ΤΕ_ )

Σ

= - σ 2(ΤΕ) Ν

Σ

The

% p ro b a b ility

95

For

a s in g le

lo t

Ln

For

a set

N lo ts

c o n fid e n c e

in te rv a ls

: Τ Ε ρ

= m(TE)

± 2 σ( ΤΕ)

: TE

= m (TE)

± —

are

:

2

In

th e

m itte d nary

of

lin e a r

on

means

and

m(VEn )

= m(TE)

m( VE

= m( VE

τ

)

(see

lo t

L

on

3 0 .2 .)

a set

expressed

= m (TE)

% p ro b a b ility

95

s e c tio n

and

n

v a ria n c e s

)

η

The

range

a s in g le

σ( ΤΕ)

as

of

th e

s e ttle m e n ts

N lo ts

fo llo w s

are

a 2 (VEn )

= σ 2(ΤΕ)

and

a2 (V E J σ

1 = — a2 (V E J N η

in te rv a ls

For

a s in g le

lo t

:

VEn

= m (TE)

± 2 σ( ΤΕ)

For

a set

N lo ts

:

VE„

= m(TE)

± —

VEn a n d

v a ria b le s

V EZ c o m ­

w ith

s ta tio -

:

and

c o n fid e n c e

errors

random

are

1 =— σ 2(ΤΕ) N

th e re fo re

:

2 of

Expressed hand,

th o se

These very

re la tiv e TEn a n d

fo rm u la s

la rg e ,

standard the

in of

d e v ia tio n

b ia s ,

im p o rta n t

p o rta n t

can

become

3 0 .4 .1 .

R e la tiv e

s h a ll

random e r r o r s ,

present

buyer.

know n

to

We s h a l l rie n c e ) is the

assume th a t

σ 2(ΤΕ)

re a d e r must

e xam p le

s ix th a t

th e

and

s a me o r d e r

of

w ith

v a lu e s

s m a lle r w h ils t

(s e c tio n

VΕ Σ

on

th e

one

3 0 .4 .1 .)

and

o f N lik e ly

to

be

c h a ra c te riz e d

by

th e

th e

unchanged.

of

und e rsta n d

σ 2(ΤΕ).

u ran ium

th e lb s

te rm

in v o lv in g

We s h a l l on

show

a no the r

how

on one unim ­

th o se

order

to

of

100

lo ts

c a lc in e s ,

was

s ign e d

c o n ta in s

pound

In

co m m ercia l

an

average

(Decem ber

errors

1978).

of

of

c a lc in e d between

lo t

(h y p o th e s is

lo ts

have

is

v a ria n c e equal

L e t's

n il

th e

is

b ia s

m a g n itu d e ).

: me an

th a t

th is

zero

p o in t,

we

uran ium

o x id e ,

each

a s s o c ia te d

s ta tis tic a l

assume

now t h a t

a g iv e n

% U^Og.

93

Each

th e

s a m p lin g

w ith

illu s tra te

to

s a m p lin g

sa m p lin g

c o n c e n tra te s .

s h ip m e n t of

in

d is trib u te d

th e

VEn a n d

random e r r o r ,

tw in -s a m p le s

to ta l

th a t

of

id e n tic a l.

rem ains

random e r r o r s

v a ria n c e

US $ p e r of

of

th e

The m a t e r i a l

s e rie s

th e

and

m(TE)

b ia s

in te rv a ls

are

c o n tra cts,

s m a lle r

th e

c o n c e rn in g

be 4 3

te rm

in v o lv in g

me an

th e

hand

random e r r o r .

a b o u t 20,000

g iv e n

s in g le

by

th e

A co n tra ct c o n ta in in g

th e

u n im p o rta n c e

ch a ra c te riz e d

lo n g

te rm

be

c o n fid e n c e

o th e r

becomes

by

can

th e

th e

th e

on

th e

σ( Τ Ε ) ,

c h a ra c te riz e d

how

By

th a t

m ore,

e xam ple

are

v a lu e ,

TE^, o n

show

100 o r

σ( ΤΕ)

S

Σ

The to

p a rtie s

w ith

w e ig h ts th e

p ric e

be or

w e ll each

of

kept

in

U^Og

th e

a is

by

a

re serve .

su pp o rte d of

and

re p re s e n te d

(p h y s ic a l

tru e

s e lle r

by expe­

tw in -s a m p le s w e ig h ts

(unknown)

of

c ritic a l

373

co nte nt p^n

a^n o f

= 2 0,000

th e

to

02 ( a σ2 (ΤΕ)

u n b ia s e d , lo t

Ln

to

be

= --------------=

tru e

$ o r,

in

we

kn ow

(unknow n) round

th e

about ± 0 .4

95

%

% U^Og,

m arket

fig u re s

0 .8

p ro b a b ility fro m

w hich

p ric e

of

x

$

10^

is

c o n fid e n c e we

can

in te rv a l

e s tim a te

4.62 X

10

σ(ΤΕ)

and

= 2 .1 5

x

σ(ΤΕ) have

th e

we

can

= 0 .8

x

assumed

chances

com pute

:

106 x 2 .1 5

th e

are,

x

sa m p lin g

as w e l l

1 0 ' 3 = 1720

to

fo r

be

th e

$

c o rre c tly s e lle r

or

c a rrie d th e

out

buyer,

and

and

th e re fo re

fo r

a s in g le

:

%

to

g a in

or

lo o s e

more

5160

$

2 .1 5

%

to

g a in

or

lo o s e

between

2 and

3 σ(ρ^ η )

or

:

between

3440

and

5160

$

13.59

%

to

g a in

or

lo o s e

between

1 and

2 σ(ρ^ η )

or

: between

1720

and

3440

$

34.13

% to

g a in

or

lo o s e

between

0

1 σ(ρ^ η )

or

:

0 and

1720

$

T h is

c a lls

As of

tiv e

v a lu e ,

2)

th e

s o me

S ince

th e

fo r

lon g

5 ,000

dard

th e

s a m p lin g

1

The

pLz

at

now c o n s i d e r

1

= —

a(VE

a(VE

& to ta l

or

v a lu e

seem t o is

be

very

u n b ia s e d ,

: m ore

th a n

between

of

800,000

e x c e s s iv e . s m a ll

th e

maxim um p o s s i b l e

re pre se n ts

%

0.6 5

in

re la ­

in d e e d .

i t

is

th e re fo re

or

lo o s e

co n tra ct

of

100

to

$,

It

w in

lik e ly

an

e q u ita b le

w hich

a p p re c ia b le

means

th a t

p ro p o rtio n

of

sta k e .

of

w h o le the

cu m ula te d

1

= 2 .1 5

x

error

10

lo ts

(N

co m m itte d

= 100). on

th e

The

set

of

is

:

re la tiv e 100

lo ts

s ta n ­ is

:

-4

*

10

p ric e

of

= 80

a b s o lu te

is is

= ------ σ(ΤΕ)

)

and

co m m e rcia l not

th e

)

n

= 100 p Ln The

nobody

3 σ(ρ^ η )

:

p ro b a b ility

money

d e v ia tio n

σ ( VE )

th e

$ does

its

run,

th a n

re m arks

and

amount o f L e t's

tw o

com pared w it h

loss

in

:

10 J

0 .1 3

1)

:

(93)2

= σ (TE)

= PL n we

= 799,800

The

(0.2)2

)

a Ln As cr ( VEn )

S in ce

% ^O g.

e x p e rie n c e ,

co nte nt

= ----------—

σ ( ρ 5 η)

93

% U308

0.2

=

is

x 43

past

c ritic a l

a (aSn)

Ln

x 0.9 3

A c c o rd in g of

lo t

x

th e

100

lo ts

is

of

d e v ia tio n

of

th e

th e

order

of

:

1C)6 *

sta n d a rd

c u m u la tiv e

p ric e

p Lj;

1 σ ( pS z } Th e s in g le cept or

=

pL e

g(VE

re la tiv e lo t,

%of

as

= 100

s ta n d ard

w h ils t

3 σ(ρ^ Σ )

0.0 6 5

}

th e

th e

th e

p Ln

d e v ia tio n

a b s o lu te

maxim um

to ta l

Χ ^

σ( ΤΕ) is

te n

sta n d a rd

p o s s ib le

co m m e rcia l

lo s s

v a lu e

= 10 ö ( p S n ^ tim e s

s m a lle r

d e v ia tio n of of

e ith e r th e

= 1 7 ’ 200

is

fo r

te n

p a rty,

*

100

tim e s it

co n c e n tra te s

lo ts

th a n

la rg e r.

am ounts o b je c t

of

fo r

I f

to

a

we a c ­

5 1,600

th e

$

c o n tra ct.

374

Such

a maxim um

u n re a lis tic neous

m a te ria ls

a q u ite

b a b ility

fo r

th e w ho le

such

pass

d u stry

is

What

is

w ith

th e

is out

to

mes

b ia s

s a m p lin g

o fte n

succeed.

3 0.5 .

CONCLUSIONS

Who c a n th e of

a ffo rd

in tro d u c tio n a much m o r e

o n ly

co rrect p lin g

and a p t

th a n

u s u a lly but

in

to o

s a m p lin g

to

d e liv e r

te c h n ic a l

s m a ll

data

is

of

th e

la rg e

to

enough.

in

h is

n a tu re

ch ap ters

17,

to

s a me

th e are

g a in

th e

th e

very

le t's

is

o n ly

any

n ow

loss

w o u ld

% pro­

a 0 .1 3

F u rth e rm o re ,

amount o f

on

homoge­

assume

th e re

is

money s i n c e

ra n g in g fa c t

th a t

u n lik e ly or

lo o s e

s a m p lin g

fro m

is

1 to

th e

to

in ­

observed).

% as

0.0 6 5

2 % c o u ld

u ran ium

be

b ia s e d

com pared

and

i f

p o s s ib ility

your

and

fa v o u r.

The

tru th

th e

is

of

th e

tin

m a tte r

m ine

co m m ercia l A c c o rd in g

fa c ilitie s

be

can

perhaps

number o f fro m

in

can

in s p e c tio n

th e

th e lon g

d e te ct of

s till

th e s e

m e n tio n e d

to

re ga rd e d le s s

our as

in

in e v ita b le run.

th e se

a

i f

th e

sa m p lin g

and

c h e ckin g

w h e th e r

re s p e cte d ,

e x p e rie n c e

1 to

sam­

2 % b ia s

s ta tis tic a l

th e

number o f

p re p a ra tio n

rem ains

is

random e r r o r s

a s p e c ia lis t a c tu a lly

in

b ia s

a b s o lu te ly

A c a re fu l

b ia s e s

th e y

co m m e rcia l

in tro d u ce

a

s o m e ti­ th a t

s a m p lin g ,

random e r r o r s .

31)

are

lik e

YOU ? I n

d e trim e n ta l

27

th ro u g h

c o n d itio n s

sa m ple s,

by

money

sa m p lin g

d is tin g u is h e d

out

lo o s in g

o wn

A fa ir

chapter

of

th e ir

p ro d u c tio n

or

u n b ia s e d

s a m p lin g .

18 a nd

th e re

b ia s e s

b ia s e s

? Can

th a n

d e v ic e s

c a rrie d

but

exceeded.

e m p h a s iz e

2 % i f

to

p art

A c ritic a l

when

and

th e

th a t

ris k

im pose

9 % of

(see

g a in

fo u n d e d

are

Th e m aximum p o s s i b l e

s lig h tly

to

such

1 or

o wn

c le a rly

re p ro d u c ib le

v a lu e

not

e q u ita b le .

th e ir

s ig n ific a n tly re s u lts

m ent and m e th o d s , s ta te d

be

very

is

c o n ce n tra te s

s id e .

to

e ig h th

of

n e v e rth e le s s

a n a ly s is

in

tre a ch e ro u s

a m in o rity

lo o se

try

lo o s e th is

or

we s h o u l d

w rong

c o n c lu s io n

in te rv a l.

o cca sio n s

s e n s itiz e d

th e y

to to

be

one w he re

to

so

th a t

th e

to

Our

u ran ium

to ta l

a sym m e trica l

th e

s a m p lin g

to

is

chance)

v a rio u s

of

b ia s e d

th e

reached

(or

(th o u g h

on

are

be

o n ly

ris k

be

Some C o m p a n i e s

to

on

th e

s y s te m a tic

th a t

sa m p lin g

assumed

im p o rta n c e

happens

triv ia l.

happens

%of

0 .1 6

ris k

u nd e te cte d

th e

so

th e ir

a lo s s

p ro b a b ly

a b s o lu te ly i t

1 % U^Og c o n f i d e n c e

an e q u a l

p o in te d

e a s ily

:

$ or

procedure

N o w , we

is

th a t ±

129,000

h y p o th e s is

p a rty

and

im p ro b a b le

n o t exceed

by

ris k

h y p o th e s is

th e

th e

e q u ip ­

c o n d itio n s

best

b ia s

de­

te c to r. L arger

b ia s e s

e x p e rim e n ta l are

im p le m e n te d

p o i n t w hich

in

w ill

o p p o rtu n itie s te d w hich

are

lik e ly

m e thods

does

to

be

of

arouse fo r

accordance w ith d e a lt w ith

c ome

not

to

c h e ck in g

across

m ean

th a t

in

re m ain

c e rta in ch a p te r

b ia s e s th e y

s u s p ic io n b ia s

be

la te r

in e ffic ie n t,

n a tio n a l 32.

ra n g in g

can

sooner o r

or

but

in te rn a tio n a l

S till

la rg e r

b ia s e s

between

10 a n d

20

e a s ily

up

co rrected .

to

e s p e c ia lly

%)

(we are

5 %, when

th e y

S ta n d ard s, had

a

v a rio u s

e a s ily

dete c­

375

Our

c o n c lu s io n

concern b ia s you

sh o u ld

be

(ch a p te rs destroy

s ib le o n ly

you

tio n ,

e x tra c tio n

few

anyway

w ith

p ie c e s

try

and

money

or

be

a d v ic e

to

is

th e

be

no

lo n g

c u ra tiv e

co m m ercia l of

p o s s ib le

e ra d ic a tio n .

you.

and

If

There

money

c h a ra c te riz e d

reduced

s a m p lin g ,

s a m p lin g

when

by

b ia s e s ,

is

is

th e

fir s t

p re p a ra tio n

com prom ise w i t h

d e stro y

random e r r o r s

o n ly

le ft, th e

e ith e r

one

then

pos­ but

v a ria n c e

s u p p re s s in g

th e

σ 2( Τ Ε )

d e lim ita ­

b ia s e s .

p e rm is s io n ,

buyers

s p e n t on sources

run

s ig n ific a n tly

p re p a ra tio n

re a d e r's

to

p o s s ib le

There in

reduce

to

is

a ll

27). w ill

and

lik e ly

the

of

th e y

p re v e n tiv e

can

is

i f

18 a n d

or

:

w hich

Now,

th a t

e lim in a te

17,

them

s tra te g y th e n

is to

and

we w o u l d

s e lle rs

of

lik e

to

m in e ra l

c lo s e

th is

ch a p te r w ith

co m m od ities

a ll

over

a

th e

w o rld . 1)

Never

hammer a n d w ith

re ly shovel

a c o rre c t,

2)

on

W h e th e r

a n o n -p ro b a b ilis tic

m e th od .

u nb ia sed

s e lle r

v e rn in g

sa m ple s

are

c e rta in

th ro ug h

an

or

v a lu a tio n

fo rg e t

th a t

th e re

d e v ic e s

a n d we

fo re ig n

C o u n try

fo r

fe g u ard

a g a in s t

b ia s .

The 1) or

fa c t no

o n ly to

way

never

by to

of

th e

3)

to

agree

-

p lin g check

syste m s

of

a n a ly s is ,

the

sure

th a t

agreem ent p a rty ,

th e

s a m p lin g

your

th a t

is

a b ig

th a n

fo r you

a c c o rd in g u n le s s

in sta n ce w ou ld

th e

Company

are

of

rig h t

sw orn

to

you

th e

spend

not

u s in g to

be

w hich

are

fa c ilitie s

in te re s ts

i n c o r r e c t ways

your

and

expert

agreem ents

in te re s ts

s a m p lin g

and

th e

to

be

used

g o in g

to

be

co rrect

is

to

im pose

s a fe ly

p rotecte d

are

je o ­

s a m p lin g in

d e fin ite ly

try in g

go­

a b s o lu te ly

re p re se n te d ,

s a m p le r

are

fa c ilitie s ,

co rrected

w here y o u r

th e

th e

as

o th e r of

p a rty

one

of

e s tim a te d

d iffe re n c e

i f

rig h ts

as

e s tim a te d th e

c ritic a l

d iffe re n c e s sa m p lin g

its

a

no

sa­

c o n d i­

to

have

th e s e

is

:

d esig n ed

necessary,

are

on

not

s tric tly

between

e q u iv a le n t

to

between

:

o b ta in e d

o b ta in e d

between

e s tim a te d

gross

d iffe re n c e

fro m

a u n iq u e

p rim a ry

assays.

sa m ple s

a s s a y in g .

s y s te m a tic

m a g n itu d e s

tw in -s a m p le s

on

as

a c c e p ta b le

fo llo w in g

d iffe re n c e s

component

and

an

between

of

check

on

th e

:

assay

money

as

scheme.

o th e r

a c c re d ite d

p a rty

content

of

more

such

p a rty,

w ith

of

are

m ethod

e ith e r.

a s a m p lin g

content

c ritic a l

o th e r

th a t

th a t

an

an

th e

e v id e n c e by

own w e i g h i n g

s ig n

check

w e ig h t

th e

and y o u rs

c ritic a l :

in s ta n c e ,

make

o th e r

e s tim a te s

sam ple

to

have y o u r

th o s e

h is

th a t

s u ffic ie n t

sa fe g u a rd

e v a lu a te d 2)

have

and

much

sa m p lin g

s ig n

by

out

Never

is

never

o b ta in e d

la id

p a rd iz e d .

The

b uyer,

th o se

and

d e s ig n e d

3)

lo o s e

p ro b a b ilis tic

e x p e rt's

c o rre c tly

tio n s

You m i g h t

s a m p lin g

and

ta re

fro m

sa m ple s

by

both

and

tw o

p a rtie s

w e ig h ts ,

independent

sam­

assays. in d e p e n d e n tly

m o is tu re

s a m p lin g

:

and

376

-

s e ttle m e n t

c lu d in g

A number o f out

of

p a rtie s

both

to

our 4)

-

w he th e r th a t, (see

5)

case

o n ly

fo r

of

our

d ete ct

are

o n ly

m in o r

th e

s ta tis tic a l

d iffe re n c e and

v a lu e

i f

th e

s y s te m a tic

upon

s o me

or

reason

and

i f

the

a rb itra tio n

are

w hich

la rg e r

E Q UI T Y

U n lik e ly c le v e r

model

b o n n a ire , 814

-

of

th e

in

of

order

d is a rm

s m a lle r

between

th e

in ­

assays

la b o ra to rie s

d e fin ite ly

in s u ffi­

w h ic h ,

a c c o rd in g

d is c re p a n c ie s .

ch ap ter

than

both

is

31

to

d e c id e

on

a

e s tim a te s ,

s m a lle r

the

a rith m e tic w e ig h t

is

th e

la rg e r

of

at

th e

le a s t

com ple te

agreem ent

th a n

th e

m aximum a g r e e d

m aximum a g r e e d

m ean

of

c ritic a l

both

upon,

e s tim a te s

component

but

th e

w id e ly

p ric e

o f your

to

take

court

by

one

c o n tra c t w ritte n

s a tis fa c tio n be

reached

in te re s ts

re le v a n t

one

of

or

d e c id e as

th e

s e ttle m e n t

p ro d u c tio n

:

th e

or

th e

in

both

such

te rm s

p a rtie s . s e lle r

con­

expe rt

If

and ask

fo r

buyer, fo r

P a ris .

expenses

to

upon,

in d e p e n d e n t

je o p a rd iz e d ,

in

lo s s e s

case

an

between

are

Commerce

im p o rta n t

fro m

m aximum a g r e e d

of

Chamber o f

much m o re

th e

out

can

th a t your

c o s tly

lia b le

than

a rb itra tio n

c e rta in

v a rie s

are

must

th e

be

cost

n e xt,

is

co nsu m p tio n

c o n s id e re d of

such

not

a

lik e ly

whereas

m uc h

p la c e .

L O U I S - L E - D E B O N N A I R E 1S S P L I T T I N G METHOD i t

may a p p e a r ,

s p littin g

T h is a

la rg e

h is

draw

m a le the th e

e q u ity

son

was

co n c e iv e d

re s tle s s

between it

th e

We k n o w

a g a in s t

1 % of

b ia s e s

an

w ith

may seem

o b v io u s ly

840.

c o v e rin g

is

In te rn a tio n a l

p o lic y

of

L o u is

to

is

by

d is c re p a n c ie s

s a m p lin g

d e s c rib e d

e x is ts ,

co nte nt,

a m ic a b le

p ra c tic a lly

though

re ig n ,

to

no

C h a rle m a g n e 's to

fo r

p a rtie s .

o th e r

th a n

e m pire

L o th a ire

d iffe re n c e sam ple

a proce d u re

a ss a y in g

e x is ts

adopt

d iffe re n c e

fir s t

a few y e a r s

a procedure

im p o rta n t

3 0.6 .

from

both

in s u ra n c e

c o n tro l,

fa c to rs

ma y b e .

p ro v id e by

to

c ritic a l

case

fo r

be

the

th e

and w o r k i n g

more

of

as

s h o u ld

an

re le v a n t

3 above).

s y s te m a tic

S uch

m e th od

d iffe re n c e

d iffe re n c e

you

p rim a ry

com pared w i t h

d iffe re n c e

a s y s te m a tic

the

agreed

Such lo t

a ll

im p u ritie s .

a c c e p ta b le

a same

lis t) . by

between

p e n a liz e d

a c e rta in

as

the

6)

and

fro m

lo t

i f

tra c t

d iffe re n c e s

o b ta in e d

s p lit

p ric e

to

th e

p ro v id e

a s y s te m a tic

i f

p o in t

g o v e rn in g

as

can

of

com ponents

:

upon

(firs t

i t

im ple m en t

b a sis

check

tw in -s a m p le s

e x p e rie n c e , to

sound

to

on

because

:

v a lu a b le

c o n tra cts

c a rrie d

c ie n t

p ric e

secondary

and

of

je a lo u s y

n in th

more

h o ly ,

:

of

ce ntu ry

a c c u ra te ly

ru le d

to d a y 's

p ro je c t

o ffs p rin g s

lin e

or

h e ir,

n e ith e r

p art th e

the

over

nor

roman

w estern

and m i s t r u s t s p lit

sons,

e m p ire ,

p ro v id e s

"o v e re q u ity ".

Towards

and th e

"H o ly

th e

between

a very

L o u is -le -D e Roman

n e v e rth e le s s

e m pire

e ld e s t,

h is

th e

h is to ry

s o -c a lle d

b u t was

h is

th e

of

of

Europe.

s p littin g

L o th a ire ,

t h a t w ou ld

th e

French

end two

E m p ire "

a m ig h ty of of

h is th e

C h a rle s -le -C h a u v e . o ld

kn ow ing

Em peror th a t

most In

in v ite d

C ha rles

w ou ld

377

th e n

have

It

does

m e rc ia l of

tia l

of

not

tin g

th e

th e re

fo r

s tre tc h

and

th e

from

th e

a tte m p ts rem ains

a 50

in

th e

b ia s

in

L o u is -le -D e b o n n a ire 's be

ment o f

th e

in

the

fro m

a m oral

tra d e

of

th e

transpose p a rtie s

th is

to

th e

model

to

com­

c o n tra ct,

each

s p littin g

o p e ra tio n s

th e

p riv ile g e

of

in

c h o o s in g

th e

th e

p re ­

poten­

sa m p le . th e

:

fo r

approach, and

to

th e

idea

suppose

sam ple

e x p e rie n c e d

ch ea t w h ic h ,

a b o u t m o ra ls

have

24,

% p ro b a b ility

to

On e o f

perform

but

th e

p a rtie s

cares

im a g in a tio n

a c tu a l

ch ap ter

w hereas

e q u a lly

p o rtio n .

w ou ld

s e le c tio n

to

of

w o u ld

la tte r

become out

h is

co m m o d itie s.

in s ta n c e ,

p o in te d

o p e ra tio n

c h o o s in g

a huge

t h a t w o u ld

a lre a d y

of

o f m in e ra l

o th e r

o p e ra to r

ours,

need

tu rn s

sam ple

As

p riv ile g e

s a m p lin g

them by

sence

th e

th e

in th e

now t h a t

to

fo r

se pa ra te

some

a p ro b a b ilis tic b ia s

a ssum ing

c le v e r,

th e

s ta n d p o in t,

m ine ra l

rem ains

to th e

b ia s is

w ork

re aso n

in

fa v o u r

?

o th e r

such of

re p re s e n ta tiv e s o n ly

s p lit­

or

approach

can w o rk

h ig h ly

co m m o d itie s

th e

to

as

th e of

both

th e

s a tis fa c to ry .

cheat

d e tri­

B u t who