Galloway (A Treatise On Probability - 7ed Encyclopaedia Britannica) (1839) PDF

GIFT OP
JULIUS WAHOENHEIM
TREATISE
ON
PROBABILITY:
FORMING
THE ARTICLE UNDER THAT HEAD IN THE SEVENTH EDITION OF
THE ENCYCLOPEDIA BRITANNICA.
BY
THOMAS GALLOWAY,
M.A. F.R.S.
SECRETARY OF THE ROYAL ASTRONOMICAL SOCIETY.
EDINBURGH:
ADAM AND CHARLES BLACK, NORTH
BRIDGE,
BOOKSELLERS TO HER MAJESTY FOR SCOTLAND.

1839.
EDINBURGH:
BALFOIR
JACK, Printers, Niddry
Street.
CONTENTS.
INTRODUCTION
Page
SECTION
I.
GENERAL PRINCIPLES OF THE THEORY OF PROBABILITY.
of
Measure of mathematical
5, 6.
events.
compound
an event which
8.
Example
may happen
in
P. 16.
Chance and Proba
1-4. Definition of the terms
ARTICLES.
bility.
1.
7.
probability.
numbers.
in different ways.
9.
10.
Probability
Probability of
Identity of the
formulae for simultaneous and successive events.
SECTION
II.
OF THE PROBABILITY OF EVENTS DEPENDING ON A REPETITION OF

TRIALS, OR
COMPOUNDED OF ANY NUMBER OF SIMPLE EVENTS, THE
CHANCES IN RESPECT OF WHICH ARE KNOWN A PRIORI AND CON

STANT
P. 27.
ARTICLES.
may happen
theorem.
case of any
11, 12, Probabilities of the different
in a series of trials.
13.
Example in a particular
number of simple events.
14.
combinations that
Application of the binomial

case.
15.
1621.
Artifice for abbreviating the calculation.
G00735
Extension to the
Examples.
24, 25.
22, 23.
Determination of
CONTENTS.
IV
the
number of
trials
required to render the probability of an event
equal to an assigned fraction.
SECTION
III.
OF THE PHOB ABILITY OF EVENTS DEPENDING ON A REPETITION OF

TRIALS, OR
COMPOUNDED OF ANY NUMBER OF SIMPLE EVENTS, THE

PRIORI, AND VARY
CHANCES IN RESPECT OF WHICH ARE KNOWN A

IN
THE DIFFERENT TRIALS
ARTICLES.
P. 51.
26. Expression for the probability of the different pos
sible results in particular cases.

case.
29.
Examples.
30.
27, 28.
Extension to the general
Solution of a
question
proposed by
Huygens.
SECTION
IV.
OF MATHEMATICAL AND MORAL EXPECTATION
ARTICLES.
31. Definition of the term

expectation. 32.
of mathematical expectation.
plicable in particular cases.
Formula
P. 59.
to express the value of a moral

expectation.
sequences of the formulae.
Example
33. Mathematical expectation not

ap
34. Hypothesis of Bernoulli.
35, 36.
40.
3739. Con
Application of the theory of Ber
noulli to the subject of insurances.
41.
SECTION
Petersburg problem.
V.
OF THE PROBABILITY OF FUTURE

EVENTS DEDUCED FROM EXPERIENCE
P. 75.
ARTICLES.
42. Hypotheses
respecting the causes of an event. 43.
ermination of the probabilities of the
different hypotheses.
44.
bility of a future event deduced from the
probabilities of the
>eses.
45. Extension of the formula to
finite
causes or hypotheses.
46.
Sense
in
number of
any
which the term cause is used in
CONTENTS.
Examples of the
47, 48.
this theory.
the formulae to physical or moral events.

formulae to the case of an infinite
50.
Application of
Extension of the
51, 52.
number of
SECTION
49. Case in which
formulae.
are not equally probable.
the different causes
different causes.
VI.
HUMAN
OF BENEFITS DEPENDING ON THE PROBABLE DURATION OF

LIFE
P. 90.
ARTICLES.
puted.
single
53.
54, 55.
life.
56.
Principles on which the probability of life
Method of computing
Of
nuities on joint lives.
59
life
58.
annuity for terms of years.
On
the survivor of any
57.
SECTION
Of an
number of
Methods of computing the values of assurances on
63.
com
is
the value of an annuity on a
lires.
lives.
VII.
OF THE APPLICATION OF THE THEORY OF PROBABILITY TO TESTIMONY,
AND TO THE DECISIONS OF JURIES AND TRIBUNALS
P. 100.
64, 65. Expression for the probability of an event at

by a single witness on an assumed hypothesis. 66. Case in
67. Case in
which the event attested is extremely improbable.
ARTICLES.
tested
which the character of the witness
is
altogether
unknown.
68.
Ex.
pression of the probability of an attested event, regard being had

to the a priori probability of the
event.
events attested by several witnesses.

conflicting testimony.
76.
72
75.
69, 70.
Probability of
71. Formulae for the case of
Successive testimony, or tradition.
Application to the verdicts of juries.
77
81.
Probability of
acquittal and condemnation under different hypotheses.
Probability of a verdict being correct

majority.
84.
when pronounced by
a giren
Numerical expression for the error of a verdict when
arbitrary values are given to the constants. 85,86.

stants,
82, 83.
Values of the con
deduced from the records of the criminal courts in France.
CONTENTS.
VI
SECTION
VIII.
OF THE SOLUTIONS OF QUESTIONS INVOLVING LARGE NUMBERS. P. 128.
ARTICLES
Form of the
87-89.
theorem
the
for
approximating to
92.
formula.
series
of
90.
trials.
numerical value.
Stirling s
Example of
91.
a large
Probability of any assigned result of
93.
trials.
its
of
expression for the probability
number of
the most probable event in a large
Example of
the rapid diminution of the terms
of the binomial towards the beginning arid end of the series

the exponent
is
94.
large.
rences of an event falling within given limits.
of the limits for an assigned probability.
96.
95. Determination
Method of computing
the value of the integral which expresses the probability.

neral consequences of the formulas.
99.
98.
Method of
100
Application to numerical examples.
physical cause.
105.
To
Ge
Extension to
103.
Applica
the determination of a
determine the limits of loss or gain in a
multitude of pecuniary risks.

the ultimate results
To
104.
97.
using the table.
102.
the case of probabilities deduced from experience.

tion to a question of mortality.
when
of the number of occur
Probability
of play.
106.
General deductions relative to
108, 109.
Digression relative to a
question respecting the duration of play.
SECTION
OF THE MOST PROBABLE
IX.
MEAN RESULTS OF NUMEROUS DISCORDANT
OBSERVATIONS, AND THE LIMITS OF PROBABLE ERROR
P. 164.
ARTICLES.
110. Object of the
111.
application of the theory.
Statement of the problem.
112123. Solution of the problem,
and expression for the probability of the
of a
average
of observations falling within
124.
assigned limits.
large
number
Connection of
the special constant to which the

average continually approaches,
with the centre of
125126. Cases in
gravity of a curve line.
which the number of possible values of the

phenomenon
is
limited.
CONTENTS.
Vll
127130. Examples when the law of the probability

131
132. Form of the function which
is
be known.
supposed to
expresses the
law of probability, on the hypotheses of positive and negative errors

being equally probable.
133.
Illustration
by means of a curve.
134. Determination of the constants on this hypothesis.
135142.
Investigation of an expression for the constant of the law of facility,

in terms of the results of observations already
ent expressions for the limits.
mining the relative values of
tables.
precision, weight, probable error,

servations.
149.
made.
143.
Use of the formulae
144.
145
Differ
in deter
148. Expressions for the
and mean error of the
results of ob
Determination of the most probable value of a
function of several quantities, separately determined by observation,
with the precision and weight of the determination.
SECTION
X.
OF THE METHOD OF LEAST SQUARES

ARTICLES.
P. 200.
151-154. Method of forming equations of condition.
155-157. Determination of the most probable values of the
ele
ments from the equations of condition, and the probable amount of

the remaining errors.
158. Agreement of this solution with the
method of
ment.
least squares.
160.
Example
159.
Example
in the case
in the case of a single ele
of sereral elements.
substitute the following List of

requested to
that at the end of the volume.
Errata
The Reader
is
for
ERRATA.
Page
40, note, for Essai sur
les
Probabilites, read Theorie Analytique
des Probabilitcs.
read adapt.
43, line 7 from foot, for adopt
In the note, read Traite Elewith.
49, line 13, read beginning
mentaire.
54, line 7, for a
81,
first line, for
89, line 9, for
TO,
m, read
TO
135, last line, for
188
line 8, for
read a
(1+2+3..
m
+ ),
.
m+m
(n+ar)-+^-J,
e~\
read
read
(1+2+3.. .+n).
read
(+*)-"-*-
-**.
193, line 8, delete products of the.

198, line 9, for e e", read e e".
,
203, line 19, for kc

line
209, last
3 from
read kc.
foot, for (1), read (2).
line, for les
read des.
210, line 2 from foot, for 1833, 1834, and 1835, read 1834,
1835, and 1836.
211, last
line, for
Abhandhmgen, read Nachrichten.
PREFACE.
IN drawing up the following treatise for the Ency

clopaedia Britannica, my design has been, to pre
sent a general view of the principles, applications,
and more important

theory
results
of Probability,
of the
as laid
down
mathematical
in the
best
and most recent works on the subject

particu
those
of
and
Poisson
and, without
larly
Laplace
;
entering into the details of mathematical
difficul
ties, to explain the methods of applying analysis to

the solution of the principal questions, as fully as
the limits of the space which could be appropriated

to the article would permit.
In the prosecution of this design, questions con

nected with lotteries and games of chance, the sub
to which the
earlier writers on Probability

confined
themselves, and which are frequent
chiefly
ly supposed to form a principal part, if not the whole
jects
of the science, occupy but a small portion of the work ;

indeed they are only introduced as furnishing exam-
PREFACE.
The
pies to illustrate the principles or the formulae.
expectation has been developed

at some length, because it admits of important
The subject of
applications in the affairs of life.
theory of moral
mortality,
and the computation of annuities and
other pecuniary interests depending on

(the
gencies,
common
most
contin
life
application
of
the
theory,) having been treated in detail under other
heads in the Encyclopedia, scarcely came within the

scope of the article ; but the method of applying the
principles of the science to the data furnished
by the
mortality tables, has been explained, and the formulae

of most frequent practical use have been deduced.
The
section
on the probability of testimony and the
verdicts of juries,
is
more expanded, and contains
formulae which resolve the principal questions the
having
subject presents,
causes
of uncertainty.
regard to the different

In the following section,
formulae are given for obtaining approximate values

of expressions which involve very large numbers.
Such
more
occur
expressions
applications
interesting
and give
chances,
mathematical
in
all
of
higher and
the doctrine of
the
rise to the greater
part of the
which attend the subject.

The probability of the result of a large number
of
difficulties
experiments
or
within given limits,
observations being
is
contained
expressed by a definite inte
which occurs in various other applications of

in order to facilitate the
analysis
computation of
gral,
such probabilities, a table of the values of this inte-
PREFACE.
gral
is
XI
given at the end, by which, and by a simple

we are enabled to solve a mul
arithmetical operation,
titude of questions of the most interesting
and other
In the investigation of the most

of a quantity, to be deter
mean
value
probable
mined in magnitude or position, from a series of ob
wise
difficult kind.
servations liable to error,
and the determination of
the limits of probable uncertainty, I have followed the

very general and elegant analysis of Poisson. The last
section contains the application of the general theory
to
determination of
the
method
the most
advantageous
and an
of combining equations of condition,
explanation of the celebrated rule of least squares,

is shewn to coincide with that
pointed out by
which
the theory as giving the most probable values of the

elements involved.
The principal application which has hitherto been

made of the theory explained in the last two sec
has been to questions of astronomy ; but
may be applied, with equal advantage, to every
branch of physical inquiry, or statistical research,
tions,
it
where the object

rate
mean
results
is
to deduce
precise
and accu
from multiplied observations, or
to detect the existence of general laws, or the opera

tion of constant causes, among
large masses of phe
nomena.
Although considerable pains have been taken to

treat the subject in an elementary manner, it was im
possible, in
some
sections, to avoid
places, particularly the three last
discussions which belong to the
PREFACE.
more
abstruse
parts
higher and more
of the
useful
mathematics.
parts of the
admit of being treated in
not
unless the reader
is
prepared to
The
theory, do
any other way,

abandon the ac
curacy of exact science, and to accept of asser
have endeavoured, however, to

place the demonstrations in their clearest light, and
to give the expressions in their simplest forms ; and
tion for proof.
nothing in the present work be

yond the reach of those who have a moderate know
ledge of the differential and integral calculus, and
it is
hoped there
is
are acquainted with the
first
principles of the theory
of finite differences.
It
may be
proper, perhaps, to give a reason for de
viating from the usual notation, by writing fractional

This
expressions in the form of ratios or quotients.
practice
space
article
was adopted
in printing,
was
purpose of economising
was perceived that the
for the
when
it
likely to exceed the prescribed limits.
may be
considered objectionable on the ground of

its being less familiar to the reader ; but as it is here
used it can hardly occasion any real embarrassment.
It
LONDON,
February
12, 1639.
PROBABILITY.
THE
is an extensive and very im

of
mathematical
branch
science, the object of which
portant
is to reduce to calculation the reasons which we have for
doctrine of probability
believing or expecting any contingent event, or for assent
ing to any conclusion which

it is
is
not necessarily true.
considered that the whole edifice of
human
When
science, with
the exception of a few self-evident truths, such as the axioms

is nothing more than an assemblage of propo
which can only be pronounced to be more or less pro
bable, the importance of a calculus which enables us to ap
of geometry,
sitions
preciate exactly the degree of probability existing in each
be readily understood.
Our reasons for judging an event
case, will
to
be probable or im
;
first, an
probable, are derived from two distinct sources
a priori knowledge of the causes or circumstances which

determine
are
its
occurrence
and, secondly,
when
the causes
unknown, experience of what has already happened in
the same circumstances, or in circumstances apparently

milar.
si
Suppose, for example, a hundred white balls to be
placed in an urn along with
fifty
black balls, and that a per
draw a
son, blindfold, proceeds to
ball,
there
is
who
to us,
are acquainted with the contents of the urn, a determinate

probability that the ball which
balls
being supposed
to
be
is
drawn
will
the
is
same chance of drawing any one
ing any other

balls for
same circum
we assume
stances with respect to facility of drawing,
there
The
be white.
all in precisely the
that
ball as of draw
and, consequently, since there are two white
each black
ball,
and therefore two chances of draw
colour for each chance of drawing one
ing one of the
first
of the second,
we conclude
drawing of a white
ball to
the event which consists in the
be twice
as probable as the
op
In this case
posite event, or the drawing of a black ball.
our knowledge of the contents of the urn enables us to
judge of the probable result of the drawing.

ever, that antecedently to the drawing,
Suppose,
we were
how
entirely ig
norant of the contents of the urn, but that after a great
number of
ways
trials
have been made, (the ball drawn being al

urn after each trial, in order that the
replaced in the
circumstances
may be
the same in
all
the
trials) it
has been
observed that a white ball has been drawn twice as often

as a black ball,
many
as
we presume
white as black
balls,
that the urn contains twice as
and consequently
affords twice
many chances of drawing a white ball as of drawing a
presumption becomes stronger in propor

tion to the number of instances included in the observation.
black
In
and
this case
this
experience makes up for the want of a priori
knowledge, and
affords a
measure of the probability of the
result of a future trial.

It is
all
only in a comparatively small
number of
the possible ways in which an event
known a priori, and
cases that
may happen
are
in which, consequently, the ratio of the
HISTORY OF THE SCIENCE.
number of chances favouring the event
to the
ber of existing chances
In
is
questions of this class to
determinate.
whole
fact,
which the calculus can be
are connected with lotteries
num
most of the
and games of hazard.
applied,
The
re
obtained from the analysis of such questions cannot be
sults
considered as being of any great value in themselves, but
they frequently throw light on subjects of far higher impor

tance which present analogous combinations.
It is true that
the mathematical theory comes in aid of moral considerations,
and demonstrates the ruinous tendency of gambling even
when
the conditions of the play are equal, mathematically
speaking
but, unfortunately, those
who
indulge a passion
seldom capable of appreciating the force of

such arguments. The principal advantage which has resulted
for this vice are
from the application of analysis to games of chance is the ex

tension and improvement of the calculus to which it has led.
The calculation of the probabilities of events, the chances

of which are not known a priori, but inferred from experi
ence,
is
founded on the presumed constancy of the laws of
nature, in obedience to which events depending on constant
though unknown causes, are always reproduced in the same

when considered in large numbers. Among the vari
order
ous phenomena of the physical and moral world, nothing is

more remarkable than the constancy which is observed to
prevail in the recurrence of events of the
If
same
kind.
The
of male to female births furnishes a noted instance.
ratio
we
consider only a small
number of births, nothing can be
but taking a very large num

ber, as those of a whole kingdom in the course of a year,
the proportion of males to females is found to be almost inva
more uncertain than the result
riable,
man
and nearly
life
as 21 to 20.
affords another familiar
The mean
example.
duration of hu
Notwithstand-
PROBABILITY.
of con
ing the proverbial uncertainty of life, the differences
are
mankind
to
which
accidents
various
the
and
stitutions,
num
exposed, the average duration of the lives of a large
ber of individuals living in the same country
is
always found
be very nearly the same, insomuch that pecuniary risks

depending on it, if undertaken in sufficiently large numbers,
to
are
the least uncertain of all commercial speculations.
among
A similar
constancy is remarked in the results of statistical in

of crimes of the same spe
quiries of every kind. The number
cies committed in a year, the ratio of the number of acquit
tals to
the
number of
ships lost in
trials,
the
number of conflagrations, of
a particular trade, of letters which pass through the
post-office, of patients admitted into the public hospitals ;
in
ev
ery case the numbers in a given time are observed to fluctuate
between very narrow
limits, and to
approach nearer and near
more extended, to fixed mean values.

This constant approximation to fixed ratios, which is
er, as the observation
proved by
same
to
experience, in the recurrence of events of the
kind, enables us to apply the calculus of probabilities
many
social
all
is
of the most interesting questions connected with our
and
political institutions
and
to
determine the aver
age result of a series of coming events with as much precision

as if their chances were determinate, and known a priori, like
that of obtaining a given point with the throw of a die.
What
ever be the nature of the phenomenon under consideration,
whether
it
belong to the physical or moral order of things,
the calculus
is
equally applicable
when
the requisite data
have been determined from experience.

The foundations of the mathematical theory of probabili
ties were laid by Pascal and Fermat about the middle of
the 17th century.
Among some
to chances, the following
other questions relating
was proposed
to Pascal.
"
Two

persons
sit
down
first
gains three
The
first
to play
games
on the condition that the one who
shall
be the winner of the
stakes.
having gained two games, and the second one,

they agree to leave off and divide the stakes in proportion
to their respective probabilities of winning
what share is
:
each entitled to take ?"
Pascal solved the question, but by
a method which was applicable only to the particular case.
Fermat, to
whom
was communicated by Pascal, employ
it
ed the direct and general method of combinations, and gave

a solution which could be applied to the case of any
num
His reasoning, however, did not at first

appear to Pascal to be satisfactory, and a correspondence
on the subject took place between these two illustrious
ber of players.
geometers, which
is
preserved in their respective works, and
throws some light on the history of mathematics
About the same period Huygens composed

Ratiociniis in Ludis Ale<z, which was
first
in that age. 1
his tract
Exercitationes Geometricce of Schooten in 1658,

the
first
De
published in the
This was
systematic treatise which appeared on the doctrine
of chances.
It
contained an analysis of the various ques
which had been solved by Pascal and Fermat, and

the end five new questions were proposed, the solutions
tions
at
may now appear, were then attend

The analysis of two of them
difficulty.
of which, simple as they
ed with considerable
was
in fact
given for the
first
tury after their publication.

into English
marks
and published
in
time by Montmort, half a cen
Huygen
1
relative to the advantage of the
of Pharaon, in an Essay on the
OBuvres de JBlaise Pascal, tome
Fermat, Toloste, 1679.
tract
was translated
692, with some additional re
banker in the game

edited and
Laws of Chance,
iv.
Paris,
1819; Opera Petride
PROBABILITY.
of
supposed to have been written by Motte, then Secretary
the Royal Society.
James Bernoulli appears to have been the first who per

ceived that the theory of probability may be applied to much
more important purposes than to regulate the stakes and ex
both of the
pectations of gamesters, and that the phenomena,
moral and physical world, anomalous and irregular as they
in
appear when viewed in detail, exhibit, when considered
renders
their
which
of
succession
a
large numbers, constancy
occurrence capable of being submitted to numerical
The Ars
mation.
Conjectandi, published in
a
years after the death of the author, contains
number of in
teresting questions relative to combinations

series
but the most remarkable result which
esti
1713, seven
it
and
infinite
contains
is
theorem respecting the indefinite repetition of events, which
may be
said to form the basis of all the higher applications
of the theory.
It consists in this, that if
a series of
trials
be
which must either happen or

each trial, the probability becomes greater and greater,
instituted respecting an event

fail in
as the
number of trials
number of times
it
is
increased, that the ratio of the
happens, to the whole
number of trials,
a priori probability in a single trial ;

and that the number of trials may be made so great as to
will
be equal to
its
give a probability, approaching as nearly to certainty as we

ratio of its occurren
please, that the difference between the
ces to the
its
number of trials, and the
a priori
tity.
fraction
probability, will be-less than
which measures
any assigned quan
Bernoulli informs us, that the solution of this impor
tant theorem
had engaged
his attention
during a period of
twenty years.
In the interval between the death of Bernoulli and the
appearance of the Ars Conjectandi, Montmort published

his
Essai
edition
cT Analyse
was
1708
in
tended, and enriched
sur
les
derable merit
now
it
forgotten,
considerably ex
and Nicolas
possesses consi
games of chance, many of whicli
has lost
much
of
its
original interest.
Demoivre began to turn his atten

the subject of probability, and his labours, which
About the same

tion to
The work
first
but being chiefly confined to the examina
tion of the conditions of
are
is
several letters of John
Bernoulli, appeared in 1713.
The
Jcux de Hazard.
the second, which
by
time,
were continued during a long life, contributed greatly to the

advancement of the general theory, as well as the extension
of some of
first
sura
its
most interesting applications.
publication on the subject was a Latin

/Sortis,
in the Transactions
of the Royal Society
1711. His Essay on the Doctrine of Chances

in
1716
a second edition in 1738
Demoivre
dated 1756.
appeared
but the third and most
elegance
but
on Lives,
This work contains a great variety of ques
tions relating to chances, solved with

;
for
first
valuable, including also his Treatise on Annuities

is
memoir De Men-
it is
much
clearness and
chiefly remarkable for the theory of re
curring series, there given for the
first
time,
which
portant use in investigations of this kind, and
equivalent to the methods employed in the

for the integration
is
modern
is
of im
in fact
calculus
of equations of finite differences having
constant co-efficients.
Of the
particular results obtained
by
Demoivre, one of the most important in reference to theory,

is an extension of the theorem of James Bernoulli, above
mentioned.
It follows
from Bernoulli
theorem, that
if
we
have a given probability that the ratio of the number of

occurrences of an event to the whole number of trials, will
approach to the a priori probability of the event within
certain given limits, those limits will become narrower and
PROBABILITY.
8
narrower, as the
number of trials
to complete the theorem,
it
is
multiplied
but
merical value of the probability that in a large

future
the
trials,
signed limits.
number of occurrences
For
this
of the natural numbers

trials
cation,
becomes very
trials is
10,000
purpose
1, 2, 3, 4,
an operation which,
industry.
order
will fall within as
we must
&c.,
nu
number of
up
find the product

to the
number of
direct multipli
attempted by
even when the number of
laborious,
inconsiderable,
for
if
in
necessary to assign the
is
and when the number
is
great, as
altogether beyond the reach of human

example,
A formula was however discovered by Stirling.
is
by means of which an approximate value of the product is

found by the summation of a few of the first terms of a series
which converges the more rapidly as the number of trials is
greater.
With
the aid of this formula,
Demoivre was en
abled to assign the probability in question, and thus give a

practical value to the theorem of Bernoulli.
The
objects and important applications of the theory of
probabilities having
been made known by the works now
mentioned, the subject has ever since been regarded as one

of the most curious and interesting branches of mathemati
cal speculation,
and accordingly has received more or
attention from almost every mathematician of eminence.
and especially
the volumes of the
great variety of questions connected with

relating to lotteries, are interspersed in
less
it
Paris and Berlin Memoirs, (particularly the latter,) by John

and Nicolas Bernoulli, Euler, Lambert, Beguelin, and others.
D Alembert
has likewise treated of the theory in several of
the volumes of his Opuscula; and

that in
some instances
its first
it is
not a
little
remarkable,
principles should have
misunderstood by so ingenious and profound a writer.

the St. Petersburg Memoirs, (vol.
v.) there
is
been
In
an interesting
paper by Daniel Bernoulli on the relative values of the ex

pectations of individuals who engage in play, or stake sums
on contingent benefits, when regard is had to the difference
of their fortunes
is
a consideration which, in
necessary to take into account
value of a
on
ly
On
sum
of
money to an
it is
individual,
many
cases,
it
obvious, that the
depends not mere
absolute amount, but also on his previous wealth.
its
this principle Bernoulli
expectation, which admits of
has founded a theory of moral
numerous and important
cations to the ordinary affairs of

the
for
Royal Society
for the years
life.
appli
The Transactions of
1763 and 1764, contain two
papers by the Rev. Mr. Bayes, with additions to the latter
by Dr. Price, which deserve to be noticed, inasmuch as the

principles on which the probability of an event is determin
ed,
when
the event depends on causes of which the exis
tence and influence are only presumed from experience,

are there for the
first
time correctly laid down.
The
ques
and solved by Bayes was this a series of ex

periments having been made relative to an event, to deter
mine the presumption there is, that the fraction which mea
tion proposed
sures
its
probability
falls
within given limits.
One
lity
of the earliest applications of the theory of probabi

was to determine, from observations of mortality, the
average duration of
interests
human
depending on
its
life,
and the value of pecuniary
continuance or
particular application appears to
have been
failure.
This
first
thought of,
be carried into practical effect, in
Holland, by Hudde and the celebrated pensionary De Witt
or at least attempted to
but the
first
tables of mortality, with the corresponding va
on single lives, were constructed by our

countryman Dr. Halley, and published in the
For the history of
Philosophical Transactions for 1693.
lues of annuities
illustrious
PROBABILITY.
10
this
articles,
AN
in the seventh edition of the
En
branch of the subject, we refer to the two
MORTALITY
NUITIES and
cyclopaedia Britannica.
We
may
remark, however, that
although the English writers, who have expressly treated

of it, have almost without exception confined themselves to
the explanation of the methods of computing annuity tables,
and of determining from them the values of sums depend
ing on
nomy
life
contingencies, the aid which this branch of eco
derives from the general theory of probabilities,
no means confined
questions.
to the consideration of
The number of
is
by
such elementary
observations necessary to in
risks
spire confidence in the tables, the extent to which
dif
of
the
be
safely undertaken,
comparative weights
may
ferent sets of observations, and the probable limits of de
parture from the average results of previous observations in a
given number of future instances, are
all
questions of the ut
most importance, which come within the scope of the calculus,

and cannot, in fact, be justly appreciated by any other means.
The
application of the theory of probability to the subject
of jurisprudence, and the verdicts of juries and decisions of

tribunals, has
been discussed by the Marquis Condorcet
rious articles in the Encyclopedic
in
va
but more
Methodique
Essai sur F application de fAnalyse a la
Probabilite des Decisions renducs a la Pluralite des Voix,
;
especially in his
Paris 1785; a
work of great ingenuity, and abounding with
interesting remarks on subjects of the highest importance to
humanity.
James
Bernoulli,
it
appears, had intended to
branch of probability in the Ars

premature death prevented that work
treat jurisprudence as a
Conjectandi, but his
from being completed.
There
is
a memoir on the subject
by his nephew Nicolas, in the Leipsic Acts for 171 1. The

most important questions to be determined, are the number
1 1
of jurors of which a jury ought to consist, and the majority

which should be required to agree in a verdict in order to
afford, on the one hand, the greatest probability that an ac
cused person will not be wrongly condemned
and, on the
other, to give to society the greatest security that its inter

ests will not
be compromised, by allowing too great facili

This important subject has
ties for the guilty to escape.
been treated more profoundly, and with numerical elements

derived from much better data than existed in the time of
Condorcet, in a recent work by Poisson, to w hich
r
we
shall
presently allude.
Another of the moral subjects to which the theory of pro

bability has been applied, and connected with the preceding,
is
the appreciation of the evidence of testimony.
ters of this kind,
it is
In mat
easy to see that the calculus must be
founded almost entirely upon hypothetical data. The vera

city of a witness can scarcely be made the subject of direct
experiment ; and by reason of the complicated circumstan
ces with which the facts forming the subject of testimony
are usually accompanied, and the numberless ways in which
mankind are influenced by

norance,
it is
their passions, credulity, or ig
perhaps equally impossible to deduce an aver

number of state
age value from the comparison of a great
ments which have been ascertained
to
be true or
merical results can therefore only be obtained
Nu
false.
by having
re
course to hypotheses, and consequently must be considered

as only probable approximations.
which
is
The knowledge, however,
thus obtained of the various combinations of the
quantities concerned, affords important aid in guiding our
judgments
cide
upon
in
complicated cases, and
conflicting testimony.
when we have
to
de
Approximations deduced
from a train of accurate and systematic reasoning, are
al
12
PROBABILITY.
ways to be preferred to the most specious arguments drawn

from any other source.
The analysis of probability has been applied with signal
advantage in
researches of Natural Philosophy, but
many
especially in appreciating the
Owing
mean
to the imperfections of sense
errors of observations.
and of instruments, phy
magnitudes are only susceptible of being measured with

and where the last degree of
in certain limits of accuracy
precision is indispensable, as in practical astronomy, it is on
sical
by means of a very great number of measures, compared

with one another, and combined according to the methods
ly
which
this calculus points out, that
we can obtain
est approximation to the true values
are capable of giving.
the near
which the observations
The mean errors of observations were
treated as questions of probability by Lagrange in the Turin
Memoirs
for
1773
but
it is
to Laplace that the theory
owes
The me
thod of combining numerous equations of condition now
universally followed, known as the method of minimum
its
principal extension and most important results.
squares, and which Laplace has demonstrated to be that
which leaves the
least probable
amount of error in the
final
equations, was made known by Legendre in an Appendix to

his Nouvelles Mcthodes pour la Determination des Orbits
des Comctes, published in 1806.
A similar method, however,
or rather the same, (for they are identical in principle,)
had
been discovered by Gauss, and employed by him for several

1
years before the work of Legendre made its appearance.
Laplace
bilites, first
great work, the Theorie Analytique des Proba-
published in 1812,
is
one of the most remarkable
productions that has ever appeared in abstract science.
Theoria Motus Corpcrum Ccelestium) p. 221.
The
13
principles of the calculus, as well as the peculiar
of analysis which
far
it
and the most interesting and
requires,
questions which
difficult
methods
presents, are here discussed in a
it
more general manner than had been attempted by any

and it may be said, accord
former writer on the subject
have placed the theory under an entirely new aspect.

much to be regretted that so little pains have been taken
ingly, to
It is
by the
author to render the work intelligible to the
illustrious
generality of mathematical readers. Consisting for the greater

part of separate
Academy
or order,
memoirs presented
at different times to the
of Sciences, arranged without regard to symmetry
it
abounds with repetitions which only serve
to
em
barrass the student; while the deficiency of explanation
com
bined with the subtlety of the analysis, and the inherent
intri
cacy of the subject, render
it
often a painfully difficult task to
seize the force of the demonstrations. Notwithstanding these

defects,
however,
it
forms one of the most splendid creations
of mathematical genius
and
is
alike admirable,
whether we
regard the extension which has been given to the calculus,

or the results which have
lofty philosophy in
been arrived
at,
or the tone of
which subjects bearing on some of the
most important concerns of mankind are treated.
Next
to the Theorie
Analytique of Laplace, the most im
portant work which has hitherto appeared on the subject of

probability
is
the recent one of Poisson, entitled, Recherches
sur la Probabilite des Jugements, (Paris
might be inferred from the
title
that this
837.) Although it
work relates only to
a single though very interesting application of the theory, the
greater part of
it is
devoted to the development and demon
stration of the general principles,
and the discussion of the
principal questions which present themselves in the differ

ent applications
and it is only in the last of the five books
;
PROBABILITY.
of which
refers
is
it
consists that the special subject to
taken into consideration.
to the decisions of tribunals,
which the
title
In applying the theory
Condorcet and Laplace had been

from the want of authentic
unable to obtain positive results

data
but the recent publication by the French government
of the Comptes Generaux de VAdministration de la Justice
Crimmellc, in France, having furnished an immense collec

tion of facts
from which the requisite data could be obtained,

to consider the subject anew, and the results
Poisson was led
of his investigations, which are of singular interest, are given

in the
work now mentioned.
theory of the
to the
It is
mean
Poisson had already given a
errors of observations in the Additions
Connaissance des Terns
for
1827 and 1832.
in these two works of Laplace and Poisson that the
higher and more abstruse parts of the theory of probabilities

must be studied.
very clear exposition of the principles,
accompanied with
many
interesting remarks on the uses
applications of the theory,
able
little
bilites,
is
and
given by Lacroix in his valu
work, Traite Elementaire du Calcul dcs Proba-
Paris 1822.
Since the time of Demoivre, the English treatises on the

general theory of probability have neither been numerous,
nor, with one or two exceptions, very important. Simpson s
Laws of Chance
examples, in
(1740) contains a considerable number of
the solution of which the author displays his
usual acuteness and originality, but as they belong entirely

to that class in
which the chances are known a priori, they
give no idea of the most interesting applications of the theory.
Dodson s Mathematical Repository contains a

of the same kind.
The Essay
in the
large selection
Library of Useful
Knowledge, by Mr. Lubbock, gives a more comprehensive

and philosophical, though an elementary view of the sub-

ject
but by
far the
most valuable work
in
15
the language
is
the Treatise in the Encyclopedia Metropolitana, by Pro
De Morgan, 1837- In this very able production, Mr.

De Morgan has treated the subject in its utmost generality,
fessor
and embodied, within a moderate compass, the substance of

the great work of Laplace.
Within the
limits to
which the present
article
must be
would be hopeless to attempt giving a complete

view of a branch of science which embraces so many com
confined,
plicated
it
and
and which re
intricate subjects of research,
quires the aid of
some of the most abstruse and recondite
theories of the
modern mathematics.
which
order to be
In the higher appli

cations of the theory, the analysis of many of the questions
arise, in
made
intelligible,
would require
an extent of development and a parade of mathematical
for
mulae altogether incompatible with the plan and scope of

this work.
All that we can propose to ourselves, therefore,
is
to explain as briefly as
may appear consistent with
cuity, the general principles of the theory,
manner
and
perspi
to give
an
which these are applied to some cf

the more important questions which have been investigated
by Laplace and Poisson. The examples will be selected
outline of the
in
with a view to shew the nature of the principal results of the
mathematical theory, as well as the peculiar methods of ana

which are of most general application.
lysis
PROBABILITY.
16
SECTION
I.
GENERAL PRINCIPLES OF THE THEORY OF PROBABILITY.
1.
The term
in reference to
probable., in its popular acceptation,
any unknown
our judgment the event
in
more
is
likely to
used
be true than
not, or more likely to happen than not to happen.
out attempting to
is
or future event, to denote that
With
make an accurate enumeration of the va
rious circumstances
which are favourable or unfavourable
to its occurrence, or to balance their respective influences,
we suppose
there
is
cordingly pronounce
a preponderance on one side, and ac

to
it
be probable that the event
lias
occurred, or will occur, or the contrary.

2.
to
If
we can
see no reason
why an event is more likely

we say it is a chance whe
happen than not to happen,
or if it may happen in
ther the event will happen or not
more ways than one, and we have no reason for suppos
;
ing
it
will
another,
happen
we
say
in
it is
any one of these ways rather than

a chance whether
it
will
happen
way or in any other. Suppose, for

unknown number of balls of different colours
assigned
in
an urn, from which a ball
person blindfold.
Here we
that the ball about to be
than another,
that
it
any
example, an
to be placed
about to be extracted by a
have no reason for supposing
is
drawn
will
in
in
will
be of one colour rather
be white rather than black,
GENERAL PRINCIPLES.
or red
and accordingly say
come
will
it is
17
a chance whether the ball
out of a particular colour, or a different.
instance, then, the
In this
term chance denotes, simply, the absence
of a known cause. If, however, we are made acquainted

with the number of balls in the urn, and the number there
are of each of the different colours, the term
is
used
in a
For instance, suppose the urn to contain

definite sense.
ten balls, of which nine are white, and the remaining one
black,
we
white
ball,
black
ball.
say there are nine chances in favour of drawing a
and one chance only

Chance, in
in
this sense,
favour of drawing the
denotes a
way of hap
may arise
pening, or a particular case or combination that
number of other
out of a
possible cases or combinations
and
an event becomes probable or improbable according as the

number of chances in its favour is greater or less than the
number
Chance and presumption are also

against it.
used
quently
synonymously with probability.
3.
The mathematical probability of any event
is
fre
the ratio
of the number of ways in which that event

to the
whole number of ways
in
which
it
may happen
may either hap
Thus, recurring to the previous example, the

event, namely, the drawing of a ball from an urn con
pen or
fail.
taining 9 white balls
and
1 black,
may happen in 10 dif
any one of the 10 balls may be
drawn but in one only of those ways will the event be a
and therefore the probability of drawing the
black ball
ferent ways, inasmuch as

;
black ball
is
1
-y ^.
In like manner, as there are 9
ent ways in which
a white
ball
may be
differ
drawn, or 9
chances of drawing a white ball, and ten chances in all, the

9
probability of drawing a white ball at the first trial is T S
It follows immediately from this definition, that the proba
bility
of drawing a ball of either colour will remain the same,
PROBABILITY.
18
however the number of
urn
balls in the
may be
increased,
provided those of each colour are increased in the same pro
For instance, suppose the number of white balls

balls to be 5 ; the num
portion.
be 45, and the number of black
to
ber of chances in favour of drawing a black ball

while there are 50 chances in
bability of a black ball being
all,
drawn
is
/y=r^
manner, the probability of drawing a white

the same as before.
trary events, that
is
Generally, let
is
5,
consequently the pro
ball is
and
to say, such that the
In the same
|^=r-j g
be two con
one or the other
of them must necessarily happen, and both cannot happen
together
and
let
a be the number of chances or combina
which produce the event E, and b be the number of

combinations which produce the event F, or cause the fail
tions
ure of
then the probability that
and the probability that

happen
is
will
will
happen
happen, or that
is
will not
In future, the term probability will be us
ed only to signify mathematical probability.

4.
It
is
to
be carefully remarked, that the
different
chances or combinations which form the elements of pro

If this equa
bability are supposed to be perfectly equal.
lity
does not hold, and there
is
any circumstance respect
ing the event under consideration which renders one

bination or set of combinations
more
com
likely to occur than
another, the different combinations must be multiplied
numbers proportional
which the units
many
distinct
event
will
to
their respective
facilities,
by
after
each multiplier may be regarded as so

chances, from which the probability of the
in
be found by the above formula.
lent to saying that a combination or
This
is
chance which
equiva
is
twice
GENERAL PRINCIPLES.
happen as another, must be regarded
as likely to
equal and
19
as
two
similar combinations in comparison of that other
a proposition which
5. It follows
is
sufficiently obvious.
from the above
definition, that the probabi
measured by a fraction less

and 1. It
than unity, and may have any value between
of
two
which
that
the
sum
the
fractions
mea
also,
follows,
lity
of any contingent event
is
sure the probabilities of two contrary events

unit,
which
is
equal to
is
the measure of certainty, inasmuch as either
the one or the other necessarily occurs.
example, the probability of the event
the contrary event F
-4
is
+b
and
Thus,
is
-,
in the last
and that of
7= 1
ab*a+b
Hence
denote the probability of any event E, and q the pro

This
bability of the contrary event F, we have q~ 1
p.
if
consequence of the definition
is
of great importance in the
calculation of probabilities.
6.
We
cessarily
to
have here supposed the result of a

one or other of two events E and F
imagine the
trial
to
trial to
;
be ne
but
it is
easy
it
may
give
be of such a kind that
any one of a number of events E, F, G, H, &c. each

having a given number of chances in its favour. This case
rise to
is
represented by supposing an urn to contain balls of as
many different
colours or sorts as there are different events.
Let the urn be conceived
to contain
balls of the sort
which produces the event E, b of the sort which produces

F, c of the sort which produces G, and so on ; and let a-fb-\-c
+ d,
&c.
the urn.
H, &c. are
The
k,
so that k
is
the whole
number of balls
events E, F,
probabilities of the different
then, respectively,
by
the definition,
in
G,
20
PROBABILITY.
abed
T T T T
sum of which =1.
the
it
In
all,
different sorts contained in
and consequently the sum of all the probabilities
amounts
to unit or certainty.
When
7.
a ball be drawn at
fact, if
must be of one or other of the
the urn
&c<
an event
is
compounded of two or more simple
events independent of each other, the probability of the
compound event
is
equal to the product of the probabilities
of the several simple events of which
it
is
compounded.
Let us imagine two urns, A and B, of which A contains a
white balls and b black, and B contains a white and b
Make a-\-b= c, and a
black.
pound event whose probability

drawing of a white
the c balls in
ball
+b
is
=:c
to
and
from both urns.
may be drawn
the
let
com
be determined be the
Now,
as each of
with any one of the c balls
whole number of ways in which the balls in A

be
differently combined by pairs with the balls in B,
may
or the whole number of possible cases is cc
But the num
in B, the
ber of cases favourable to the compound event
number of different ways
the
drawn from
to
aa
A with a white
Hence by
in
ball
which a white
is
evidently
ball
may be
from B, and therefore equal
the definition (4), the probability that a

ft ft?
white ball
will
be drawn from both urns
is
cc
p denote
Now,
if
the probability of drawing a white ball from A,
and p that of drawing a white ball from B, we have by the

a
aa
a
whence -pp.
definition p
,and
:
In general, let
p
on
that of another event

;
cc
denote the probability of an event E,
p" that of a
third
E", and so
then the probability of the concourse of the events E,
GENERAL PRINCIPLES.
21
E", &c., OY the probability that they will
&c.
all
happen,
is
an event
to say, the probability of
p Xp Xp">
compounded of any number of simple and independent
;
is
the product of the respective probabilities of
is
events,
that
the several simple events.
The
probabilities that the several simple events E,
&c., will
E,"
not
all
some of them will hap

determined in the same man
happen, or that
pen and others fail, are easily

ner ; it will be sufficient to indicate their several expres
sions.
Suppose there are only three simple events, of which
q=l
Ip,
p, q
q"=l
p and p" and let

The product pq q" ex
the probabilities are respectively p,

p".
compound event which con

and E" both failing qp q" is
presses the probability of the
sists in
happening and
the probability that
both
fail
pp p"
pp p"
they will
all
8.
will
happen, and that
the probability they will
is
the probability they will not
one of them at
that
not
is
all fail;
three
fail,
least will fail
and
all
qq q"
all
and E"
will
three happen
three happen, or
is
the probability
qq q" is the probability they will

or that one at least of them will happen.
1
As an example of the
application of this rule, suppose
were required to assign the probability of throwing aces,

at one throw, with two common dice.
As a common die
it
has six symmetrical faces, there are in respect of each die

six
ways equally
The
happen.
one die
possible, in
which the simple event may
probability therefore of throwing ace with
p-=\. In respect of the second die,

hence the probability of the compound
event, or that aces will be thrown is j0p
J X
5V The
is
J, that
we have also p
is,
=J
probability that aces will not be thrown at

is
therefore (5)
aces at any given
=f
gs
J
trial
any assigned
trial
and the odds against throwing
are 35 to
22
PROBABILITY.
Again, suppose two numbers, each consisting of 7 di

at random, (for instance from a table of
gits, to be taken
logarithms), and let
be proposed to assign the proba

that the substraction of the one from the other will
bility
it
be performed without
its
being necessary, in any case,

Here, as each digit may
the upper figure.
to increase
to 9 both inclusive,
have any one of the ten values from
as each of those values in the upper line may be com
and
bined with any one of them in the lower
line,
there are 100
different combinations or equally possible cases for each par
Now, if the upper figure be 0, there is

those
favourable to the event, or which
of
cases
one
only
tial
will
substraction.
admit of the substraction being performed, namely,

If the upper figure be 1,
the figure below is also 0.
when
there are two cases favourable, namely,
the under figure
or
is
are three favourable cases, namely,
or 2.
is 0,
1,
gits,
the whole
Hence,
for
Proceeding
each
in this
when
any one of the figures
cases
=c.=yg5
there
the di
all
found to be
is
55 favourable
therefore (4) the probabi
in the
upper
line
than the corresponding figure in the under line

/
2,
the under figure
partial substraction there are
we have j9=//=//
be
way through
number of favourable
cases out of 100 possible cases

lity that
those in which
If the upper figure
is
is
not less
-f^
for the probability of
and
each
of the seven simple events or partial substractions, whence,
by
(7), the probability
of the
compound event
=:(-55)
which
9-
is less
When
than ^\, and greater than
an event
may happen
is
=0152243,
^.
in several different ways,
GENERAL PRINCIPLES.
23
each independent of the others, the probability of the event

the sum of all the partial probabilities taken in respect
is
of each of the different ways.
Suppose there are n different urns A t A 2 A 3 ...... A w

each containing balls of two colours, white and black, and
,
let the
whole number of balls in each urn respectively, be

2
1?
and the number of white
and
the event
let
......
balls in
n,
each be
be the extraction of a white
drawing a ball from any urn at random.
are
for it
happen,
any one of the urns.
drawn from any given urn,

be drawn from
;
therefore,
drawn from
is
therefore
this urn, the probability
by
1
.
c2
therefore
-.
trom A.
by
of
(7), the probability of
is
facility
from
probability that the ball will be
its
and
if
it
being white
is
a white ball being
In like manner the probabi-
Cj
of a white ball being drawn from
lity
which the event
may be drawn with equal

The
ball in
In this case there
different ways, all equally probable, in
may
ci
to
be
a~
cs
A2
and so
is
on.
shewn
to
be
Denoting
the whole probability of the event E, the
proposition affirms that
To prove
to a
this, let
the fractions
common denominator, and
tions to
>
&c. be reduced
suppose the equivalent frac
be
>
&**.#
j**
rr*^/-^
a
..-.
24
PROBABILITY.
We may
777
now
conceive the urns
A 1? A 2 A 3 ...A n to
,
be re
placed by others, each containing the same number, y, of

balls, and of which the first contains a l white balls, the se
cond
ball
a 2 ,and so on;
cisely the
the
and
being drawn from
first
same
as
it
it is
evident that the chance of a white
this
new system of urns will be pre
was
Now
system.
for a white ball being
drawn from
the probability of drawing a white
from the new system will not be altered by placing the

whole of the ny balls in a single urn, for they may still be
ball
conceived as arranged in groups, disposed in any manner

whatever, each group containing the same
number of balls,
and the same proportion of white to black as were

and as each group contains the same
separate urns
;
in the
num
ber of balls, the chance of laying the hand on any one group
is
the same as that of laying
lity
it
on any other.
The
of drawing a white ball from the single urn,
is
probabi
therefore
the same as for drawing
it from the
group of separate urns
which contain each the same number of balls. But the pro
bability of drawing it from the single urn is the ratio of the
number of white
balls contained in the
urn to the number of
both colours, therefore (this probability being p)
p=
whence, substituting
i+
for
O,
O.Q
,
n
>
&c.,
......
,7
values,
+3
we have
+o);
c.,
their
respective
we have
As a particular case suppose
three urns A, B,
C to be placed
GENERAL PRINCIPLES.
together, of which
contains 2 white balls and
C 4
3 white balls and 2 black, and
and
let it
25
be required
to
black
white and 3 black,
determine the probability
white ball being drawn from the group by a person
As
ignorant of the contents of the different urns.
of a
who
there
is
is
no
reason for selecting one urn in preference to another, the pro

bability that
he
he draw from
be drawn
is
and 3 cases
fore \
in
put his hand into the urn A is ^ ; and if

urn the probability that a white ball will
will
this
there being 2 cases favourable to that event,
The
all.
probability of both events
is
there
In like manner, the probability of the ball
being drawn from B is \ ; and if drawn from B the proba

bility of its being white is f ; therefore, the probability of
this
compound event
is
of the ball being drawn from

the probability of
bability of this
compound event
Lastly, the probability
is
being white
its
the proposition
by
and
is
is
bability of the event
/>=
drawn from C
therefore, the pro
now demonstrated,
if
^\
Hence,
the complete pro
is
+i+A=
*ff-
If all the balls had-been placed in a single urn, the proba
of drawing a white ball would have been T9j, for there

are 3
5
15 balls in all, of which
7
bility
+ + =
2+34-4=9
But T9^
J-J| a fraction which differs sensibly
from Uf, the measure of the probability of the same event
when the balls are distributed in the manner above supposed
are white.
amongst the
cases
10.
of a
is
The
different urns.
distinction
between the two
important.
The
rule laid
down
in (7) for finding the probability
compound event applies
alike
whether the simple events
are determined simultaneously or in succession.
when
In
fact,
the simple events are entirely independent of each
26
PROBABILITY.
other, the chances
which determine the compound event are
not influenced in any
way by the intervention
of time.
Sup
compound event to be the throwing

of
of a certain number of points with a given number
pose, for
example, the
dice
the chances for and against the event are obviously
the same whether the

die
is
thrown
m dice are thrown at once, or a single
times successively.
But
as the determin
ation of the probability of a
compound event is in general

facilitated by supposing the simple events to_be decided one
after the other,
it
will
be convenient
to
view the subject
in
method of forming the differ

ent combinations of the chances by which the probabilities
this light in explaining the
of
compound events
are determined.
EVENTS DEPENDING ON REPETITION.
SECTION
27
II.
OF THE PROBABILITY OF EVENTS DEPENDING ON A REPE

TITION OF TRIALS, OR COMPOUNDED OF ANY NUMBER
OF SIMPLE EVENTS, THE CHANCES IN RESPECT OF WHICH
ARE KNOWN A PRIORI, AND CONSTANT.

Suppose an urn to contain a -f- b balls, a white and
b black, and let a ball be successively drawn, and replaced
1 1
in the
urn after each drawing,
in order that the
favour of drawing a ball of either colour

in
every
trial,
and
let it
chances in
may be the same
be required to find the respective
probabilities of the different possible results of any
number
of drawings.
Let us
event
suppose the number of trials to be two.

happen in any of these four different ways
The
first
may
white, second white;
cond white
first
first white,
second black;
black, second black.
first
first
black, se
Assuming
to re
present the simple event which consists in the drawing of a
white ball, and B that of a black ball, and supposing the or

der of the arrangement of the two letters to correspond with
the order of succession of the simple events, the four pos
sible cases or
combinations will be represented thus
WW, WB, BW,

Now
trial
BB.
let the probability of drawing a white

be />, and that of drawing a black ball be
ball in
q,
any
(whence,
23
p=.
-1>
->
compound events
the probabilities of the four possible
are
by (7) respectively
probability of
of
WW
y>
as
-=.
under
p-
= pq
ofBW = q Xp=pq
of BB = ? X
=?
WB = p
</
If we disregard the order of succession,
and consider the two
arrangements WB and BW, which are equally probable, as

forming the same compound event, namely, a ball of both
colours in the two trials, the probability of this event, by
(9),
becomes 2 pq.
The sum
possible arrangements
whence
it
is
of the probabilities of
all
the
therefore
appears that the probabilities of the different ar
rangements
in
two
trials
are respectively the terms of the
development of the binomial, (p qYLet us next suppose the number of
The
different arrangements that
pimple events in three
trials,
respectively, are as follows
be three.
to
may be formed of the
with the probability of each
WWW, probability of which
~ppp~p~
WWB,
WBW,
BWW,
=ppq=p
WBB,
BWB,
BBW,
BBB,
It
trials
..........................
..........................
..........................
...........................
...........................
-qpq-pq^
..........................
^qqp=pf
..........................
=q(/q=f/-<
thus appears that the probability of obtaining two events
of one kind, and one of the other,
is
order they succeed each other, and,
the same? in whatever

in fact, is
independent
CHANCES KNOWN A PRIORI, AND CONSTANT.

of the order.
29
Disregarding, then, the order of succession,
and considering the combination of two white balls with

one black, in whatever order they may be arranged, as the
same compound event, the probability of its occurrence
any order whatever, being the sum of
its
in
probabilities in
each particular order (9), is fypq. In like manner, regard

ing the combination of two black balls with one white, in any
order of arrangement, as the same
compound event, its pro

The compound event resulting from three
is
3pq
must then happen
bability
in one of four different ways,

namely,
3 white balls; 2 white, combined withl black, in any order;
trials
2 black, combined with one white, in any order
sum of the
3 black; and the

cases
or, lastly,
probabilities of these different
is
Hence
the probabilities of
nations in three
trials
all
the different possible combi
are respectively given
by the deve
5
lopment of the binomial (p-{-q) denote
the
1 2. In
probability of any simple
general, let^?
event E, then the probability of E happening twice in two

2
of happening thrice in 3 trials jo 3 , and of hap
trials is j
,
pening
times in
successive
trials, //".
the probability of the contrary event

the probability of
F happening n times
In like manner,
being q
in
(jo-f^r
n successive
1),
trials
Hence (7) the probability of E happening m times,

and then F happening n times in succession, in m-\-n trials,
m n
But the probability of these events happening in
is p q
any assigned order is the same as that of their happening in
m n
any other assigned order ; therefore p q is the measure of
the probability that E will occur m times, and F will occur
n
is
n times
determinate order. Now, let m + n=h, and let

number of different ways in which m events E, and
in a
U be the
30
in h trials, and P be the proba

one of these combinations whatever, or the proba
occurring m times, and F occurring n times in h
n events F, can be combined

bility of any
bility
of
trials,
without regard to the order in which they succeed
we have then
each other,
P=U/y.
In order to determine the value of U,
the events in question to be so
many
we may suppose
different things repre
sented by the letters A, B, C, D, E, &c. of which there are
of one kind, and n of another, and
make m--n=h then

;
by the algebraic theory of combinations, we have

1
2 3
*
u=
This value of
may be
2 .3
is
mxl
symmetrical
.2 .3
in respect of
and
n,
and
otherwise written in either of the two following
forms,
h(h\)(h
2)
1.2.3
jr
_k(hl)(h
m-f-1
m
h
2)
1.2.3
n+l
n
which shew that the probability P, or the product Upm q"

is the
(m-j-l)th term of the development of the bino
mial
(p+q)* arranged according
to the
increasing
of^?, or the (w-fl)th term of the
powers
same development ar
ranged according to the increasing powers of q. Hence

we conclude that when p and q remain constant, the pro
babilities of all the different
compound events which can
be formed by the combination of the simple events E and

F in h trials, are expressed by the different terms of the
formula (p-)-q) h expanded by the binomial theorem.
The whole number of possible cases is evidently A-f-1,
for in h
1 times,
experiments, E may occur h times, h

h
h times;
2 times
which the contrary event
different cases are unequally probable,
greater or smaller
being the case in
this last
occurs in
the
all
The
trials.
both by reason of the
number of combinations by which they
produced, and which in reference
may be
31
to
each case
is
represented by U, and by reason of the inequality between
p and
h
q.
It will
be shewn afterwards, that when pz=q, and

is that in which
a whole number, the most probable case
is
the occurrences of
and
are equal
and
if
is
number, the two most probable cases are those

the difference in the
number of occurrences of F
13.
in
is
light, let
first
E will
in
namely,
h-n+ 1
1.2.3
term p h expresses the probability that the event

h
every one of the h trials. The second term hp ~^q
at the first or last or
determinate succession
+ q) \
times the event
is
E in
occur h
will
once, without distinction of order
happen
now demonstrated
_expresses the probability that
that
is
times,
to say
any intermediate
by suppressing the
third
term
[.2
If a
proposed, for example, that of
succession,
coefficient h,
and
F may
trial.
and F
in the
next
the probability of the event in the assigned order
The
which
and the
us consider separately the different
terms of the development of (p
The
in
unity.
In order to place the proposition
a clearer
number of occurrences of
an odd
and
is
is
consequently p
V*~V expresses
trial,
found
h
q.
the probability
EVENTS DEPENDING UPON REPETITION.
32
that the result of
be h
trials will
and twice the event F, without

particular order
be assigned,
ring in that particular order
necessary to suppress the
it is
general term
ph
is
-2
h(hl)(h2)
~2
it
trials will
be
F in any
and n times F
n) times the event E, and n times the event
The
order.
in
+~ph~n qn ex-
presses the probability that the result of

(h
Jf a
and the probability of the simple events occur
coefficient,
The
2 times the event E,
distinction of order.
probability of (h
an assigned order
14. If
we suppose
in favour of its
pq-=.^,
is
~n
n) times
n.
E to be such that the
the event
failing are equal, that
happening or
the different terms of the binomial Q?
pressing the coefficients,
chances
become
all
equal
+ #)
>
is,
if
on sup
so that a parti
cular order being assigned in each of the possible cases or
combinations,
all
the cases
become
equally probable.
Thus,
suppose a shilling to be tossed 100 times in succession, the

l
.
The
probability of head turning up in every trial is (^)
probability of 50 heads and 50
100 if
50
50
is(i)
x(i) =(i)
m heads and n tails
of
Hence
is
also
the probability of any
tails
in
any assigned order
wi+w=100,
n
the probability
w
(l)(l) =(i) +-(r)ioo.

compound event formed by
the combination of two simple contrary events succeeding
each other
in
an assigned order, and each having the same
probability,
independent of the ratio of the simple events,
and depends only on the number of trials. Before the trials,
it is an even
wager that head will be turned up in succession
is
100 times, and that the result of 100
and 50
trials will
be 50 heads
a given order of succession, or any proportion

of heads to tails in an order arbitrarily chosen.
This con
tails in
sideration
is
frequently lost sight of in reasoning about those
33
events of the natural world, which are termed extraordinary
and miraculous.
If in tossing a shilling
100 times into the
number of heads turned up is found nearly equal to

number of tails, the event excites no surprise some
the
air,
the
was expected. On the contrary, if the diffe

rence between the number of heads and the number of tails
thing like
is
it
considerable, the event
is
termed extraordinary
and
if
head turned up in every trial without exception, we should

scarcely be persuaded that such an event was entirely the re
sult of chance,
and independent of a special cause. Never
a priori probability that every trial will give

same as the probability of throwing
of
and tails in an assigned order of
number
heads
any given
succession.
It will, however, be proved afterwards, that if
theless, the
head,
is
precisely the
such an event as throwing head 100 times in succession were

actually observed, the probability of a special cause having
intervened, would approach very nearly to certainty.
15. Hitherto we have supposed the compound event to
be formed by the combination of two simple events only,
E and F, one of which necessarily excludes the other. Let
now suppose
us
Ea E5
,
p3
there are any
number of simple
events,
E 15
&c. of which the respective probabilities a.rep 19 p,^
&c. and such that one or other of them necessarily hap
pens in each
trial,
so that p^ -\-p 2 -{-p 5
+> &c.=
1,
and de
termine the probability of any assigned combination of them

in a given number of trials.
This case may be represented
by supposing an urn
to contain a
number of balls of as many
different colours as there are distinct events
the event E,
and its proba

be the drawing of a ball of the colour
is
will
numerator
be the fraction whose
bility p
equal to the
will
number of balls of the colour

number of
balls in the urn.
i,
and denominator the whole
Now
the probability of the
34
event Ej happening
that of
E3
E2
times in succession is//? by
happening n times
in succession
happening r times in succession
formed by the occurrence of m times

,
der,
p\
that of
p^
compound event which
fore (7) the probability of the
E5
is
and so on. There
E 1? n
times
E2
r times
and so on, these events succeeding each other

is
the product
p p p5
&c.
But the
is
in or
probability of
the simple events succeeding each other in any particular
order
is
the same as that of their succeeding in any other
assigned order (12) ; consequently, if

events
ber of different ways in which
events
and
E3
denote the
E 1?
n events
we
Assuming h=m+n-\-r+, &c. we have
in
also
__
We
&c.
being the coefficient of the term which has for
multiplier
tinomial
any order
by the theory
1-2.3 .................. h
~1 .2.3...mxl .2.3...nxl.2.3...rx
TT,-
its
have,
of combinations,
the factor
&c. can be combined, or succeed each other,
be the probability of the compound event
whatever,
num
E2
p p
p*, &c. in the expansion of the mul
A
(Pi+Pv+Ps + & C *) whence
shall
>
now proceed
to give
some examples of the ap
plications of the preceding formulae.

16.
Let
it
be proposed to assign the probability P, of
throwing ace once, and not oftener, in four successive throws

of the same die.
Simpson, p. 15.
Here, the chance of throwing ace in a single trial being

we have jo=}, and consequently 5 =^, and also 7<t=4.
Now the compound event being the occurrence of the sim,1,

whose probability
pie event E,
trary event
event
is jo,
35
once, and of the con
three times, the probability of the
compound
that term of the development of (p-\-q)* which
is
multiplied
by pq
If,
therefore, in the formula,
1.2.
T)
"~
we make
p^
(?
^
B>
is
w?
m=l,
7^=4,
ii
?i=3,
we
m fJ n
*
shall
have
1 .2.3.4
/5\ _12_5
X
X J_
6
\6/ "324
""1X1.2.3
which
the probability required, and the same as that of
is
throwing one ace, and not more than one, at a single throw
with 4 dice.
The
probability of the contrary event, that
probability of either not throwing an ace at
all,
to say, the
is
or of throw
ing more aces than one
is 1
1||
^|| ; and therefore the
odds against throwing one ace and no more in 4 throws of a
common die are 199 to 125, or 8 to 5 very nearly.
17. If in this
example
it
had been proposed
probability of throwing ace once at
least,
to assign the
instead of once
it would have been

necessary to have includ
ed those cases in which the ace occurs twice, or three times,
and not more,
or in each of the four trials.
the
first
The
binomial (jo-f-^) 4 gives
term of which expresses the probability of throw

the second that of throw
ing ace four times in succession
ing ace three times, and another

that of throwing ace twice,
and a
number once
the third
different face twice
the
fourth that of throwing ace once, and a different face three
times
and the
of the four
fifth
trials.
that of throwing a different face in each
But
events, excepting the
as every
last, satisfies
one of these compound

the condition of ace be
ing thrown once at least, the whole probability of that event
36
must be the sum of the

by which
probabilities of the different events
may be produced
it
and
(9)
is
consequently
W V^YW V VU;^W
+T6") T+TS-J Uv +4 U;UJ
5
671
\*
sum of the
In general, the
first n-\- 1
terms of (p
presses the probability of obtaining not less than h
the probability of each of which
or not
is/?,
trary events, the probability of each of
Since
p +<?= 1,
the
sum of all
1298
+ 9 )" ex
r
"
n events,
more than n con
which
is
q.
the terms of the series pro
duced by the expansion of (p+q) h
is
equal to unit, and
sum of any number of the terms is equal to

This
diminished by the sum of the remaining terms.
therefore the
unit
consideration frequently gives the
means of abridging the
preceding example, instead of

order to find the proba
>v expanding the binomial (^-fu) 4
bilities of throwing 4 aces, 3 aces, 2 aces, and 1 ace only,
calculations.
in the
Thus,
we might have sought the probability

The probability of not throwing
throwing ace at all.
in a series of 4 trials,
of not
ace in a single trial

ing it in 4 trials is
is
|j,
and therefore (7) that of not throw
Hence
() =T 6/^5-
the probability of
the contrary event, namely, that ace will be thrown once

or oftener,
18.
is
Let a
-f^fy^-fifo
be tossed
shilling
more than 3 heads
will turn
what
up
case,/>=i, g=, A=10

=:()io(l -f l)i. Now the
this
the same as before.
is
therefore
in the first
10 trials?
In
(p+q) =(i+i) 10
term of this development

expresses the probability that head will not turn up in any
one of the ten trials ; the last but one, the probability that
it
will turn
up once
it
will turn
up twice
ity that
it
will turn
last
the last but two, the probability that
and the
last
up three times
but three, the probabil

;
therefore the four last

terms include
37
the different ways in which the ten trials
all
give not more than three heads
and
their
sum consequently
expresses the probability that not more than 3 heads will
Now
be thrown.
expan:ion of
-f-
the last four (or

1
1Q 9
10
gives yVy^j
whence the
n34 = M
and the odds
than three times in 10
more than 3 heads
probability of the contrary event
up
more than 3 heads
;
8
i
176, which multiplied by (i) 10= To35^
is
for the probability that not
will turn
or that
1.2.3
1.2
and their sum
four) terms of the
are
in
first
will
be thrown,
trials
is 1
JQ"^
throwing heads more
in favour of
are 53 to 11.
engage in play; the probability of A s

a
is
winning game
p, and the probability of B s winning a
19-
and
A s winning m
wins n games, the play being supposed to
terminate when either of those events has occurred.
game
is
required the probability P, of
games before
It is
evident that the question must be decided at the
by the (m + n l)th game for supposing m-\-n 2

games to have been played, there is only one combination
according to which the match can remain undecided, name
latest,
ly,
which
that in
in this case the
Suppose
m+
A has won m
bility that of these
and
Bn
games
and
games
m have
been won by A, and x by
m +*
represented by the term of the binomial (p-\-q)
B,
is
in
which the factor
m
p q* occurs (13) which term
1.2. 3 ......... m + x
But
1,
next game necessarily decides the match.

x games to have been played. The proba
last
is
.2 .3
cannot win
wins the
game,
games out of m-\-x exactly unless he

must have won m games
for otherwise he
EVENTS DEPENDING ON REPETITION
38
out of
m+x
1, if
therefore that
not out of a smaller number.
A may
win
m games
necessary in the first place
is
Now
game.
of m-Jf-x
1
.
any order (13)
in
that
the probability of his
2 .3 .........
he
exactly,
1
it
out of
also the
winning m
next
games out
is
m+x
1X1
2.3...m
In order
m+x
wins m
out of
any order, and then that he wins
in
m-\-x
.2.3...x pm
_j
x
5
and the probability of his winning the following game

probability of both events is (7)
is
p,
whence the
1.2.3 ...... m x 1
1.2.3 ...... m 1X1 .2.3 ......
As
winning
which
which, therefore, expresses the probability of
games out of m + x exactly.
If
we suppose a?=0,
the probability of
.*:=: 1, it
As
this
formula becomes
winning
games
bability that
If
x= 2,
it
A wins m games
--123
out of
P *"^
and so on.
till
games out of w-f-3
we
^G
probabilities of
all
the different
Now
number not greater than

it
decided by
If
1)
the pro
x=3,
it
be-
wins
Continuing this process

m x
by p q the sum of the
,
As
winning
is
games
m-\-x.
has been shewn, that the match
(m + n
compound events
which expresses the probability of

out of a
m + 2.
Probability that
arrive at the term multiplied
is
If
A wins m games
becomes
m(m+\)(m4-2)
comes
in succession.
becomes mp mq, the probability that
outofw-j-1.
jo
is
necessarily
games; consequently the
solution

of the question
obtained by substituting n
is
39
for
in
the last formula, which will then express the probability
of
As
winning
m games
greater than m-\-n
any order, out of a number not
in
On making
1.
this substitution,
we
obtain
-j
l+mq+
?=/>"
m(m -f- 1 ).~.m

.
The
vour of B, or that
into q,
and
is
will
match
probability Q, that the
greater than m-\-n
-f-
.2
,.-}.
be decided
will
win n games out of a number not

found by changing m into n, and
1, is
therefore
m + n-2
n(n+l)
1.2
m\
As an example,
n=2.
The
in fa
let us
suppose
probability of
As
-, q-=. -,
m=4,
and
winning the match, or the
value of P, becomes
and the probability of
Bs
winning the match, or the value
ofQ,
/l\f
2.3/2v
^3^1.2MU
In this example the

great as that of B,
won by him in
number
as the
2.8.4/2v) _
l-2.3W
131
j""243
A is
supposed to be twice as
and the number of games that must be
skill
of
order to gain the match
required to be
won by B
is
also twice as great
in
order that
B may
40
gain
one might therefore suppose, that when they begin

chances in favour of each are equal. But the
to play the
result
shews that the chances
in favour of
those in favour of B in the proportion of

it
appears that
who
it
would be unsafe
to
A are fewer than
12 to 131
wager
whence
that a player
has two chances in his favour while his adversary has
only one, will gain four games before his adversary shall
have gained two.
Suppose A and B, engaged in play, agree to leave off be

match is decided, it is evident that the stakes ought
be shared between them in proportion to their respective
fore the
to
and consequently the share of each

found from either of the above expressions for P and Q.
probabilities of winning,
is
This was one of the questions proposed by the Chevalier

de Mere to the celebrated Pascal, to which allusion has al
ready been made.
An
20
urn contains n-\-l
balls,
marked with the num
n\ a ball is successively drawn and re

1, 2,
placed in the urn, so that the chance of drawing any given
bers 0,
number remains the same
in
that in h trials the sum of the
The
in
each
trial,
what is the probability
numbers drawn will be equal to*?
problem depends on the number of

number s can be formed by the addition of
solution of this
which the
ways
h different numbers, each of which may have any value from
If we suppose the numbers marked on the balls to
to n.
be indexes of a certain quantity #, and develope the expres

2
n h
sion (x-\-x l -\-x
the coefficient of any term of
-\-x )
,
the development will indicate the

in
which the
balls
may be drawn,
number of different ways

sum of the num
so that the
bers drawn in h trials shall be equal to the

1
Demoivre, Miscellanea Analytica,
Probability, p. 253, et seq.
p.
196
sum of the
in-
Laplace, Essai sur
les

dexes of x
term.
in that
coefficient of that
of the indexes
is
is
we denote by
therefore,
the
term of the development in which the sum

will be the number of cases fa
s, then
But the whole number of
vourable to the event.

cases
If,
possible
therefore the probability of the event
(/z-J-l)*;
On
account of the particular form of the polynomial

is found without
question, the value of
difficulty.
is
in
Because x-}-x
n
-{-x ...... -\>x =:
2
_x
-}-x
......
+xn ) h =(l
l
,
therefore
x n + l )\l
pressing these two factors in series,
h.
Now, ex
xn + l )h
x)~
we have (1
1.2
and the
coefficients of the several
these two series in which the
be found as follows
terms of the product of

the indexes is s will
sum of
(1.) Multiply the
first
term of the
first
cient of the product will

(2.)
be
/i(/*-fl)(/i
v
+ 2)
1.2.OQ
,
Multiply the second term of the
term of the second
series
which has
for its
by that term
x 1 the coeffi-
series
of the second series of which the argument
is
......
h+s ~1
......
by that
s~ n ~ l
x
argument
first
series
the coefficient of the product will be
,_
k(h
+ \)(h + 2)
1.2.3
(3.)
...... k
+ sn
......
2
1
Multiply the third term of the first series by that

series which has for its argument
term of the second

*_2(n+i).
h e coefficient of the product will be
42
h(h\) X A(A+ !)(/* + 2) ...... h+s2n3

~T72~
......
1.2.3
(4.)
the
new
and
is
Proceed
in the
same manner with the fourth term of
and so on with the
first series,
others, advancing at each
multiplication one term to the right in the
terms to the
n-{- 1
reached
equal
left in
in the first series, the
or greater than
to,
exponent of x
The sum
s.
in
which
is
of the several products
We
thus obtained will be the value of N.

......
first series,
the second series, until a term
have therefore
h+s-l
1.2.3
h
.2.3
h(h
)(h
2) ......
1.2.3
1.2
......
1
s2n
......
_&c.
The series now found for N may be changed
into another,
having a more elegant form, by reducing all the terms to

1.
others having the common denominator 1.2. 3 ...... h
This will be accomplished by leaving out of the numerator
and denominator of the first term all the numbers after h 1
to
5,
(including
serting the
s),
when
s is greater
than h
numbers between s and h
when s^Ji
1,
or
by
in
(the last included),
by leaving out of the numerator and deno

all the numbers from h
1 to
minator of the second term

s
n,
or
by inserting those numbers
other terms.
1.2.
If
3.../
we then make
l=/e,
we
n )(s
shall
the
and so on with the
common denominator
have
) ...... (s
n -f //
2)
43
_&c.
to
be continued
comes
till
the last factor of one of the terms be
or negative.
we
If
also
make
+ h2=f(n+ !)=/
h3=f2(n + )=/"
2n 4.
s3n+h4=f3(n+ !)=/"
&c.
and write the factors

order, the
each of the terms in the reverse
in
above value of
N will become
+ &C.
As an example of the application of this formula, let
be required to assign the probability of throwing the point
16 with 4 common dice. (Simpson, p. 53.)
21.
it
die having
no face marked
0, it is necessary, in
order
formula to this case, to suppose the number of

points on each face to be diminished by unit, which is equi
to adopt the
valent to supposing s
to
be substituted
bers are then 0, 1,2, 3, 4, 5, and
)=12.
Hence
we have
for
s.
rc=5,
The num
^=4, and
44
f =f-2(n+})=3
f"=f 3(n+l )=3,
and A
we
Substituting these values in the formula,
3>
find
336)
9.8.7Xg(=
or
N= 125.
and
Now the probability of the event is N-*-(w + 1)*;

A
in the
=6
=il296 ; consequently
present case (ra-j- l)
the probability required, namely that of throwing the point
16 with 4 dice,
is
TO^T-*
22. In the numerical solution of questions of this sort,
it
sometimes happens that the labour
may be abridged by
the
of
computing
probability
throwing a different point from
that which is proposed, but which has the same number of
in its favour.
For example, let it be proposed to
determine the probability that in throwing 10 dice the sum
of the points will be 50. In this case, the smallest number
chances
of points that can possibly be thrown
60; and the chances
in favour of
and the greatest

throwing 10 and of throw
is
10,
The probability of throwing

ing 60 are obviously equal.
any given number of points above 10 is also evidently the
same as that of throwing the number which is as much under
60
and consequently the probability of throwing 50
same
as the probability of throwing 20, these
at equal distances
from the extremes.
bability of throwing
20 with 10
the
numbers being
Now to
dice, or,
is
which
find the pro

is
the same,
the probability that in 10 successive drawings from an urn

containing 6
balls,
marked with the numbers
sum of the numbers drawn
the
be
=1
negative,
Substituting these numbers in the series
2... 9
N, and observing that since f.
for
10,
0, 1, 2, 3, 4, 5,
we have ^10,
5=10; whence /=19,/=13,/"=7,/ "
7*=5,
and
will
45
7i-f
2=
the third
1,
term becomes negative, we have
19.18.17.16.15.14.13.12.
11
*=-1.2.3.4.S.6.7.8.9
(=923
8)
13.12.11 .10.9.8.7.6.5.
1.2.3 .4.5.6.7.8.9 (=-7150

and consequently N= 85228. Dividing this by (n +
=6
= 60466176, the probability of throwing 20, or of throwing
-
50, with 10 dice
23.
drawn
The
in
is
QOOQ
found
not exceed s
is
between
- and
number of
points
or
probability that the whole

trials will
10
found by substituting
for s the different values 0,
1, 2 ...... s in the series for N,

and taking the sum of the results. This labour, however,
may be avoided by means of a property of the figurate num
It is well
bers.
known
that the
sum of the
bers obtained by giving n every value from
series of
n=
to
num
n=v (v
being any number whatever) in the formula
*(*+!)(* + 2) ...... (n + u)
is
expressed by
1-2
this other
formula
(p
or,
which
tained
to
&=
is
+ 2)
......
w+1
......
the same thing, that the
sum of the
series
ob
by giving x successively every value from x=l-{-u

1
-f u
-{-
v in the formula
46
EVENTS DEPENDING OX REPETITION-
l)(g2) ...... (xu)

1.2.3
w+1
gfo
is
expressed by this other formula

(x
+ l)x(x\)
in
which
......
xu
.2.3 .......... w + 2
#=1 -j-w-}-i\
N (20) with
be evident, that on giving s every
value successively, from
to s in the value off, and denot
the
sum
of
all
the
results, we shall have
ing by
Comparing the
different terms of the series for
these last formulae,
it
will
y,
,(
1.2.3.4
...
l)(/-2)...(/-A+2)
h
...
1.2.3.4
//
l)(/
//
-2)-(/ -^+2)
1.2.3.4
A(A-l)
1.2
...
&c.
for the probability that the
not be greater than
number of points thrown
s.
As an example, let/*=lO, s=5, rc=5; we have
f=s +
will
= 14, / =/
and consequently the
then
+ 2=0,
O+l)=8, whence/
second term vanishes. Hence
A
18. 14. 13. IS. 11. 10. 9. 8. 7. 6
1.2.3.4.5.6
The
drawn
in
10 drawings not exceeding
3003
And
this
is
fifi >.{.>
OU"xOO
I/O
ing 10
16.
.7.8.9.10-
probability, therefore, of the
common
(Simpson,
dice the
sum of
5, is
the
numbers
3003 -f-(6) 10 or
,
also the probability that in throw-
sum of the
points does not exceed
p. 60.)
24. In the preceding questions the
number of trials, de-

noted by
h,
has been supposed to be given
ject, in every case, has
been
to
47
and the ob
determine the value of a
given term, or of a given number of terms of the series pro

duced by the development of the binomial (p-j-^y. But
there
is
a numerous class of questions in respect of which
the exponent h
is unknown, and is
required to be deter
mined from the condition that an assigned term, or the sum
of a certain number of assigned terms of the development,
For example, let it be proposed

3d
die must be thrown in
must have a given value.

to determine
how
often a
common
order to give the probability of ace turning up once at least

equal to a given fraction u.
Here the
ing ace in any throw being
Now,
probability of throw
we havep=r, and q=$>
term of the development of (p q} h except

gives a combination in which ace occurs once
as every
ing the
last,
or oftener, the question requires a value to
such, that the

shall
sum of the
be equal to
This
u.
common methods
of
first
trial
be found
for h,
h terms of that development,
may be
done, in general, by the
and error; but
in the present case,
the last term being the only one not included
among
those
which contain a chance of throwing ace, it is evident that

it is
only necessary to find the last term alone in order to
have the probability of not throwing ace in h trials, which by
u.
The last term of the development
the question is 1
is
(f
therefore
we must have
whence h log q=\og

Let
log
M=log
y,
(1
and ^=log
u,)
and <7=-, we
the equation qh ~l
shall
and log #=log b

A _log
log b
(1
then have log (1
log c;
log
u;
u) -i-log q.
u)-=.
whence
y^
log c
Substituting in this general formula the particular numbers
48
given in the question, namely b=5,
w=J, and consequently /3= 1,
yr=2,
c=6
and supposing
we have /*=.
7,-
log 5
logb
whence, by computing from the logarithmic
From
this
it
throwing ace once at

of not throwing
it
If the question
throw aces
at
h=;3
8.
greater than the probability
least, is
all.
had been
with two dice one

to
tables,
follows, that in four trials the probability of
may
to
determine in
how many throws
undertake, on an equality of chance,
at least once,
we
should have had
p=^, q=
f f and consequently 635, and c=36. Substituting these

numbers in the general formula, and observing, that in this
,
0=1, y=2, we
case also
The
get
A=
--
-=-
log 36
=24-6.
log 35
probability of not throwing aces once
is
therefore
greater than the opposite probability or that of throwing
aces once or oftener,

less
when
the
when
number
is
the
number of throws
is
24, but
25.
These two questions are celebrated
in the early history of
the theory of Probability, from the circumstance that the
Chevalier de Mere, by
whom
they were proposed to Pascal,
declared the two results above stated to be inconsistent with
each other, and thence took occasion to question the accu

racy of the theory of combinations by means of which they
had been obtained.

bility
He
reasoned thus
of throwing ace with one die
ing aces with two dice ^ of 5
given
= 35
is
J,
Since the proba

and that of throw
therefore, if there
probability in favour of throwing
be a
ace in fourthrows with
one die, there must likewise be the same probability of throw

4 24 throws in other words,
ing aces with two dice in 6 X
;
the chances in favour of an event

six times
more numerous than those
in a single trial,
being
in favour of F,
there

will
be as many chances in favour of Fin
in favour of
the
E in
The
one.
49
six trials as there are
error consists in supposing that
number of trials must increase
or diminish exactly in the
inverse ratio of the probability ofobtaining the proposed point.

25.
The
bility
general question
may be enunciated
the probability an event
Let p
it
will fail
probability =u
how many
that
will
E will happen,
as follows
q the proba
are required to give a

k
times.
happen
trials
Let r=the number required. Taking the sum of all the

x
terms of the development of (p-\-q) in which the exponent
of p is less than k, we shall have the probability that the
event does not happen k times in x
This sum must
trials.
u ; therefore, beginning
consequently be made equal to 1
the last term, and writing the terms in the reverse order,
we have
the equation
Let p-=.
e q,
and
this equation
becomes
from which the value of x may be found by the ordinary

methods of converging series.
Ifp=q^ then
consequently
x(
e=l and
u=^,
_...
Pascal, (Euvres, torn.
if
we
also suppose
the equation will
M=^, and
become
_
iv. p.
367
Lacroix, Elemental, p. 36.
50
"-JI
But the
first
side of tins equation
continued to k terms
terms
is
first
therefore, since the
last, it
follows that the
sum of the
first
whole number of terms must be
But the whole number of terms
10, then
the expansion of (1
)*
equal to one-half of the whole series, and the terms

half of the series are the same as those of the
of the
is#+l;
is
therefore
#=19
2k=zx+l, and.x=2k
hence
that head will turn
in the expansion of (1
in tossing a shilling
1.
it
up 10 times in 19 throws.
-f-
2.
1)*
Suppose k
is an even bet
CHANCES KNOWN A PRIORI, AND VARYING.
SECTION
III.
OF THE PROBABILITY OF EVENTS DEPENDING ON A RE

PETITION OF TRIALS, OR COMPOUNDED OF ANY NUMBER
OF SIMPLE EVENTS, THE CHANCES IN RESPECT OF WHICH
ARE KNOWN A PRIORI, AND VARY IN THE DIFFERENT

TRIALS.
26. Let us suppose the trials to consist in drawing balls

from an urn containing a white balls, and b black balls, and
that when a ball is extracted it is not returned to the urn.
Make a + b=c, and
W, and
event
trial
let
At
that of a black ball the event B.
the probability of
the second
the extraction of a white ball be the
trial,
the
is
(4),
and that of B,
number of balls
in the
urn
is
the
-.
first
But at
diminish
ed by 1 and the probability of drawing a white ball at the

second trial is therefore not the same as it was in the first,
;
but
is
place.
by the event which has already taken

happened at the first trial, the number of
influenced
If
white balls remaining in the urn
ber of black
The
is b,
probability of
-^-,and that of B
is
then a
the
and the number of both colours

at
-.
is
the
next
trial
In like manner,
if
is
num
c
1.
therefore
B happened
at the first trial, the probability of
at the second
is
52
and that of
which can
is
arise
Hence (7) the different combinations
from two
WW,
trials
are the following
WB,
BW,
BB,
the probabilities of which are respectively,
a(a
1)
c(c
Now
in
if
we
which
ba
ab
1)
c(c
1)
c(c
b(bl)
1)
c(c
1)
neglect the order of succession in the two cases
W and B are
combined, the probability of the
compound event which consists of the extraction of a ball

of each colour in the two
trials, is
lities
and the probabi-
r,
c(c
1)
of the three possible combinations are respectively
a(a
2ab
1)
b(bl)
c(clY c(c\y c(cl)

Comparing these with the
binations
when
same com
the chances are constant, or the ball
is
re
turned to the urn after each drawing, namely

J??
"?
"?"
"?*
the analogy of the two cases. is obvious-
After two balls have been drawn, the whole number re
maining
in the
urn
is
but the number of each colour
depends on the two events that have already occurred. If

two white balls have been drawn, the probability of drawing a white ball at the next
trial will
just seen that the probability of
Q
be -
2
;
WW
is
c(c
(7) the probability of
WWW
but
we have
q
is
) (a
^~\
C(C
i)(C
,
^; therefore
1)
~2A
4)
Thepro-

bability of
drawn
is
drawing a black
which b are white)

^
/
c(c-\)(c
two white have been
ball after
(for there are
now
2 balls in the urn, of
therefore the probability of
On
forming in
this
manner
all
WWB
the differtrials,
find
c(c
2)
l)(c
_
-- _
ab(a
^
If
is
2)
ent possible combinations which can result from three
we
53
we
disregard the order of succession in those combina
tions into
and
the same
W and B both
W twice and B once
as the
also the occurrence of
which
currence of
event
1)
J.
enter,
compound
- 3ab(b
^
c(c\)(c
same compound
once and B twice as
event, in whatever order they occur,
the probability of the former will be
the latter
and consider the oc
1)^
2)
.,
and of
54
m! -\-n f
27. In general, if
in
have been found
to
balls
of white balls in the urn will
number of
the
Hence
nf.
the probability of drawing a white ball in the next
be
trial will
if in
successively for
-,
Now
m!
to
now be a
n and the whole number of both colours
black b
have been drawn, of which
be white and n black, the number
and that of drawing a black
these two fractions
and n
and n
all
we
ball
substitute
the different numbers from
respectively, the product of the m-\-n
numbers thusobtained will (7)be the probability of drawings

white balls and
Let
72
black in an assigned order,
mm-\-n
be denoted by K, and we
this probability
(-+ l)x 6(6-1 )(6-2)

(c2)
(cmn +
ft(g-l)(a-2)
c(c
trials.
have
shall
(6-rc+l)
1
Hence, if we denote by
and n black being drawn
h trials, we shall have P= UK,
whatever the given order may be.
m white
the probability of
in
any order whatever,
where, as in (12),
in
balls
.2.3
h
.
mxl
.2.3
-2.3
the co-efficient of that term of the binomial (p-\-q) h which

has for its argument pm q n ; this co-efficient
expressing all
the different arrangements which can be formed of
things
of one kind, and n things of another.

28.
When
the urn
is
supposed to contain balls of more
than two different colours, the probability of any

proposed
number of each colour being drawn in a given number of
trials is
!
found with the same
of the
so on,
and
first
let
facility.
Suppose it to contain
a
of
the
colour,
second, a 5 of the third, and
2
al
+a
-\-a.
&c.=c then
;

in
m-^-n
first
+ rJf&LC.h
trials,
there will be drawn
n of the second, r of the
colour,
third, &c.
U xÔi
!)(!
2) ......... (!
X 2(
Xa 3 (a 5
1)(^
2) ......... (tf 2
1)(3
2) ......... (a 3
1)K
2) ......... (
&c.
55
of the
is
__
_ _
K
where, as in (15),
rj,_
1.2.3 ............ h
.2.3 ...... m*l .2.3 ...... wxl .2-3 ....... X&c.
/
The
29.
following examples will shew the use of the pre
ceding formulae.
Suppose a bag to contain
and 8 black, what
is
balls,
of which 8 are white
the probability that in drawing 8 balls
from the bag the whole of them will be white ?

Applying the formula (27) to the solution of
tion,
this
we have a=8, 6=8, cnl6, i=8, n=0, and
probability required
is
ques
as the
that of drawing white balls only, b
cannot enter into any of the factors of the numerator
hence
8. 7. 6. 5. 4. 3. 2. 1_ 1
""16.15.14.13.12.11.10. 9~~12870
and since
m=h, U=l,
the probability sought,
is
therefore
T270Let there be a heap of 20 cards, wherein are 7 diamonds,

6 hearts, 4 spades, and 3 clubs required the probability
;
that in drawing 8 of them at a venture there shall
3 diamonds and 2 hearts?
The
(Simpson,
come out
p. 21.)
probability required in this case being that of draw
ing 3 diamonds, 2 hearts, and 3 other cards which are nei-
56
may be con
ther diamonds nor hearts, the spades and clubs
sidered as forming one parcel, containing 7 cards.
then
/*
2,
r=3,
7/
=8
We
have
a 3 =7, c=20; m=3,

2 =6,
formula (28) a l =7,
in the
therefore
1.2.3.4.5.6.7.8
=l. 2. 3X1. 2 X 1. 2. 3-
and the probability required becomes,
X
The odds
7.6. 5x6
20
18
19
5x7
17
16
15
1225
14
13~3978*
against the event are therefore
2753
to 1225, or
nearly 9 to 4.
Let 4 cards be drawn from a pack of 52

?
probability of drawing one of each sort
In this case
c=52
1
also
we have a
m=l,
.2.3.4
= 13,
the
= 13,
k4, whence U =
a 2 :=13,
ra=l, r=l, s=l,
is
=rl3,
and the probability required becomes, on
=r 24,
what
substituting these
numbers
in the formula (28),
13.13.13.13_
_
X 52.
51 .50.
20825""
21.97
49""
n<
The odds against this event are nearly 8 to 1.

30. The following question, proposed by Huygens, and
solved by
Demoivre and Bernoulli (Ars Conjectandi,
p. 59)>
problems now under consideration.

urn contains 12 balls, of which 4 are white and 8
belongs to the class of
An
black.
Three gamesters A, B, and C agree that the first

shall draw a white ball shall be the winner of
who, blindfold,
the stakes.
third,
They
fourth,
also agree that
and
replaced in the urn.
so on
It is
probabilities of winning.
A shall draw first. B second,
and the
proposed
balls
drawn are not
to find their respective
57
Here the play terminates as soon as a white ball is drawn,

it must therefore terminate with the 9th trial, if not
and
sooner, inasmuch as, after 8 black balls have
been drawn,
the urn will contain only white balls, and the probability of
drawing a white
The
ball at the
next
trial will
be solved,
question will therefore
probabilities of the play ending with the
games
of
its
respectively,
the
1st, 4th,
sum of the
and
and
sum
the
2d, 3d, 4th, &c.
1st,
probabilities
7th, for the probability of
probabilities of its ending with
the 2d, 5th, and 8th, for the probability of

the
certainty.
we determine
and take the sum of the
ending with the
A s winning;
become
if
of the probabilities of
9th, for the probability of
its
Cs
B s winning
and
ending with the 3d, 6th,

winning.
For the sake of rendering the solution more general, let

a be the number of white balls in the urn, b the number of
black,
and let a -f- b=c.
The probability of drawing
a white
ending with the
first
probability of the play ending with the second
trial
ball at the first trial, or of the play
a
trial, is
The
is
then
compounded of the probability of a black
at the first trial,
and a white
at the second
ball
;
being drawn
and the proba
a
bility
of both events (26)
The
is
..
probability of the play ending with the third
trial is
compounded of three separate probabilities, namely, that a

black ball will be drawn at the first trial ; that a black ball
will
be drawn
at the third
at the
second
that a white ball will be
drawn
and the probability of the concourse of these
In general the probability of a black ball being drawn in
58
trials
!)(&
I,(b
successively,
and a white
(br + 2)a
2)
r-r?
c(c
l)(c
2)
substituting tor #.
xih
b.
is
and
a?+l
(c
c in this formula the
ball at the
numbers proposed by Huygens, we

1st, 4th, and 7th trials, or the pro
obtain in respect of the

bability in favour of
8-7.6
12
+ 12. 11
A,
8.7. 6.5.4.3
4
"
.10* 9
"12.11 .10. 9- 8. 7
in respect of the 2d, 5th,
and
in respect
bility in
and 8th
trials,
of the 3d, 6th, and 9th
4 _
"6"
77
""165
or the probability
trials,
or the proba
favour of C,
8.7.6
.5.4
12.11*10"*" 12. 11. 10.9.8
8.7.6
4_
"*"
.5.4.3.2.1
12.11.10.9.8.7.6.5
4 __ 35
*T~~
165*
The chances
in favour of
A, B, and C, are therefore pro
numbers 77, 53, 35, respectively.

If the condition of the play had been that the ball was
portional to the
be returned
to the
urn after each
trial,
to
the chances in fa
vour of the three gamesters would have been easily found

by the formula (12) to be respectively as the numbers 9>
6,
and
4.
MATHEMATICAL AND MORAL EXPECTATION.
SECTION
IV.
OF MATHEMATICAL AND MORAL EXPECTATION.

v 31. In the theory of
probability, the
term expectation
is
used to denote the product found by multiplying the value

of a casual benefit into the probability of the event on
which
nefit
it is
contingent taking place.
But the value of a be
be estimated either with respect to
may
iis
absolute
amount, or to the amount of relative advantage it affords the

individual who receives it. This consideration has led to a
distinction
between mathematical and moral expectation.
When we
place the circumstances of the individual entirely

out of consideration, and have regard merely to the abstract
or absolute value of the benefit, the product of
by the probability of obtaining

tation of the individual
it is
its
amount
the mathematical expec
but when a relative value
is
as
signed to the benefit, the product of this relative value

the probability of obtaining
tion^
because
it is
it is
called the
by
moral expecta
estimated by certain moral considerations
respecting the circumstances or fortune of the individual in
whose favour the expectation exists, on the principle that a

sum of money which may be relatively of very little import
ance to a
man
in possession of a large fortune
great importance
stanced.
tion.
We shall first
who
may be
of
favourably circum
mathematical
the
consider
expecta
to another
is
less
CO
A and B to engage in play let/> be the pro

A s winning a game, q the probability of B s win
32. Suppose
bability of
ning
and
it,
By
game.
sum of money staked on
s a
the issue of the
the definition, the mathematical expectation of
A is ps, and that of B is qs. Now if we suppose these expec

be purchased by A and B, the sums they ought
tations to
respectively to pay for them, or in other words to stake on
the issue of the game, must be proportional to their respec

tive expectations, in order that they
may
play on equal terms.
Let therefore a be the sum staked by A, and b the sum

we have then ps qs a b, and consequently
staked by B,
pbqa.
is
the
Now
: :
suppose a~\-b=:s, or that the
amount of the
pects to gain, and
stakes
p is
then, since b
is
sum played
the
sum
the probability of his gaining
the mathematical value of
As
expectation of gain.
for
A ex
it,
pb is
In like
manner qa is the mathematical value of B s expectation of

gain. Hence it follows, that when the sum staked by each is
proportional to his probability of winning, the mathematical
expectations of the two players are equal
so that after the
stakes have been placed, and before the event
is
decided,
they might exchange places without advantage or disadvan

It follows likewise, that since the
tage to either.
the one must gain
is
just that which the other
sum which
must
the
lose,
product qa, which is B s expectation of gain, may be regard

ed as A s expectation of loss or (if taken with a negative
;
sign) as part of
pb
Butj?>
qa.
the event
his
A s whole
is
expectation, which then becomes
^#=0; whence
decided
is
the condition of
A before
not altered by the circumstance of
having staked on the issue of the play.
33. This conclusion at

it is
first
sight appears paradoxical
certain, that after the stakes are placed,
gain the
sum
b or lose
and therefore
for
A must either
his fortune will
of

necessity either be increased
stake, or diminished
tion
by the gain of his adversary
depends on theorems which

that though
sults,
The explana
afterwards be demon
by the loss of his own.

will
strated relative to the repetition of
lose or gain, yet
61
in
a single
trial
on multiplying
trials,
from which
re
it
the player must either
sufficiently the
number of
games, a probability will at length be obtained, approaching
we
as nearly to certainty as
please, that the
sum gained
or
the long run will not exceed a certain given fraction
lost in
(which may be as small as we please) of the whole sum

staked, provided the play is undertaken on terms of mathe
But
matical equality.
zard
is
easily
this indefinite repetition
be imagined,
in
of the ha
and innumerable cases may

which an individual will be guided
practically impossible
by other considerations than the mere mathematical value

of the expectation in undertaking or declining a
risk.
person of moderate fortune would scarcely be persuaded to

risk L.500 for the expectation of gaining L.5, though the
chances might be 100 to 1 in favour of the event which
would produce that sum
but numbers would be found wil
ling enough
pay L.5 for the expectation of gaining L.500,
In both cases,
the chances being 100 to 1 against them.
to
however, the expectation would be purchased at
its
real
According to the formula of mathematical

man whose sole fortune consists of a lot
the
expectation,
has an equal chance of turning up a prize
ticket
which
tery
abstract value.
of L.20,000 or a blank,
tion as
he who
is
an equally advantageous posi

in possession of L.I 0,000 ; yet no man
of ordinary prudence,
would hesitate as
Common
loss of
to
sense will
if
is
in
offered his choice of the
two
states,
which he ought to give the preference.

prevent a man from risking a sum, the
which would be attended with great privations, even
MATHEMATICAL A^ D MORAL EXPECTATION.
when, mathematically speaking, the chances are consider

It is also obvious that two individuals
ably in his favour.
whose fortunes are very unequal cannot engage in play with

the same advantage, although the chances in favour of each,
in respect of a single game, are precisely the same.
The
one who has a large fortune can repeat the hazard so often
as to obtain a probability almost equal to certainty that his

loss will
not amount to any given
who cannot
continue the play
of being ruined.
It is
sum
whereas the other,
of
in case
loss,
runs the risk
thus evident, that in a multitude of
cases the abstract theory of probability
is
not alone sufficient
an expectation, and that in dealing with

an individual must be guided to a cer
events,
contingent
to give the value of
by considerations of relative advantage.

34. Various hypotheses have been imagined for the pur
tain extent
pose of reducing such relative or moral considerations to ac

but that which appears the most natu
curate calculation
;
ral,
and applicable
to the greatest
number of cases,
consists
in supposing the relative value of any infinitely small
to be directly proportional to
its
as the fortune of the individual
receiving
it.
who
This principle was
has an expectation of
first
proposed by Daniel
Bernoulli in the Petersburg Commentaries (vol.

there applied by him
sum
absolute value, and inversely
to the solution of a
v.),
and
is
number of ques
tions of great practical interest.
Let x be the absolute value of the
capital, or, as
it is
de
nominated by Laplace, the physical fortune, of an individu

al
ral
then, according to the hypothesis of Bernoulli, the
mo
advantage which he derives from an infinitely small incre
ment of fortune
dx>
is
measured by the expression c
being a constant to be determined by the nature of the ques-

tion.
if
No\v,
we suppose
63
the physical fortune to arise from
the accumulation of the elements dx, and denote by
moral value of the
relative or
or physical value
is
a?,
r ax
dx
yl
To determine
we
shall
=c
log.
the constant,
has a given value
the
fortune, of which the absolute
have
x -f- constant.
we may suppose yo^ when x
log. a), or
c log. a-f-constant,
this gives
/v
whence
yc
(log.
y=c
log.
and
is
it
to
be observed, that those values of x and yean never become

negative, for as Bernoulli has remarked,
son
who
is
it is
only the per
of
that
can
said
be
to
hunger
dying
possess ab
mere pos
a moral advantage, to
solutely nothing. In every other circumstance the
session of existence
which, however,
it
may be accounted
would be absurd to attempt
to assign a
nu
merical value.
From
35.
the above formula,
it is
easy to deduce a nu
merical expression for the value of a moral expectation.
Let
a be the
original fortune of the individual, and
sums
be received on the occurrence of certain contin
to
gent events, E, F, G, &c.
This being supposed,
a,
/3,
if the
y,
&c.
event
happens, the absolute fortune of the individual becomes
a-j-a,
and
mula,
is
comes
c log.
its
relative value, therefore, according to the for
c log. -f/3,
to
If F happens, his absolute fortune be
which the corresponding
and so on.
Now,
let
cessarily happen),
is
the probabilities of the
events E, F, G, &c. be respectively p, q,
p _j. q jf r 4. &c.
relative value
r,
&c. (assuming
so that one or other of the events will ne
and let
Y represent the relative fortune of
MATHEMATICAL
64
AtfD
MORAL EXPECTATION.
the individual arising from his expectation, then, since the

value of a benefit in expectation is equal to the amount of
the benefit multiplied by the probability of obtaining
it,
we
have
Y=c
Let
{?
also
formula,
log.
-*+r
we have
we
^log.
Y=c
On
to
then,
by the
comparing these two
-K
+r
"og.
log.
&c.
numbers,
g
+ + r-f- &c. =1,

X=(a + a)(a + /3)(a + y)
therefore, since
+ &c. }
get
and on passing
X
log.
log.
denote the absolute value of
values of Y,
log.
--a +fJ
log.
/?
</
In this expression
r
,
&c.
denotes the absolute value of the
and of the expectation added together if,

therefore, we deduct a from X, the difference will be the
value of the expectation, or the sum which, if it were to be
original fortune
received certainly, would procure the individual the same

relative advantage as his expectation.
36. If the
in
sums
a,
&
y,
&c. are supposed to be very small
comparison of a, so that quantities of the order
be neglected, the preceding equation becomes

-i f
whence, since p-\-q-}-r-\- &c.z=l,
&c.
may

Deducting from
the remainder
this the original fortune a,
pa~\-q[B-\-ry--8zc.
is
65
the value of the expectation, or the
sum
moral advantage. But the value of the ma

thematical expectation of the benefits a, )3, -y, &c. of which the
equivalent to the
probabilities are respectively p, q,

-f-&c. (31), therefore,
when
r,
&c.
is
also pa-{-qj3-{-ry
the contingent benefits are very
small in comparison of the original fortune, the moral ad
vantage and the mathematical expectation are sensibly the

same.
37.
From the formula X= (a -f a)p (a + $y*( a -\-y} r
&c. Ber
deduces the consequence that gambling or betting is

attended with a moral disadvantage, even when the chances
noulli
of gain or
To shew
fortune
mathematically speaking, are perfectly equal.

he proposes the following question. A, whose
loss,
this,
100 crowns, bets 50 crowns with B, on the issue
is
of an event of which the probability is ^, on these terms

if the event
happens, A is to receive from B 50 crowns ; if
:
he
it fails,
lue of
As
event
is
is
to
pay
50 crowns
what
the relative va
is
fortune, after undertaking the bet,
decided?
and before the
we have a=:100, a 50,

r=0; and the formula (35)
In this case,
50,7=0; also/?=J, <?=J,

becomes
x
whence
X=^/150 x50=87
50)
(100
and, consequently, the con
A is
worse by 13 crowns than it was before he ha

zarded the bet.
The moral disadvantage is therefore equi
dition of
valent to this sum, though the terms of the play, according

to the mathematical theory, are equal.
38.
easily
The
conclusion arrived at in this particular case
shewn
of the player,
Let a be the
to
be universally
his probability of winning,
true.
is
capital
his probability
G6
sum
of losing, and s the
In order that he
at stake.
may
play on terms of mathematical equality, the part of the

stakes contributed by himself, or the sum which he can lose,
must be ps (32), and the part contributed by his adversary,

which he may gain, must be qs. The equation in
or that
(35) therefore becomes
X= (a
and
if it
-j-
qs)P
can be shewn that
X (aps)i
this value
will follow that his condition is
of
is
less
rendered worse
than
in
it
conse
quence of having staked on the game. Now, dividing by

and taking the logarithm of both sides of the equation, we
,
=jo
get log.
rential of
log.
-}-
which (making
-f-
log.
s variable)
the diffe
1
J,
is
X
=pqds
d\og.
^
But the second
therefore
log.
rithm of X-f-a
all cases,
side of this equation
X-j-a
is
is
negative
negative,
and
therefore, the bet,
if
is
evidently negative
consequently the loga

must be less than a. In
on even terms, produces a
moral disadvantage.
39-
Another consequence deduced by Bernoulli from
theory of moral expectation,

kind
is
exposed to a
is,
this
when property of any

it is more advantageous
that
risk or hazard,
independent of each
expose
the
to
whole
at
to a single risk, al
than
once
other,
expose
in
of
loss
both
cases precisely the
be
though the probability
to
it
in parts to several risks
mer
To prove this, he takes the following example.
chant has a capital of L.4000, besides goods of the value of
same.
L.8000, which must be transported by sea.
The
lity of the loss of a vessel in the voyage being
-j^,
probabi
let
it
be
MATHEMATICAL AXD MORAL EXPECTATION.
67
proposed to find the value of the moral expectation of the
merchant
vessel,
in the case of the
and
and the other half
vessel
goods being embarked
also in the case of one half being
in a single
embarked
in
Supposing the
in another.
one
mer
chandise embarked in one ship, the absolute fortune of the
merchant
will
be increased
to L.I 2,000 in the event of the
\md will be reduced to L.4000 in the

being lost. The probability of the first of these
safe arrival of the ship,
event of
events
its
T%, and of the second
is
J5
therefore his absolute
fortune becomes, in virtue of his expectation,
X= 10751.
whence
Deducting
his other capital,
L.4000,
there remains L.675 1 for the value of the moral expectation
of the venture.
in respect
Let us next suppose the merchandise embarked in equal

In this case there are three compound
parts in two ships.
events to be considered,
1st,
the probability of which

arrive in safety
which, as
1
3C
TOO
To
capital
= L.8000
With
Too*
ê
if
;
îe ^ rst
these
if
lost
ê
St
2d,
One may
the probability of
P r k akility f wm ch
it
will
it
be L.4000 -f L.4000
will
be only L.4000.
^ x (8000)^ X (4000)T K
is
êse events happen, the
the third happen
000)
(ll)is2Xy^X
numbers the formula becomes
X=(12
become L.4000 + L.8000=
the second happen
and
two ways,
in
of the merchant will
L.12,000;
y^
and the other be
^ Ot ^ ma^
vessels may arrive in safety
is
may happen
==
To
it
Both
KT1CAL AM> MORAL KXTECTATION.
X=
w hence
to
<1
1
0:U. Deducting his other capital, which was
no risk, there remains 1*7033 for the value of the
1
This sum exceeds the former by I-.2S2
moral expectation.
and
that
soning,
in
numb*
found b\ following the same process
ea>il\
is
it
the value of the
ma
:he mor;,.
its limit,
thematical expectation, or
40.
rea
pro|>ortion
and approaches
ed,
e.f
which
f!
of LjBOOOsLJTVXK
The theory of moral expectation enables us likewise to
assign the circumstances in which
it is
advantageous or
otherwise to insure property against particular hazards.
There an* three principal questions to be considered ii
.
The amount of premium the insur

ed may pay without disadvantage 2. The ratio of his for
rence to
tins subject
1.
tune to the value of the stun exposed to risk, in order that

may be advantageous to insure at a given premium ; and
The
capital
it
3.
which the insurer or underwriter ought to pos

may insure a given risk with probable
order that he
se**, in
advantage to himself, and safety to the insured.

1 ,ei * be the value of a
cargo which a merchant emb.,
A ship,
the probability of the safe arrival of the vessel, and
n his capital independently of s.

the
premium
for insurance is
$;
mium by y* then .vis the sum the

reaches
its
lose if
does not
it
destination in safety,
;
The mathematical
we denote
for, if
insurer will gain
and
and by the theorem
for the
y) ; whence, since
merchant insures
expectation ;>y=r q(*

\ f,
therefore, the
tune becomes a
it is tlie
it
will
value of
be advantageous or
is
greater or h
oil
the vend
sum
+ </=!. y =.$.*.
^sol
ute for-
he does not insure,
r
equation X=r. (^-|-.<)
Hence
insure according as
;>
alne of
mat hen
.,nd if
qx
in the
the
*/ is
if
the pre
-^->^ ./.
\on
the logarithm
M
"i
tin
III
in
Id
lil
li
mli
ill.
d"
"i
M.
\ll.ri.|\|
M..K
<
"i.il
MM)
(/
id",
pi
,,.
ld".n illim
lli,
quh
flu
,,|
"
i,
id. di /. id
i,/
)-f-y log//,
i.Mn /
.di
equi
/ "-|--v
UK
...
<|ui nll\
in
lli.
nine
li.ii
,,
infc
//A
in in.
(,,
/>
dl
.In,
lli,
VA
..id
ii
he
If
To
.1.
I.
.1
MI
"
m,
ii.mi
ii
ululr
l,
m,
(!i;oi.l,
<
;midiiiii nl
m..\
ijii.it
mui-iillif
KHI
01
I
||
m,
||
in. ir.i-.rd
i"
mill,, m.ili
lli,
by
ii
In pruefice tbe
pre
loner.
leHM
MM.I pfiyH HIOI C
than
(/v-f..r,
but greater
llliUl tllC IHll-
he gaiim a moral
advantage by
rink,
IK n
be
In
"
,r
<
,m
fortune
rink
(I.,-
I,,.,
-.
in
(Hi-dii
..I
md
il,
tuunple,
/
M
let
ftfttr
on
prc^rniuni de-
being imured
hiniMelf,
value of
ili,
In
m..ti,
:\
A
.
ill.
.jiLilKiii
:i|ni.il
in-
O|
i.d ..--.
II,
be the
.unoutit*; then, tbe other capi-
,,
d,
let c
ion,
i|in
being
il
or noi.
..
mm
II,,
m:iK d hdin HK
il
Let Ul
ton
<
,/
.i.K.ml.i",
fortuni
.1
lli,
mud
m mm"
("| M
beCORK
nidi.il di
In-
in
.,!...
nd
nil
ulnlr
dl id,
ih,
idi
id,
Kill.
Hdl\
m.md, d
in""
m,
n,
i.,!,v,
lli.
will In-
\.diK
..I
-HI
id,
illioiil
i,
-.-IN
IK MI. in,
M,.,,,
ii
.idv.inl.i",
md
mium m;i\ InHun V

dlh.il
th<
\\
i.,
,/
p. ls
"i, .id
if li
\\
than
ifc
i.il.
I,.
i,
/,..
", n,
nd, d
.ill,
gnl
in
:.
cdiild .illdid
ih. in
in
the
ther( fbri
" i"",.
propel
.,,<
we
s)"<i\
pdHHeHH
to
//
in
it.M
value
be deternlluM
have
order that
il
him whether he
tbe value of the tncr-
=L.HOO, and
/>=};;.
Tin-
dim
nd iu
appro
:5043.
It
follow
MATHEMATICAL AM> MOK.M.
rO
t:
therefore, that unless his other capital
would
t)e
disadvantageous
(which
The
is
third question, the
amount
premium
its
being
lost.
becomes
The
it
although the
amount
L.300.
of capital the underwriter
his capital will
in
become
of the vessel arriving in safety, and b
fore
L.5043,
the same way.
After accepting the risk of the
Let b be his capital.

t\
to
the mathematical value of the
determined precisely
tor the
F.CTATIOX.
^ X L.UU)00=L.500) by
to
ought
to neglect insuring,
premium demanded exeeed

risk
sum
>
the case
h-\-e in
s-{-e in the case of
formula of the moral expectation there
X=(&+ e)p (b
s-f-f)
and
in
order that there
may be neither advantage nor disadvantage in undertaking

the risk, this value of X must be equal to his original capi
tal,
Supposing, therefore,
l>.
,v,
p, q> to have the
e,
nifications as above, the equation
termined
Unless,
is
b=(b+8W)$*(b
therefore,
from which b
9200)
25
,
is
same
sig
be de
to
whence b= 14243.
the capital of the insurer amounts to
L.I 4,243, there would be a moral disadvantage in undertak
a cargo worth L.10,000 tor a pre

ing the risk of insuring
mium of L.800 ; and it is easy to see, that if a smaller pre
mium were demanded,
On making e=600,
the capital ought to be
(which
still
still
greater.
exceeds the mathematical
value of the risk), the value of b becomes L. J; 1^878.

it follows, that a company possessing a large capital
lence
may
not
in speculations which might prove

only with safety engage
ruinous to another whose resources are more limited, but
1
even derive from them a sure profit.
41. The theory of moral expectation which we have now
been considering had

1
See the
Thhrie
its
origin in a
problem proposed by
Commcntarii Acad. AfropofitaMB,
dcs Prob. p. 4. 3 2
torn. v. ; Laplace,
Lacroix, Trait J Ekmcntaire, p. 132.
MATHEMATICAL AXD MORAL EXPECTATI

Nicolas Bernoulli to Montmort^ which, from
71
having been
its
discussed at great length by Daniel Bernoulli in the Peters
burg Memoirs* has been usually called the Petersburg prob

lem. It is this A and B play at heads and tails. A agrees
:
to pay B 2 crowns if head turn up at the first throv,

crowns if it turn up at the second, and not before 8
if
it
turn up at the third, and not before
2 n crowns
turn
if it
up
at the nth throw,
and, in general.
and not before
B s expectation ? Here the proba

of head turning up at the first throw is ^ the proba
of its turning up at the second, and not at the first, is
required the value of

bility
bility
Xi
-:
the probability of
or second, and of
= J,
and
Hence
so on.
16
2"
its
l
turning up at the third,
the probabilities of
-^ -p
the mathematical value of
Y X2+-5-X4+ -i
as
no
not turning up either at the
X-^xi
receiving 2,
crowns
are respectively
Now.
its
limit can
Bs
consequently (31)
expectation
is
X-+^Xl6... + _Lx2
be assigned to
n,
crowils .
inasmuch as
it
is
head may not turn up till after a very great,

or any assignable number, of throws, this series, of which
each term is unity, may go on for ever, and consequently
possible that
the value of
Bs
expectation becomes infinite.
Yet
it is
ob
no one would pay any considerable sum for the

This disagreement between the dictates of
expectation.
vious that
common
sense and the results of the mathematical theory,
appeared to Montrnort to involve a great paradox ; although

the question differs in this respect from no other question
of chances in which the contingent benefit
is
very great, and
72
the probability of receiving
it
very small. If the play could
be repeated an infinite number of times, B might undertake
to pay without disadvantage any sum, however large, for his
A result, however, more in accordance with or
expectation.
dinary notions,
is
obtained from the principle of Bernoulli.
Let a be the amount of
Bs
fortune before the play begins,
x the value of his expectation,

turn up at the
first
throw,
or the
Bs
3
j
and
events being respectively

for the moral expectation
Now
the
if
But the
so on.
z=a
A in con
If head
x.
fortune becomes z~\-2;
the second, and not before, 2-J-2

before, ;z-J-2
sum he pays
and make
sideration of the agreement,
-^-
probabilities of these
........
if at
a t the third, and not
the formula
2i
becomes (35)
sum which B ought
to
pay
will
be determined by
the value of his moral expectation, after the bet,
making
and before the play begins, equal
we have therefore a=X, that is,
The
/
22(
term of
general
1+
z \ 2n
J
this
the equation
to his previous fortune
series
may be
and since the logarithm of the
first
2n
being (z-J-2 )
put under the form
factor of this expression

2 log
we have
2,
=2 log 2+
log
log(l
+y)+
from which a value of z may be found by

for
the
trial
a.
Suppose z=lQO ;
10 terms of the series there results
whence
(since
x=a
-j-
log
and error
on computing
any given value of

first
73
z)
#=7
89
that
is
a= 107-89,
to say, if
sessed only 100 crowns before beginning the play,
it
pos
would
be morally disadvantageous for him to risk 8 crowns for the

expectation, although its mathematical value be infinitely
If we suppose z
1 000, the sum of 1 1 terms gives
great.
a=:1011, nearly; so that if B possessed a fortune of 1011
crowns, the value of the moral expectation would, to him,
be about
1 1
crowns.
It is scarcely
necessary to remark, that the results de
duced from the principle of Bernoulli are of a character

widely different from those which are calculated according
to the
mathematical expectation.
The
latter gives the pre
cise value of a contingent benefit, without
any assumption or
hypothesis respecting the personal circumstances of the indi

vidual
who may gain
or lose
it
of relative advantage, of which
whereas the considerations
it is
the object of Bernoulli
theory to take account, are entirely arbitrary, and by their
very nature incapable of being

computation.
or appreciate,
It is evidently
all
made
the subject of accurate
impossible to have regard
the circumstances which
may
to,
render the
same sum of money a more important benefit to one man than

to another; and consequently every rule that can be given for
the purpose must be liable to numerous exceptions.
principle,
however,
is
thus far valuable, that
it
The
gives in the
most common cases a plausible and judicious estimate of
74
the value of things which are not susceptible of exact ap

preciation
and
it
mitted to analysis.
has the advantage of being readily sub
different principle, proposed
by the
celebrated naturalist Buffon, consists in making the value

itself
of a casual benefit, instead of
its infinitely
small ele
ments, inversely proportional to the fortune of the expec

tant ; but as this hypothesis has seldom been adopted, it is
unnecessary to discuss
it
in this place.
FUTURE EVENTS DEDUCED FROM EXPERIENCE.
SECTION
75
V.
OF THE PROBABILITY OF FUTURE EVENTS DEDUCED FROM

EXPERIENCE*
42. In the preceding part of this article
sumed,
in every case, that the
it
has been as
number of chances favour
able and unfavourable to the occurrence of a contingent
event
is
known a
priori,
and consequently, that the proba
bility of the event, or the ratio of the
cases to the whole
number of
But
lutely determined.
number of favourable
cases possible, can be abso
numerous applications of the
in
theory of probabilities, and these, generally speaking, by far

the most important, the ratio of the chances in favour of an
event to those which oppose it is altogether unknown \ and

idea of the probability of the event except
we can form no
ing from a comparison of the

it
number of instances
has been observed to happen, with the whole
instances in which
it
in
which
number of
has been observed to happen and
fail.
In order to assign the probability of a contingent event in
such cases,
it is
necessary to consider
or combinations of circumstances
possibly be produced,
and
to
all
the different causes
by which the event could
determine
its
probabilities suc
cessively on the hypotheses that each of these causes exists

to the exclusion of all the others.
The comparative
facili
which these hypotheses give to the occurrence of the

event which has actually arrived, will then enable us to deties
PROBABILITY OF FUTURE EVENTS
76
termine the relative probabilities of the different hypothe
and consequently
ses,
sum
is
their absolute probabilities, since their
necessarily equal to unity
and when the probabili
ties of the different hypotheses, and of the occurrence of the

event on each hypothesis, have been determined, the pro
bability of the event occurring in a future trial will be found
by the methods already explained.

Taking a simple case, let us suppose an urn to contain
4 counters, which are either white or black ; that the num
43.
ber of each colour
unknown, but
is
in four successive
draw
ings (the counter drawn being replaced in the urn after

each trial) a white counter has been drawn three times, and
a black one once

bability of
and
let it
be proposed
to assign the pro
drawing a counter of either colour at the next
trial.
In the present case three hypotheses
1st,
2d,
may be formed
number of white and black counters
lative to the
in
re
the urn.
The urn may contain 3 white counters and 1 black

It may contain 2 white and 2 black
3d, It may con
tain
white and 3 black
for a counter of
ing been drawn, the other two possible
each colour hav
cases,
namely, that
black, are excluded
by the observa
be
the probabilities respectively of
Now,
p lt p2 pv
drawing a white counter on each hypothesis, and q^ q.2) q3) the
they are
all
white or
tion.
let
probabilities of
thesis to
be
all
,
drawing a black.
true, or that the
Supposing the first hypo

compound event which has been
observed was produced by the cause indicated by that hypo

thesis, we havejpjrrj, <7i=i ; and the probability of the ob
served event, or that 3 white counters and
drawn, (12)
pt\,
gives
q3
>
is
The second
4pf gr^fj.
whence 4p23 q a =^-
3 =J, 23=|> whence
4/? 3
The
3=^.
black would be
hypothesis gives
third hypothesis
The
probabilities
DEDUCED FROM EXPERIENCE.
77
of the observed compound event, on each of the three hy

and the
|,
J, fa
potheses, are therefore, respectively,
;
question
now
how
arises,
are the probabilities of the differ
ent hypotheses to be estimated

priori, for
determining
As we have no
thi> question,
data,
we must assume
probabilities of the different hypotheses to
the
be respective
ly proportional to the probabilities they severally give of
compound event
the observed
in other words,
assume the probability of any hypothesis

according as
it
to
we must
be greater or
affords a greater or smaller
less
number of com
binations favourable to the event which has been observed

to take place.
if
Thus,
C and Qbe
from which an observed event
and C furnishes 20
two independent causes

may be supposed to arise,
different combinations out of a given
num
ber, favourable to the occurrence of E, while C\ furnishes
only 10 such combinations out of the same number,

turally infer that the probability of the cause
rated to produce E,
is
we na
C having ope
twice as great as the probability that
the event was produced by the operation of the cause
Applying
lities
this principle to the present
C^
example, the probabi
of the three hypotheses are respectively proportional
to the three fractions
|,
fa, or to the
numbers 27,
16,
no other hypotheses are admissible, the sum of

their probabilities must be unity ; therefore, making sr the
3
and
as
probability of the
TT3
first
that of the third,
hypothesis,
we have
16
27
i=46
44.
w2 that of the second, and
Having found the
*- 2
=46>
^=46-
hy
potheses, that of drawing a white counter at the next trial

is
obtained without difficulty
for according to
what was
78
shewn
in (9), the probability of this

simple
equal to the
sum of its
event must be
probabilities relative to the different
hypotheses, each multiplied into the probability of the hypo

thesis itself.
Now it has been seen that, on the first hypo
of drawing a white ball is | on the

second f and on the third ; and that the probabilities of
thesis, the probability
the hypotheses are respectively J j, f -fa ; therefore the

probability of a white counter being drawn at the next
,
trial is
-^
^+-*^+l
X 46 T 4 x-546"184
1
5x
T4
1 46
In like manner, the probability of a black counter being

drawn at the next trial is
IX
4
27
46
4_JL X 1?
46
+T
+A
T X JL46~~184
68
and the sum of these two fractions
is
unity, as
it
ought to be,
drawn must necessarily be white or black.

45. The reasoning which has been employed in this par
since the counter
ticular case
is
Let
of general application.
event, simple or
unknown, but which may be ascribed
C2 C3
causes, C,,
happened, are
E be an
observed
compound, of which the particular cause
all
CM
to
is
any one of the
which, before the event has
equally probable, and such that the opera
any one of them excludes that of the others, so that

the event E is produced by one of them alone, and not by the
joint agency of several of them. Let the probabilities of the
tion of
observed event E on the hypothesis that it has proceeded

from each of those causes be respectively P,, P 2 P 3 ,..,P W
so that if the cause, for instance, C were the true one, the
,
probability of the event E, previous to the observation,
would be
and
let
the probabilities (as determined by the
79
event) of the existence of the different causes be respective

ly
v lt w 2
3,
.........
wn
From
the principle laid
down
in
the preceding paragraph, namely, that the probabilities of

the different causes or hypotheses are proportional to the
probabilities they respectively give of the observed event,
we have
whence, making
ing that
P i -f P 2
1 -j-z3- 2 -j-
+P
...... -f-
......-f-wa
=1
P n = 2P,., and observ

(since
assumed
it is
no other causes than those specified from

which the event could arise), we have
that there are
whence
it
appears that the probability of each hypothesis re
specting the cause of the observed event
is
found by divid
ing the probability of the event on the supposition that that

particular cause alone existed,
ties in respect
of
all
probabilities of a future event
with
by the sum of its probabili

Let us now assume the
the causes.
(which
or different, but depending an the
may be
the
same
same causes)
in
respect of the several hypotheses, to\}e,p 19 99 p s j ..... ./?* ;

so that if the particular cause C, be the true one, the proba
bility of
of
all
is
pi and let
;
U be the probability of E in respect

n will be equal to the sum of
the causes, then by (9)?
the probabilities
p^p 2 p 3 ......pn
,
relative to the different
hypotheses, each multiplied by the probability of the hypo

thesis ; that is to say we shall have
or n=2p,w,, the symbol

ferent values of jo
Cp C 2 C 3
,
......
and
Cn
indicating the
sum of all
the dif
in respect of the different causes
80
46. It
may be worth while
not here used in
its
to
remark that the word cause
ordinary acceptation to denote the
is
com
bination of circumstances, physical or moral, of which the
In the sense we have

event is a necessary consequence.
used the term, the cause C is that which gives rise to the
determinate probability P, that the event E will happen
;
but so long as
this probability falls short
of certainty,
existence also implies that of another probability,
that the contrary event
will
occurrence of
word cause
is
and
happen.
C would
the existence of the cause

it is
If we
its
P,
make P=l,
necessarily involve the
in this particular sense that the
In the theory of probabi
ordinarily used.
the causes of events are considered only in reference
lities
to the
number of chances they
those events which they

sarily,
may
afford for the occurrence of
possibly, but
do not neces
produce.
The following example may serve to illustrate the

method of applying the preceding formulae.
An urn
contains n balls, which are known to be either white or
47.
black.
ball is
drawn
random and found
at
to
be white
required the probability of drawing a white ball at the
next
In
trial ?
this case, the
number of hypotheses
is n
respecting the contents of the urn,
that
;
for
may be made
we may sup
pose that it contained one, or two, or any number of white

from 1 to n, and each of these cases may be consider
balls
ed as a distinct cause of the observed event E.

causes or hypotheses be
C 15 C 2 C 3
pose the true cause was
C,,
On
balls.
event
is
whence
P =f
Let these
C n and let us
,
or that the urn contained
this hypothesis the probability of the
sup
white
observed
and therefore, making i succes-

sively equal to
1, 2, 3,....w,
we have
But the sum of this arithmetical
2P<
81
n-(!-{^-
series is
2-[- 3.. +);*.)
therefore
2,P=J(ra -[-!), and consequently,

Wf
which
is
_P _
the probability of the assumption that the event
proceeded from the cause

white
2?
~2P,~W!~
balls.
If
probability that
we suppose
all
or that the urn contained
C,,
n we have
the balls are white
sr B =r
and
if
for the
we
also sup
pose n 3, this becomes ^ ; whence if an urn contain 3 balls

which must be either black*or white, and a white ball be
drawn
that
at the first trial,
all
48.
it is
an even wager, after the
trial,
the balls are white.
Having found, from the observed event E, the pro
babilities of the different hypotheses,
mine the probability

white ball) at the next
of the event
Here two
trial.
selves ; according as the ball
we have now
is
to deter
(the drawing of a
cases present
replaced in the urn, or
them
is
not ;
or in general, according as the law of the chances remains
constant during the series of

1st,
urn.
trials
Let us suppose that the
ball
has been replaced in the
In this case the probability of the event
pothesis that the urn contains
say
or varies.
pf==.
found above,
But the
is
white
probability
therefore
zrt
balls, is
on the hy
;
that
is
to
of this hypothesis, as
/?^,-^
whence
82
the general formula (45)
Ze 2
2-
Now
n= 2/?, w,.becomes n=E -%-
Si 2 =2z(*+
But by the pro
2z.
perty of the figurate numbers referred to in (23), the sum

of the series of numbers obtained by giving i every value
from
to izzn in the formula
We
~
2
is
expressed by
therefore S?
have also as above
consequently
.
_n(
3
and therefore
2.3
2d, Suppose the ball which has
placed in the urn.
urn at
first
been extracted
is
contained
white
ing a white ball at the next
balls,
trial is
the probability of draw
----;
ft
that
is,/^^^n 1
and the probability of the hypothesis is the same as
and consequently 11=27^=
1)
in
\\ n (n+l}
2 ^ 2 "~ 1 ^
Now
be found by writing
the above expression for 2i(+ 1 ) whence
the value of 2i(i

for
in the for-
2i(t
not re
In this case, on the hypothesis that the
1) will evidently
~v
2z(z
i\
)=
v
(
w
l) w x
( - +l)
-
l
r
*uand, therefore, in this case
n_
y (*-l )*(*+!)_
A
-3
-(rc_l)ra(ra+l)
When n is a very large number, the ratio of 2n -j- 1 to

3, the value of n in the former case, does not sensibly
from f and therefore in both cases ri f
Hence it
follows, that if an event, depending on unknown causes, can
differ
happen only
one of two ways, and
in
it
has been observed
happen once, the odds are two to one in favour of
to
in the
happening
49.
The
same way
expression for
&
at the
its
next occurrence.
(45) was determined on the
in
supposition that previously to the experiments being made,
we
are entirely ignorant of the relative
sorts of balls in the urn,
numbers of the two
and have no reason
hypothesis more probable than another.
happen
to
know,
ferent causes
to suppose
one
however,
we
If,
previously to the experiment, that the dif
C 1? C 2 C 3
,
&c. have not
all
num
the same
ber of chances in their favour, or that the probabilities ojf

the different hypotheses have relative values, it becomes ne
cessary to introduce those relative values, in consequence
of which
Z3-
1?
w2
&c., will receive a modification.
Let us
conceive a number of urns, each containing balls of two

colours, black and white, to be distributed in n groups,
Aj,
Ag A
the
number of white
is
An
in
such a manner that the ratio of
balls to the
number of black
balls
the same in respect of each urn belonging to the same
group, and consequently that the probability of drawing
ball
in the
of either colour
group
it
is
may happen
the same from whichever urn
drawing a white
ball
be drawn, but
to
respect of the different groups
and
different in
let the probabilities
of
from each of the different groups be re-
84
P 1? P 2 P 3
spectively
are
and
a 1 urns
in the
let s
......
group A
Now,
2 in the
number of
the whole
-}-a 3 ...... -\-a n
Pn
T,
then, if we
let
us suppose there
group
A2
and so on,
urns, so that s^â l
make
s
rzXj,
=X 2
and
so
be the a priori probability that a ball drawn from

any urn at random, will be drawn from the group A x X 2
the probability it will be drawn from the group A 2
and,
on, A! will
it will be drawn from the

group
This
A,.
being premised, suppose a trial to be made, and
that the event E is a white ball ; the probability w of the
in general, X f the probability
hypothesis that the ball was drawn from the group A,
The a priori
found as follows.
ing drawn from the group

ally
is
X4
and
if
the ball
drawn from that group, the probability of
white
is
P,
is
its
actu
being
therefore the probability of both events

X
>,P
is
probability of the ball be
and consequently (45), w,r=-
is
P
=5-,
the symbol ofsum-
mation 2 extending to all the values of i from z= 1 to z=:w.

50. In the applications of the theory to physical or moral
may be
C 15 C 2 C 3 &c.
events, the different groups of urns here imagined
regarded as so
independent causes
many
by any one of which the event E might have been produced

W| is the probability that the event was produced by the par
ticular cause C,
had alone
and
X,-
is
,
f
the probability that the cause C,
is
if it
would have produced the observed event
be the
XP
2X
P,
the probability, previously to the experiment,
Q would
that
existed,
therefore,
efficient cause.
The formula
*r t -=:
shews that the probability of any one of
the possible causes (C
of an observed event
is
equal to the
product of the probability (P,) of the event taking place
if
85
that cause acted alone multiplied into the probability A, that
the cause C.is the true one, and divided by the

all
sum
(2A,P,) of
the similar products formed relatively to each ofthe causes
from which the event can be supposed to arise.

51. The formulae now obtained can only be used when
the
number of hypotheses
is finite
but in the applications
most frequently happens that an infinite

number of hypotheses may be made respecting the causes of
of the theory
it
an observed event, as would be the case
ample
if
the
number of balls
in the
in the
above ex
urn had been unknown.
In such cases, in order to find the values of
-sr
and n,
it
be
comes necessary to transform the sums 2 into definite in

tegrals, which is accomplished by means of the theorem
X dx, where X is a function of x. Suppose a ball
SXi^y^
to have been drawn a great number of times in succession
1
from an urn (the number in which is unknown) and re

placed in the urn after each drawing, and that the result
has been a white ball
times and a black ball n times, the
and thence the probability

a future trial will be found as
probable constitution of the urn,

of drawing a
w hite
r
Assume
follows.
ball at
the hypothesis that the ratio of the
ber of white balls to the whole number in the urn
and
& be
let
the probability of the hypothesis.
num
is
On
I,
this
hypothesis the probability of drawing a white ball in any

trial
is
x,
and that of drawing a black
ball
x,
and
m white and n
n
x) by (12). We have there
consequently, the probability of drawing
black in
m -f n trials is Vx m ( 1
fore for the probability of the observed
P=Uxm (l
whence
compound event
consequence of the above

x) ;
formula for transforming a sum into a definite integral
m
n
2P=U/â? (l x) dx{\3 being independent of x) and
therefore
in
86
"
p
SP
ar
*)"
(l
The
tion
is
value of the integral in the denominator of this frac

obtained by the usual method of integrating by parts.
Since
therefore
we getjx m +
In like manner
Continuing
this operation
x] becomes n
(1
(l
x)
n ~ l dx
n times, or
till
the exponent of
n=.0, the last integral will be
m + n +i
/>+&:=
m-f/z+l
therefore, collecting the several terms into one sum,
/-m/i
y.,7
y?
we have
1-1
1)(^_ 2) ......... 2.1
When #=0, all the terms of this series vanish, and when
x~\ they all vanish excepting the last therefore between
;
the limits #.=0 and ar=l, the value of the integral

last
term of the
series
when x
in that
term
is
that
the
is
to
say,
~(i+l)(f
For the sake of brevity,
let
the symbol [#] be adopted to
DEDUCED FROM EXPERIENCE*

represent the continued product
numbers from
87
3...X of the natural
whence by analogy [#+?/] will repre

sent the continued product of the same series from 1 to the
1
to x, 1
number denoted by x+y* Multiplying,

tor
then, the
and denominator of the above expression by
numera
.2.3...
...?wi=[m],
we
whence the
probability of the hypothesis, in consequence of
get
the equation above found, becomes
From
this value
of zr
we
are enabled to deduce that of n,
the probability of drawing a white ball at the next
(45) n=2zrp.
since
Now,
white balls in the urn

lours in the ratio of
is
to the
to 1,
the probability of drawing a
tuting
(1
xm +
(l
x) dx will evidently be obtained by substi
w-f-1 for
whence,
But the
n
x) dx.
n=2zrx
ofy^
of
whole number of both co
white ball is#; consequentlyj0=:#, and therefore
By
trial.
number
by hypothesis the
in
C xm
l
the expression found for
This substitution gives
observing that [m-\- 1]
-i-
=m-f n-}-2,
[m~\
= m-^
1,
and
we have
m+n + 2
1
This convenient notation
lias
been adopted by Mr.
De Morgan.
The
black
probability of the contrary event, or of drawing a
n=
ball, is 1
-n-le- 1
As
the numbers
and n
m-)--t-2
become
these two fractions approach nearer and
larger,
-- and 7?fc
nearer to their limits
Tfc
p and q of the respective events when
probabilities
of the number of white balls in the urn

balls as
is
the ratio
to that of the black
to n.
The
52.
which are the apri or i
probability of drawing
black balls in
n future
-\-
white balls and n
found
trials is
man
in a similar
ner, and the problem may be thus stated. E and F are two
contrary events, depending on constant but unknown causes;
and
has been observed, that in m-\-n-=.h successive in
it
stances the event
will
Assume,
of
occur
has occurred
required the probability that in
Fn
times and
Fn
times and
times,
-{-n =zh future instances,
times.
as in the last case, the facility of the occurrence
to that of
to
be
in the ratio of
to
we have
then, as before, for the probability of the hypothesis,
x m (\
lity
of
in the
whence the
next h?
x)
n.
on
this hypothesis the
next instance
probability of
trials
We
/?,
=
!
and that of F
is a?,
times
being denoted by
<!-*), making U
)+
Now
have therefore
and n times
we have
.
To
find
its
probabi
is
u/
a?,
in the
p=U
2 .3...h
m xl
,
...
for the probability of the
this hypothesis.
(12)
w=
xm
=
.
3...,,,
+*
~WM "
compound event on
probability
on the
infinite
89
number of hypotheses formed by supposing x to increase by

to a?=l, we have
infinitely small increments from x
n=2zrp=: J"Q
found,
it is
by
we
get
On
zrpdx.
n==U
O]
substituting for Tsp the value just
m m
ft x +
JQ
(\
n n
x) + dx, and
\_n~\
manifest that the value of this integral will be obtained
substituting
# m (l
m+m
for
m, and n-\-n
n
x) dx found above.
for
in the value
This substitution gives
whence we conclude
[^
The most probable hypothesis will be found by making

maximum, or its differential coefficient equal
the value of w a
to zero. Differentiating the equation
and making
=0, we get
QX
.
The most probable
ing the contents of the urn
7??(1
zr=
cc)
ww
*~
= nx,
xm (\
x}\
whence x
supposition, therefore, respect

is,
that the two sorts of balls are
same proportions as have been shewn by the previous

We shall have further occasion for these for
drawings.
in the
mulae
when we come
are large numbers.
to consider the cases in
which
and
BENEFITS DEPENDING ON THE
90
SECTION
VI.
OF BENEFITS DEPENDING ON THE PROBABLE DURATION OF
HUMAN
LIFE.
53. In applying the principles of the theory of probabi

lity to
the determination of the values of benefits depend
ing on
life,
the fundamental element which
to determine from observation
is
is
it
necessary
the probability that an in
dividual at every given age within the observed limits of
the duration of
for instance
will live
life,
one year
for
over a given portion of time,
when
this has
been determined
each year of age, the probability that an individual, or

any number of individuals, will live over any assigned num
for
ber of years,
is
easily
deduced.
that an individual A,
Thus,
whose age isy,
if
the probabilities
will live
3...X years, be denoted respectively by/^, />,

if </i>
<?
a>
9s
<?*
so
on
1, 2,
and
in respect
T/+ 1 years; r 1 ,r ,r 3 ...r^. the

respectof an individual whose age is y-j-2 years, and
in
;
z .*>px
denote the same probabilities
of an individual whose age
same
over
is
<J
then, since the probability
p2
which
has of living
obviously compounded of the probability/*,

of his living over 1 year, and of the probability q that, hav
ing attained the age y+l> he will live another year, we
over 2 years
is
have, by (7),
will live
/> 2
=/
91
Again, the probability
ps
that
over three years, being compounded of the proba-
PROBABLE DURATION OF HUMAN

he
bility /? 2 that
will live
LIFE.
91
over two years, and of the proba
having attained the age y-f-2 years, he will

we have /> 3 =r/) 2 r l =p q r
In like
r
*
so
on
that
and
so
the
;
lf
4 =/) i q^
probabili
i
bility r l that,
survive another year,
manner
ties jo 2 , j 3 , /> 4
/?*
are successively derived from j^, q lt
r 15 s lt &c. which are supposed to be the data of observation.

If a large
number n of
individuals, all
born in the same
year, were selected, and if it were observed that the num

ber of them remaining alive at the end of the first year is
nv
at the
third
end of the second year
and
ra
at the
fl ,
so on, then the probabilities
end of the
p l p 2 p3
,
c.
would be given directly by the observation, being respec

tively equal to the quotients
But the most
&c.
accurate observations of mortality are furnished by the ex
perience of the annuity and assurance

are not
made on an
isolated
where they
number, diminishing, and con
offices,
sequently giving a less valuable result every year, but on a
comparison of the numbers which, in a series of years, en

This observation
ter upoii and survive each year of age.
gives p 19
q^
r 1? st
&c, whence
as above, for every year of

54.
sums
The
to
pv p
l2 ,
lives,
p3
are found by
com
&c. with the rate of in
r= the rate of interest, that is to say,

for a year, and v= the present value of
L.I to be received at the end of a year,

r=l-t-(l-f-7*)
understood in
For
&c. are found,
money. Let
the interest of L,l
values of annuities on lives, and of reversionary
be paid on the failure of
bining the probabilities

terest of
p^ p 3 p 4
life.
we
shall
then have
Now
this
an annuity, payable yearly, is always

sense, that the first payment becomes due
further details on this subject, see
MORTALITY,
vol. xv. p.
550.
92
end of a year after the annuity is created. Suppose

then the annuity to be L.I, the present value of the first
at the
were to be received
if it
payment,
sum
alive at
certainly,
is
but the
contingent on the annuitant being

the end of the year, the probability of which we sup
receipt of this
pose to be p\
is
therefore (7) the present value of L.I sub
ject to the contingency,
is
In like manner, the present
vp r
value of L.I to be received certninly at the end of
years
but the annuity will only be received at the end of

the #th year if the annuitant be then living, the probability
is
if
of which ispx
lar
payment
therefore the present value of that particu
is
Hence
v*px .
of the annuity, or the
sum
the future payments,
all
2 including
sc
parts,
The
have
shall
If the annuity be
viously =. a"2v p x
55.
we
denote the present value
hand which
values of a? from x
all
which p=0.
if
in
=aA
equivalent to
is
A=2v /?
s
a pounds,
its
jr
the
sum
number
the
1 to x-=.
value
is
for
ob
denoted by Ev*p x may be divided into two

^v np n -if- 2?ry; z where n is to be taken from 1 to n, and
series
z from n -j-
to the number for which
p vanishes. The first gives
the value of the temporary annuity on the given
life for
years, and the second the value of the deferred annuity, that
is to say, of the annuity to commence n years hence if the
individual shall
be then
remainder of his
on the
life,
for
A
n
n)
Let
to continue
during the
be the value of the annuity

years for the whole of
now aged y
the value of a temporary annuity on the same
years,
years on the
To
life.
of a person
life
and
living,
find
and
A^ the value
same
A* dn>
life,
let
we have then Ar=A
A n be
n)
-f
A (dM
>.
the value of an annuity on a
If the person
now aged y
years.
life
years lives over
aged
n years, the value of an annuity on the remainder of
y-\-
life
of an annuity deferred n
his
PROBABLE DURATION OP HUMAN

life will
then be
The
n.
pn
it is
93
present value of this sum,
if it
# n A n, and the probability

n
therefore its value is v p n k n
Hence
were to be received certainly,

of receiving
LIFE.
is
A( dn ^=v np n A n and
,
A>=A
vnp n A n
-,
temporary and deferred annuities are

from
tables of A and p for all the diffe
readily computed
so that the values of
rent ages.
The
56.
A=A<* W
equation
which the values of
(rt)
,
n=l
Let
ther.
>
+ A< dn
>
gives a formula
we have then A=rA^ 1) -|-?^ i A
the value of an annuity for one year,
value of the
life
given
first
by
are readily deduced from one ano
payment
be received
to
surviving one year. Its value
we have consequently A =vp
-}-vp
is
is
in the
But
merely the
event of the
therefore vp\
A 1? orA=?y
and
(l-}-A 1 ).
This formula, which gives the value of an annuity at any

age in terms of the next higher age, and greatly facilitates
the computation of the annuity tables,
The value of an annuity on
57.
ber of individuals, that
is,
life.
A, B, C, D, &c. and
to Euler.
the joint lives of any
num
to continue only while they are all
living, is calculated precisely in the
nuity on a single
due
is
same manner
as the
an
Let there be any number of individuals,

let the probabilities
one year be respectively
p l9 q^ r^
s 1}
of each living over
&c. and
let
the probability that they will all live over one year
P = piX
P2 = p2 X
l
X
q2 X
q-i
x
rQ X
rl
PI be
then
s 19 &c.
ss,
c.
PS
p x X qx X r x x f# &c.
and the value of an annuity of L.I on the joint lives is
2t-*P a from #=1 to x= the number which renders any one
.,
of the probabilities^,
58.
The
q, r, 5,
&c. nothing.
value of an annuity on the survivor of any
num-
94
her of given
them
lives, that
is,
to continue so long as
The
exists, is thus found.
alive at the
probability that
any one of
A will be
end of the #th year being /?,, the probabi

be alive at the end of that time is
will not
lity that he
px
The
probability that all the lives will be extinct
end of the #th year
at the
is
therefore
and the probability that they will not all be

at least one of them will be in being, is
extinct, or that
i-0-/>.)0-?.)(i-^)0-*.) &*
which becomes by multiplication
p xq*p*r*
+P* q* rx +p x
qx
P* q* r x sx
&c<
sx ......
+qx rx
-.s
+&C.
4-&c.
Multiplying each of the terms by v*, and taking the sums of
the respective products from
x
px q
^.v
x is
*Zv x
and B,
C, and so on,
The
of lives
lives,
observing that
pxqxrx
that on the joint lives of A, B, and
we have
this rule
value of an annuity on the survivor of any

is
equal to the
sum of the
threes,
sum of the
value of the annuity on the
59* Let
V denote
pair of joint
annuities on the joint lives taken
When
and so on.
number
annuities on each of the
minus the sum of the annuities on each
lives, plus the
by
#=1, and
the value of the annuity on the joint lives of
there are only two lives, the
life
of the survivor becomes
the value of an assurance on the
life
of A, or the present worth of L.I to be received at the end

of the year in which
shall die.
In respect of any year,
the #th, after the present, the probability of
A dying
in the

course of that year is/V-i
lity that
year, then
life
is
ing over one year
living over
ing over
the probability of a
therefore
years,
px
Pxi(l
over one
will live
life
of that age not
liv
being the probability of A

u) is the chance of his liv
px
years,
Now v*
years and not survive x years.
be received certainly
to
u be the probabi
let
95
and dying in the following year (7).

u)=px_i px-i u ; and by (53), px î u=pf ;
i
px is the chance that A will survive x 1
But px-i (1
therefore
For
p*
years older than
LIFE.
at the
is
the value of L.I
end of the #th year
there
fore in respect of the xth year the value of the expectation

is
PX)> whence we have
V X (PJI
for the value of the as
surance
fromvr=l
we
to
#=r the number which makes j=0.
observe that
==1, and
/?
obvious that Zv x px _\-=.v(\
2v
p x by A, (A being
on the given
life),
paid at the
fail),
x
-\-*Lv
it
will
if
be
whence, denoting
as in (54) the value of the annuity
we have
V=r(l+A)
The
60;
Now$
Z.v*p. -.i=.vZv*- p.e__i,

c
A;
or
V=*>
v)A.
(1
values of assurances on joint lives, (that
end of the year
in
which any one of the
or on the survivor of any
is,
to
be
lives shall
number of joint lives,
are cal
culated from the corresponding annuities by means of the same

formula.
Thus,
let
be the value of an annuity of L.I
the value of an assur

on any number of joint lives, and
V =w (1 v)A
then
L.I
on
the
same
of
ance
joint lives,
If
A"
be the annuity, and
the survivor of any
V"=v
(I
V"
the assurance on the
number of given
lives,
life
we have
of
still
)A".
61. Assurances on lives are usually paid not in single pay
ments, but by equal yearly payments, the
first
being made
96
at the time the contract
and the succeeding
entered" into,
is
ones at the end of each future year during the
The
assured.
sum which
present value of the
contracts to pay
is
therefore equal to the
first
of the
life
the assured
payment add
ed to the value of an annuity of the same amount on his
and
if
the assurance
this
lity,
life
made on terms of mathematical equa
is
sum must be
precisely equal to the value of the
assurance in a single payment.
Therefore,
if
y denote
the
amount of the yearly payment, we have the equation

y(l+A) = V; whence ?/=V-T-(l-f. A).
The
62.
value of a temporary assurance for n years, that
of an assurance to be paid only in the event of the indi
is,
vidual dying before the end of
V be
n years
is
Let
thus found.
the present value of L.I, to be paid on the death of a
now aged y years, and V n the present value of L.I,

be paid on the death of a person now aged y-\-n years.
At the end of n years from the present time, the value of
person
to
assured on the
L.I
be
will
L.l to
Vn
he be then
life
Vn
continuing n years,
tract this
from V, we
is
continue n years
The
that as the
first
to
payment
be made
in
is
amount
for
premium by
years
1 years.
there
vnp n
Vn
If,
therefore,
we sub
v np n
Vn
found by observing,
made immediately, and n pay
all,
after the first is that of a
have the value of the temporary
shall
equivalent annual premium
ments are
is
subject to the contingency of
assurance in a single payment, namely
years
living.
life will
fore the present value of
now aged y
the probability that the
the
of a person
life
But the present value of

be received certainly at the end of n years is?;" and
,
if
is
the value of
all
the premiums
temporary annuity of the same

Denoting therefore the annual
u,
and the value of a temporary annuity
by
<*"
>,
the value of
all
the premiums
for
is

n
M-fwA<
>
=rw(l -f)A<
=yv
The
63.
n
n
m
n1
whence
while
is
probability of
of
sum
sum of money to be
which
in
receiv
provided he die
dies,
living.
W=r
Let the sum be L.I,
lity
of frequent occurrence.
is
following question
Required the present value of a
ed at the end of the year
97
and we have consequently
>
LIFE.
living over
living over
years.
its
p x -=.
present value,
the
and q x
the probabi
The chance of receiving the
years,
end of any given year, the #th, depends on two

contingencies ; 1. A may die in the course of that year, and
at the
live
over
it
2.
and
B may
both die in that year,
dying first. The probability of

dying in the a?th year
has been shewn (59) to be px_i px ; whence (7) the
probability of the first contingency is (p x
p x )q x The
B
that
and
will
in
both
the
.rth year is
die
probability
\
p x )(q x \ q x ) and for so short a period as one year,

be
considered an even chance whether A or B will
may
whatever
be the difference of their ages therefore
die first,
(p *_!
it
the probability in respect of the second contingency
##)
%(p x \ Px)(q x
the sum being received
Hence
at the
end of the #th year,
p
+ ^(px-ipx)(q*-iq )=l;(p*-ip
fe+i + $ *)> which being developed, and multiplied by
x
(Px-l
is
the whole probability of
)q x
is
x)
vx
,
becomes
and the sum of
#=1,
It
all
the values of this expression from
gives the value of
W.
has been already shewn (59) that
(1
v)A, where
1 v x (p xî
l
A= the annuity on the
life
/?*)=
of A. In like
98
we denote by AB the value of an annuity on the

we shall have *Zv x (p x-.iq x-.ip*q*)
if
manner,
joint lives of Ajind B,
=-v
#)AB, which
(1
jt/^rrjt/j />.,,_!,
individual
then
first
Assume p such
dying.
will live
xx
Pi
Again,
B who
is
Pi
~q\q x
-.\>
then q x
denoting by
*Lifpx q
will live over
AB
and B.
years,
__
denoting by AB the
value of an annuity on the joint lives of

ing the different terms,
we have
A and B
Collect
therefore
v)AB + 4"A B --7
W=ifv(l
is
one year younger than B,
and I v xp x qx__i
is
over x years
the value of an annuity on the joint lives of

let
that
evidently the probability that an
x is
one year younger than A,

x
AB
the value of an assurance to be
is
paid on the death of the
AB
whence
computed from tables of annuities on joint lives.

and B are both of the same age, the two last terms
easily
If
destroy each other, and
is
equal to ^ the value of L.I, to
be paid on the failure of the joint

to be, since there
ing before
evidently ought
same chance of A dy
dying before A.
in this case the
is
as of
The formula
lives, as it
gives the value of L.I in a single
the equivalent yearly payment
is
payment
W divided by 1-fAB,
for
the contract ceases on the failure of the joint lives by the
death of either.
It
would be easy
to
extend the formula to the case of an
assurance to be paid on the contingency of the failure of
any number of
of other
lives
during the continuance of any number
lives, or of an assurance to continue only during a
stated time
but as
it is
not our purpose to give solutions
LIFE.
of the various problems of this kind which
may
99
occur in
practice, but merely to shew the manner in which the ge

neral principles of the theory are applied to them, we shall
not pursue the subject farther, but refer the reader to the
article
Milne,
ANNUITIES, and
2
in
which
it is
to the standard
works of Baily 1 and
treated in detail.
The Doctrine of Life Annuities and Assurances analytically inves

and practically explained, fyc. By Francis Baily. London,
This werk is now out of print, but a French translation of
1813.
1
tigated
it
has recently been published at Paris.

2
Treatise on the Valuation of Annuities and Assurances on Lives
and Survivorships,
Sfc.
By
Joshua Milne.
London, 1815.
APPLICATION TO THE
100
SECTION
VII.
OF THE APPLICATION OF THE THEORY OF PROBABILITY

TO TESTIMONY, AND TO THE DECISIONS OF JURIES AND
TRIBUNALS.
64.
The
case of a witness
making an
assertion
may be
represented by an urn containing balls of two colours, the

ratio of the number of one colour to that of the other
being
unknown, but presumed from the result of a number of ex
periments, which consist in drawing a ball at random, and
replacing
it
in the
urn after each
trial.
true assertion
being represented by a ball of one colour, and a false one

by a ball of the other, it follows from the theorem in (51),
that if a witness has
true and
true
is
false,
-,
m+n+2
made m-\-n
assertions, of
which
m are
the probability of a future assertion being
and that of its being
false
n ~*~
m+n+2
Let
of these fractions be represented by v, and the se

cond by w, then v is the measure of the veracity of the in
the
first
dividual, or the probability of his speaking the truth,
the opposite probability, since
v+wl.
and
In general, the
existing data are insufficient to enable us to determine the
numerical values of v and
in this
manner
in applying the formulae to particular cases,
and therefore
we must
assign
on previous
knowledge of the moral character of the individual, or on
arbitrary values to these quantities, founded
DECISIONS OF JURIES ANt>
some
more
notions,
number of
relative
or less sanctioned
by experience, of the
made by men
true and false statements
in general, placed in similar circumstances.
Having assumed v and
iv, let us suppose a witness to

an event has taken place, the a priori probabi
of which is jo, and let it be proposed to determine the
65.
testify that
lity
In
probability of the event after the testimony.
the event observed (E)
is
this case
the assertion of the witness, and
two hypotheses only can be made respecting its cause 1st,

that the event testified really took place ; and 2d, that it
;
On
did not.
the
first
hypothesis the witness has spoken
the truth, the probability of which
occurred of which the probability
is
is
probability (Pj) of the coincidence

hypothesis, the witness has testified
of which
is
is
and an event has
therefore (7) the
On
vp.
the second
falsely, the probability
and the event attested did not happen, the
probability of which is q\ therefore the probability (P 2 ) of

the coincidence is wq.
Hence, by the formula (47) z^rr:
P,-r-sP )the probability (ar 1 of the first hypothesis becomes
,
and the probability
The sum of
of the second
(zr 2 )
these two probabilities
is
unit,
vp-\-wq
a condition
which ought evidently to be fulfilled, since no other hypo

thesis can be made, and consequently one or other of the
two must be
of
w-j
and
true.
sr 2
It is to
be observed, that these values
are the respective probabilities, after the tes
timony has been given, that the event attested took place,
and that it did not.
vp
-
Since -0-,=
vp
+ wq
--
L-2
= p\v(\p)
L-L
vp
wq
wq}
L>
we have
*~-i--*/
= p(vqwq)
vp
wq
-or.
.
;
--
= p(v
but v
vp
-
wq)
vp -f- wq
w=v
+ v=i
A.MM
_Y
therefore
1,
vTION TO THE
1.
*-./>= ^-
?.
This traction being
positive or negative, according as 2r
than unity, or as r
it
cî,
then
is
bility
On
w,^/);
that
J,
greater than
is
greater or less
follows that
it
to say, the probability of the
is
the veracity of the witness
the contrary,
is
groat or or loss than i,
event alter the testimony
when
its
is
a priori proba
greater than ^.
the veracity of the witness
it
the effect of the testimony
the event less than
its
is
less
than
to render the probability of
is
a priori probability.
66. If the event asserted by the witness be of such a na

ture that
so that
occurrence
its
is
is
a priori extremely improbable,
a very small fraction, and q consequently ap
proaches nearly to unity, although at the same time the ve

racity of the witness
approaching
be great, and measured by a fraction

w, becomes nearly equal
to unity, the value of
\op-r-ic, (for on this supposition /j-j-wv/H-r
But
to w).
bility
it is
obvious, that
is
nearly equal
however great the improba
of a witness giving false testimony
the improbability of a physical event
times greater
be given to
ir,
times smaller
in other words,
the value of
may be supposed,
may be any number of
however small a value may

may still be any number of
so that notwithstanding the veracity of the
witness, the probability of the event after the testimony,
namely
tity.
arjSr/J-f-ir
On
may be
this principle
less
than any assignable quan

easily give credence
mankind do not
to a witness asserting a very extraordinary or improbable
event.
be so
The odds
against the occurrence of the event
may
no single witness, how

character, would suffice to induce be
great, that the testimony of
ever respectable his

lief.
(37.
In the case of the character of a witness being alto-
DECISIONS OF JURIES AJCD TRIBUNAL*.

gether unknown,
we may suppose v to have aQ possible values
within certain limits, and to find the value of
w l by integrat
ing the fractionyW , </; between those limits.
we
,.
tp-f-
ing
/,
^/
J vp-^wq
v and
p + (2p
/>
for
2/*
and q
which on
substitut-
become*
respectively,
the integral of which
p
*
,=:
is
l)v
]
{fj
liave fa.dc-=.
Since
2y>
1
J
being a constant, the value of which will be determined

from the assumed limits. If v be fllj l iftfil to vary between
the limits
and
if
r=0 and
rz=l, then
we assume /?=J, we have /V l <r=r|(l
which, since the logarithm
is
Jlog.3),
the Napierian logarithm,
and
Nap. log. 3=1.0986, becomes f X .4507=.676, or nearly f.

Whence we see, that on this hypothesis the probability of
the event
68.
diminished in consequence of the testimony.

credit due to the ^ftiftMFffy of a witness depends
is
The
not merely on his good faith, but also on the probability that
he is not himself deceived with respect to the event he as
serts.
The chances of a
credulity or ignorance are
witness being deceived through
much more numerous
than the chances of intentional fraud

the case more particularly
ture that
it
taken one for another
when
and
the event
in various
may happen
is
in general
this
must be
of such a na
ways which may be mis
as for instance, in the case of a lot
and the witness asserting that it

number, which might with equal proba
tery ticket being drawn,
bears a particular
bility
be any other number on the wheel.
The
following
APPLICATION TO THE
104
method of applying the
question will illustrate the
when a
An
made between
is
urn contains s
marked
at
distinction
A2
calculus
these sources of error.
of which a\ are marked
balls,
An A
a n marked
A 1?
random, a witness of the drawing affirms that the
drawn
Am
marked
is
+an
(n being the
of the different indices or sorts of balls)
we make p l =a --s,p 2 =za 2 -i-s
.pm
=an -^s,
the a priori probability that the ball drawn
marked
A pe
1?
whose index
A2
is
rent hypotheses
ball
and so on.
may be made
which has been drawn,
it
ar
is
is
belongs to the class

evident that n
diffe
to any one of
Let the probabilities of
l?
drawn
then pi
of the class
may belong
these hypotheses be respectively
assertion that the ball
is
so that if
respecting the index of the
for
spect of any particular index
it
It is
A A 2 ...A n
the different classes
ball
required the probability of the testi
mony being true.

Here we have s=:a l +a a -{-a 3
number
having been drawn
ball
is
w 2 ... WB
(that
is,
in re
the probability after the
marked A,)
and
let
the
probabilities of the assertion
P 2 ...P n
respectively P,,
ed A,, then P,
be marked
is
on each of these hypotheses be

(that is, if the ball drawn be mark
the probability the witness will assert
A m ).
it
v be the veracity of the wit

u
and
the
that
he has not been deceived.
ness,
probability
to
(1.)
is
Lastly, let
Let us first consider the hypothesis that the ball drawn

A m and consequently that the assertion is true.
marked
In order to find
Pm
the probability of the assertion being
made, there are four cases

suppose the witness
is
speaks the truth (v).

this case is uv.
2d,
tends to deceive, or
to
be considered.
1st,
we may
not deceived himself (w), and that he
The probability of the assertion in

The witness knows the truth, but in
testifies falsely.
In this case the proba-
AND TRIBUNALS.
DECISIONS OF JURIES
105
of the assertion being made, on the hypothesis under

is 0. 3d, The witness has been deceived him
bility
consideration,
but intends to speak the truth. In this case also the proba
bility of the assertion being made is 0. 4th, The witness has
self,
been deceived himself, and intends

the assertion might be
made
and
In this case
to deceive.
to find the probability of
being made we have to consider, that since the witness

has been deceived, he must have supposed some other index
its
A^
than
ceive,
have been drawn
to
and since he intends
to
de
he must assert some other index to be drawn than
that which he supposes to be drawn.
Setting aside, there
fore, the index which he supposes to have been drawn, there
remain n
when he
any one of which he is as
others,
as any other.
The
intends to deceive
is
l~(n
is
the witness
is
deceived
tends to deceive =r(l
(n
this case
).
=(l
v)
u)
w)(l
v)-~-(n
the index of the ball
case, we have P m
being made on the
first
bability of the assertion
drawn was
Am
Probabi
he
he names
assertion
is
in
Am
therefore in
this to the
Adding
1).
2. Probability
3. Probability
The probability of the
bability found in the
(2.)
=(1
the pro
compounded of the
probabilities of three simple events, as follows

lity
naming A m
Hence
1).
bability of the assertion in this case
name
likely to
probability, therefore, of his
pro
the whole pro
hypothesis that
namely
Let us now consider one of the remaining hypotheses,

ball actually drawn was marked A,, and
and suppose that the

not
Am
as attested
by the
witness.
As
four possible cases for consideration.
knows the
fact and speaks the truth.
sertion could not
be made, or
its
before, there are

1st,
The
witness
In this case the as
probability
is 0.
2d,
The
APPLICATION TO THE
106
witness knows the fact, and intends to deceive.
In this case
the probability of his asserting A m to be drawn is compound

ed of the probability that he is not deceived (u), the proba
bility that
he
testifies falsely
(1
v),
and the probability that,

A m from among
knowing the index A f to be drawn, he selects
then
which remain after rejecting A (l-=-(w
made
probability of the assertion being

fore
v]-^-(n
(1
1).
The
3d,
intends to speak the truth.

case,
it is
is
is
is
there
deceived, and
reasoning as in the last
easy to see that the probability of the assertion
being made in
ness
By
witness
The
1)).
in this case
this case is v (1
u)-s-(n
of the assertion being made in

sidering, that as the witness
4th,
1).
The
deceived, and intends to deceive.

this case will
The
wit
probability
be found by con
himself deceived, he must
is
suppose some particular index to be drawn different from
f,
(which drawn by hypothesis), for instance A the proba

bility of which is ]-r-(n
1); and intending to deceive, he
is
must
on some index different from
fix
poses to
be drawn
of which selection
also l-?-(ra
is
Am
is
\-s-(n
I)
The
1).
therefore, that the witness supposes
annnounces
which he sup
e,
and he announces A m the probability
But
it
probability,
be drawn, and
to
is
evident, that
whatever can be affirmed with respect to the particular in

dex A e may be affirmed with equal truth of every one of
,
the other
indexes, excepting A,, which
(since by hypothesis the witness
he announces,
2
therefore n
(since
is
is
actually drawn,
deceived), and
Am
which
by hypothesis he lies). There are

ways in which he may at the same
different
time be deceived, and intend to deceive, and announce
Am
consequently the probability of this announcement in any of

2
these ways is (n
Multiplying this into the
I)
2)-f-(?z
.
probability of his being deceived (1
u),
and the probability
DECISIONS OF JURIES AND TRIBUNALS.
107
-___- -
of his giving false testimony (1
v),
assertion in this case becomes -
the whole probability of the assertion, in
cluded
the cases in
drawn was
is
A,,
u(l-v)
n1
As
all
in the hypothesis that the ball actually
marked
Hence
v\f
nl
(nl)
this expression will evidently
any other of the n
assertion on
be the probability of the

hypotheses that the ball
different from A m
actually drawn was marked with an index
the
sum of
potheses
bers
the probabilities of the assertion on
2P
is
Z,
1, 2, 3,-..n,
where
all
these hy
successively each of the
i is
num
excepting m.
We have now to find wm
first
hypo
Since the hypotheses, in the present question, are not

equally probable a priori^ we must have recourse to the
thesis.
all
formula (49)ar
case
=X P -r-2X,. lp
z
l.,
and consequently in the present
we have
Pm
>
AmJ
~X m P m + SX P;
f
the sign of summation
including every value of
from
Now the value of P, being the

excepting i=m.
same in respect of each of the hypotheses which suppose
the assertion untrue, SXj^rrPiZX, ; and the sum of all the
to n,
i=Q
values of X f from
\ m ~a m -t-s,
to
we have SX =(s
t
together with the values of

r
making u
i=n
u,
v,
Pm
being
on excluding
ct^-r-s.
Substituting this,
and
above found, and
P,, as
the formula becomes, after the
proper reduction,
a m {(n
1,
l)ut>+uV}
APPLICATION TO THE
108
which
ed
is
the probability of the hypothesis that a ball
A m was drawn, or that the testimony is true.

When there are no two balls in the urn having
index, the
numbers a lt
mark
the same
a s &c. become each =1, and
a^
In this case the formula gives
s~?i.
(n
)uv -f. u v
uv
\) (u v
which, on observing that
uv+u v + uv
+ (n
-\-u
2)u v
v =.\, becomes
by reduction
This
is
the probability of the truth of the testimony of a
witness,
who
affirms that the
urn which contains n

vious, that
and n
is
balls,
when u and v are
is
I, 2,
drawn from an
3... n.
It is
ob
fractions approaching to unity,
a considerable number, the second term becomes
very small, and
may be
becomes simply
wm =M#.
69.
number
numbered
neglected.
The
probability then
We now proceed to consider the probability of an event
attested
by several witnesses and first let us suppose the wit

The measures of the ve
;
nesses to agree in their testimony.
racity of the several witnesses being respectively v lt v a , v s ,

&c.,
and the a priori probability of the event being
have by (58)
first
for its probability after the
p,
we
testimony of the
witness,
In order to find the probability of the event after the se
cond witness gives
mony
of the
first
his testimony,
we may suppose
the a
testi
be changed from p to
by the
formula
and
the
same
witness,
gives
priori probability to
109
Let a third witness now come forward, and give testi

in favour of the same event. Its probability after his
mony
testimony will become in like manner
In general, let
-sr
be the probability of an event
after
it
has been attested by x witnesses, and let vx be the veracity

of the last witness, then Trx \ being the probability of the
event after x
If
l
eyewitnesses have each testified in
its fa
we have
vour,
we suppose
the witnesses
=v 2 =v 3 ...=:v*,
this
all
equally credible, or that
becomes
V*
v
it
Now, if v=-|, then (1 v)-?-v=:l, and w,=/?; whence

appears that the probability of an event is not increased
by the testimony of any number of witnesses, when the vera

city of each
is
only ^; but
becomes more probable

and when v
is
when v is greater than
as the
a considerable fraction,
creases very rapidly with the
70.
When
given value.
required to
number of
the values of v and
the last formula
|,
the event
number ofwitnesses is greater,
may be found
its
probability in
witnesses.
are given, that of x in
so as to render
-srx
of any
Hence we may find the number of witnesses

make it an even wager, whether an event ex
ceedingly improbable, and in favour of which they give una-
110
APPLICATION TO THE
nimous testimony, has happened or not. For

example,
the odds against the event be a million million to one,
let
tllat is
let
P=
i,ooo,oo.Uo,ooi
veracity ofeach witness
we must have
Zf = 10 12
v\*
/I
(
be.
therefore
P
1
log
-p
x log
or
9= 12
In order that &*
= Tot? and let
more probable
it
than that
it
v
v
12
=l
(-Vxl0
\y/
J,
= -,and
9
whence x log ^=r

y
and therefore x
nearly, so that 13 independent witnesses
render
may equal
Now
1.
" the
12
would
=12-6
suffice to
that the event really took place
did not.
This example is given by Mr. Babbage, (Ninth Bridgewater Treatise, Note E), with a view to shew the fallacy
of
Hume s
celebrated argument respecting miracles.
What
that if
we suppose an urn
to contain a million million of white balls,
and only one black
the example proves
ball,
is
simply
this,
and that on a ball being drawn at random from the
urn, thirteen eyewitnesses of the drawing, each of
whom
makes only one false statement in ten, without collusion,

and independently of each other, affirm to A, who was not
present at the drawing, that the ball drawn was black, then
would have rather a stronger reason for believing than
for disbelieving the testimony.
But
it is
sufficiently obvious,
that the event attested in this case, though exceedingly
probable a priori, cannot be regarded

culous.
On the contrary,
with the same
facility,
the black ball
and was a priori
im
any way mira

might be drawn
as in
as likely to
be drawn,
DECISIONS OF JURIES
AND TRIBUNALS.
1 1 1
Let it be granted
of
within the range
fortuitous occurrence,
as any other specified ball in the urn.

that an event
is
and that there exists a single chance
any number of
millions of chances,
in its favour out of
may then happen in

of
trials
a
number
trial
one
may be assigned,
nay,
any
such that its non-occurrence would be many times more im
it
probable than the contrary.
number of
71. Let us next consider the case of a
nesses contradicting each other.
If the
first
nounces an event of which the probability

probability, after the testimony, of its
ar,,
and the probability that
it
wit
witness an
then the
is />,
having happened
has not happened
a-,.
is
Sup
pose a second witness now to appear, and testify that the

event has not happened, and let the probability of the truth
of his testimony be denoted by
sr 2
then
sr,
being the
probability before his testimony was given that what he as

serts
is
true,
and v 2 being the measure of
have, as in (69),
^ =V
2
2(
WT1
for the
>
\7~/ii
--
we
whence, since
tnere results
P)
measure of the probability that the event has not
The
happened.
fore
his veracity,
z?-
2,
probability that
and accordingly
if
it
has happened
be
less
than
is
-J,
there
there
a stronger reason for believing that the event happened

than that it did not. The method of forming the expression
is
for the probability of the event, after
or denied
nesses,
is
by a third witness, or any

obvious.
it
has been attested
number of successive wit
APPLICATION TO THE
112
we suppose
If
the values of v l and v 2 to be equal, the ex
pression
becomes
lity that
the event did not happen.
must be the
p, which
is
the
It is
a priori probabi
obvious that this
inasmuch as two contradictory testimonies

of equal weight neutralize each other. In general, the pro
bability of an event which is affirmed by m witnesses, and
case,
denied by n witnesses,
all
that of an event which
is
equally credible,
affirmed
by
is
the same as
n witnesses who
agree in their testimony.
When
72.
ries
a relation has been transmitted through a se

whom the first only has a direct know
of narrators, of
ledge of the event, and each of the others derives his
know
ledge from the relation of the preceding, the probability of

the event is diminished by every succeeding relation
In
order to obtain a general expression for the probability of
we may
traditionary testimony,
take the event considered
in (68), namely the extraction of a ball marked A m from an

urn containing s balls, of which
are marked A t ,* 2 mark
ed A 2 ,... n marked A n there being in all n different in
t
Now
dices.
suppose the relation to have passed through a
T T 2 ,...T.r
chain of narrators, T,
whom
the
first
t,
in
number
.r-f-1,
of
only was an eyewitness of the event, each
of the others receiving his knowledge of it from the one pre
ceding him, and communicating

ceeding, the question
ball
Am
marked
rated
Tx
by
is
it
in his turn to the suc
to determine the probability that a
was drawn,
after this event has
been nar
the last witness of the series.
73. In order to apply the general formula of (68) to this

case,
it is
necessary to remark that the event observed
is
the
T, of his having been informed by T,_i; that

drawn from the urn, the drawing of which was seen
attestation of
the ball
by T, was marked
Am
There are n
different hypotheses
113
respecting the index of the ball actually drawn, but
it
is
only necessary to consider two of them, namely, the hypo
drawn was marked A m and any

one of the other hypotheses which consist in supposing that
a ball with a different index from A m was drawn, for ex
thesis that the ball actually
ample
Let the probability of the attestation, on the hy

drawn was A m be denot
pothesis that the index of the ball
ed by y x and its probability on the hypothesis that the in

dex was A, by y ,, (yx and y x corresponding to P m and
P in (68), which express the same probabilities in respect
,
of the eyewitness T). then by (68), the probability of the

hypothesis that
Am
But since y
the
x is
dex drawn was

(68) Z\-=(s
was drawn
same
for all the
Am
from
different
a m )--s, and A m
+ (s
We
v2
have now to find
is
yx and y
hypotheses that the
SA^rrf/^ZA,
m -j-*, therefore
,
in
and by
0/
x in
terms of x.
&c. be the respective probabilities of T,
Let
T I} T 2
v,
,
v lt
&c.
speaking the truth, then the probability of T, speaking the

truth is ?;,,and the probability that he does not 1
v,, whether
because he
is
dishonest,
and intends
to deceive, or because
he has mistaken the statement of the preceding witness. Now

there are two ways in which it may happen that A m is announ
ced by T,. First, if he speaks the truth, and has been inform
ed by the preceding narrator Tj_i that A m was the index
drawn ; secondly, if he lies, and has been informed by T ,_i
A m was drawn. Assuming yx ^\

same signification with respect to T^_i that yf
has been assumed to have with respect to T,, (that is to say,
that a different index from
to have the
APPLICATION TO THE
114
the probability of the assertion being
hypothesis that the ball actually

bility
of the
of these combinations
first
spect to the second case,
nounces a
ced
of n
different
vx
on the
to
With
yx_\*
be observed, that
1)
l-i-(72
and on multiplying
v, that the testimony of
yx _i
that
re
his
if
announcing
Am
indexes different from that announced by T^
we
is
Tx
A m ), the proba
T, an
index from that which has been announ
it is
him by T^_i, the chance of
to
made by
drawn was
this
Tx is false,
Tj_i has announced a
out
i
is
by the probability
and by the probability

from A TO
different index
have, for the probability of the second combination
(l_vj(l_^_ )^_(/i_l). The whole probability of T, tes

tifying that A m was drawn, is therefore, on the first hypo
1
thesis,
given by the equation,
yx =vxyx^ + (\
This
is
an equation of
O(l-^_0
finite differences
the complete integral of which
_1
C(nv,
y*~n*
\)(nv,
.(n
of the
1).
first
order,
is
l)...(^-t
!)(>,
l)i
(n\)*
In order to determine the arbitrary constant C, it is to be ob

served, that, since T/!, y^-y^ as well as v v v 2 ...vx apply to
the narrators
T T 2 ...T,
t,
respectively, if we suppose
x=o
the resulting value of the integral will be the probiibility

that A m was announced by the eyewitness T, on the hy
be
Pm
C=(nP m
n.
) -T-
If,
therefore,
we make.
1
is easily verified ; for on changing x into x
and forming the expression vx yî + (1 Uj)(l
This
tegral,
(rc
Am
was actually drawn. Let this probability

then the equation becomes P m =C-f-l-r-ra whence
pothesis that
1),
there results an identical equation.
in the in
yx
i)-r-
we
obtain,
on the
first
tually
drawn was
being
testified
P,,
we
15
hypothesis, for any value of #,
In the same manner

testimony given by
we
find the probability
of the
T*, on the hypothesis that the ball ac

.
t
The
probability of the event, after
by the eyewitness, being on this hypothesis
have,
Substituting these values of yx and
above found
for zrm ,
we
x in
the expression
obtain for the probability of the event
observed by T, and narrated by T,, the narration having

passed from one to another in the manner supposed,
74. Since
=v x----f-,
and since vx
is
always less
than unity, and n always greater than unity, each of the
terms of the series represented by X, whether positive or

negative, is a proper fraction, whence the value of X be
comes smaller and smaller

finite,
then
0,
and
bability of the event.
vr m
as
x increases.
Supposes
in
~am -t-s, which is the a priori pro
Hence we
see that the probability of
an event transmitted through a series of traditionary evi

dence becomes weaker at every step, and ultimately equal
to the simple probability of the event, independent of
any
testimony.
75. When the urn is supposed to contain only n balls,
each having a different index, the expression for wm is great
ly simplified
P m +(n
for, in this case,
1)P
consequently
=1) the
vrm
a m -=:l,sn
therefore, (since
denominator becomes n, and we have
~l -\-(nP m
l)X|-j-^, which coincides
APPLICATION TO THE
116
with the value of yx found above, that
is
thesis that
it
to say, with the
by T, on the hypo
probability of the event being testified
Laplace, in solving this
actually happened.
particular case of the problem, (p. 456) assumes that the
are,
however, as
wm are identical. They
by yx and
probabilities here denoted
evident from the above
is
analysis, quite
distinct in their nature, and their values are only equal in the
which
particular case in
to 1.
76.
am
is
to
am
in the ratio
of
(Poisson, p. 112.)
The
question of determining the probability that the
verdict of a jury
is
is
correct,
precisely analogous to that of
by one or more
finding the probability of an event attested
Let us
witnesses.
first
take the case of a single juror, and
assume w=r the probability that the juror gives a cor

rect verdict, (that is, correct in respect of the facts), and
/?= the
probability that the accused
guilty before being

put on his trial. Suppose the verdict guilty to be return
ed ; two hypotheses may be made respecting the cause of
the verdict,
he
is
first,
be condemned
that the accused
On
innocent.
On
lity
of which
will
be condemned
u.
if
probability of which
ties
guilty
secondly, that
the second hypothesis, the accused
the juror gives a
wrong verdict, the

But the a priori probabili
u.
is 1
of these causes (the guilt or innocence of the accused)
being respectively
w-j
is
hypothesis, the accused will
first
juror gives a right verdict, the probabi
if the
is
the
is
and
p we have by
9
being the probability of the
bability that the accused
given, and
*r 2
the accused
is
is
first
(49)
hypothesis, or the pro
guilty after the verdict has
been
the probability resulting from the verdict that

innocent.
117
77. Suppose the verdict not guilty to be given,
and
a\
On
hypotheses.
cused
is
and
let
be the probabilities after the verdict of the two

the
first
hypothesis, namely, that the ac
will
guilty, this verdict
be given
a wrong verdict, of which the probability
if
the juror gives
u ; and on
is 1
the second hypothesis, the verdict will be given if the juror

gives a right verdict, of which the probability is u ; and the
probabilities of these hypotheses before the verdict being
respectively
From
the above value of
1)
,
*00
according as u
by the verdict
ir,,
we
a fraction which
obtain
is
guilt of the
positive or negative
is
Hence it appears
rendered
more probable
only
greater or less than J.

is
guilty being pronounced,
when
that the juror gives a correct verdict
In like manner
&
p)
that the guilt of the accused
lity
we have
as before,
-"
>
and
shewn
it is
that
ts \
is
the probabi
greater than \.
(the presumption of the
accused after a verdict of acquittal),
is
greater
p when u is less than ^.

78. The a priori probability of the condemnation of the
than
accused before he
is
put on his
trial is
w)0
p) J
two ways in which
condemnation may
take place ; first, if the accused be guilty, and the juror give
a correct verdict, the probability of which concurrence is
for there are
up
and, secondly,
ror give a
(lp).
wrong
up-\-(\
this
if
the accused be innocent, and the ju
verdict, the probability of which is (1
Therefore, making c
dict of condemnation,
= the
we have c=up-\-(\
for a verdict of acquittal,
c=(l
u)
probability of a ver
u)p
u)(l
+ u(\
p)
and
p).
79. Let us next suppose that after the verdict ef the
first
APPLICATION TO THE
118
juror has been pronounced, the accused
is
put onhis
trial
be
u s be the probability the second

a
correct
verdict, and c 2 be the probability the ac
juror gives
cused will be pronounced guilty by him. After the verdict
fore a second juror,
guilty has
lity
and
let
been pronounced by the
of the guilt of the accused
u2
c 2 will be found
first
by substituting
c, whence c a z=u a w l
rors
evident that
it is
for w,
the above value of
The
juror, the probabi
zr^ and
is
and & l
+ (\
probability of a verdict of condemnation

is
cc 2
in
-sr^).
by both ju
therefore, (observing that w-jsrttp-r-c),
for
U Q )(\
we have
for this probability
The
probability of the guilt of the accused after a verdict
first juror
being w j,
of
a
verdict of acquittal being given by the
the probability
of acquittal has been pronounced by the
second juror
cQ
is 1
(observing that
u)p--(l
bability of a verdict of acquittal
Adding the
+u
u^-sr\
(l
= (1
2 (\
c),
wJ
therefore,
wehavefor the pro
by both jurors
probability of a verdict of condemnation
by both
by both, we have uu s ^(\ u)
the probability of both giving the same verdict.
that of acquittal
jurors, to
u2)
(1
for
This result
inasmuch
may be
is
independent of/?, and is evidently true a priori,

two ways in which the same verdict
as there are
given, namely,
when both
jurors are right,
and when
both are wrong.
The
probability of acquittal
verdict of guilty
u 2 (l
23-j
wj;
guilty
by
first,
multiplying by
their values,
the
by the
we have
first,
c,
by the second
is
juror, after a
c 2 rr (1
and substituting
U 2 )î
for c
-f-
and
for the probability of a verdict of
and not
guilty
by the second,

In like manner,
first
if
the accused has been acquitted by the
of his guilt becomes

juror, the presumption
probability of a verdict guilty by the second

.j_(l
w 2 )(l
yr\)
of not guilty by the

C)c
(1
The sum
fl
is
and the
c^rrw^
therefore the probability of a verdict
first,
= (l
and of guilty by the second

U)u 2 p
+ U(l
is
M 8 )(l-_p).
of these two expressions gives for the probability
^2) +
of a discordant verdict, u(\

80. If
119
u)u 2
(l
we now suppose wrrw 2 and make

,
w, the
probability that the two jurors will agree in their verdict,
whether they are both right or both wrong, isu^

the probability of a discordant verdict
sum of the two expressions
iu
(u-}-w)
and
The
uw+uw2uw.
u *-\-2mv-\-2v 2
<
is
and
therefore the probabilities of the different cases are respec

tively given
by the developement of the binomial (u
+ w)
2
.
pursuing this reasoning, it is easy to see that if there

be any number h whatever of jurors, or voters on any ques
By
which admits only of simple affirmation or negation, all

being supposed to possess the same integrity and know
tion
ledge, so that there
the same probability
is
u of a correct
decision in respect of each, the probablities of the different
cases are found by the development of the binomial
(u+w)
h
.
The
probability of a correct verdict being pronounced una

h
nimously is u ; of an erroneous one being pronounced una
and the probability that a correct verdict

of the jurors, and an erroneous one by
be given
1 .2.3
h
m n
is Uu w , where U
,
nimously
is
n>
up
by m
will
81.
guilty
The
by
.2.3...wxl
.2.3...T*
probability that the accused will be pronounced
m jurors,
that the value of
and acquitted by n> on the supposition

the same for each juror, is thus found.
is
There are two ways
in
which
this
event
may
take place
APPLICATION TO THE
120
the accused be guilty (the probability of which
1st, if
and
lity
is
p),
m jurors
decide correctly, and n wrongly (the probabi

m
of which is Uu iv n ) ; the probability of the condemna
tion taking place in this
way
is
therefore \Ju m w p.
n
the accused be innocent (the probability of which
n jurors decide
which
in this
way
is
q) and
m wrongly
(the probability of
the probability of the event taking place
rightly,
Uu"w m )
is
and
2d, If
is
Uun wm q.
therefore
Let
therefore denote
we have
the whole probability of the verdict, and
G= U(u mw p + u wmq).
n
Hence
the probability that the accused will be
condemned
unanimously by a jury consisting of ^jurors is v^p ufq ;

and the probability that he will be unanimously acquitted
uh q-\-whp.
82. Suppose the accused to have been pronounced guilty
by
m jurors,
and not guilty by n
jurors, the probability of
the verdict of the majority being correct
Two
formula in (49).
accused
is
guilty
is
found from the
may be made 1st, the

innocent.
The probability Pj
hypotheses
2d, he
is
of the observed event (the condemnation by m, and acquit

m n
tal by n jurors) on the first hypothesis is Uu w ; and the
a priori
bilities
fore if
two hypotheses (or the proba

and X 2 in (49), being p and q ; there
denoted by X x
denote the probability of the verdict being cor
^
73-
rect, that
is,
verdict has been pronounced,
being wrong, we
757
~~
shall
and
first
zr s
hypothesis after the
the probability of
its
have (49)
um w np
u m wnp -f un w m q
~~
u n wmq
If the verdict has been pronounced unanimously, then
h and wzrO, and the formulae become
If
pq
J,
and
w=:z,
we have then
wn +
But
this is the probability of
~~
+w*
a verdict being correct which
has been pronounced unanimously by
jurors
follows that the probability of a decision rendered
majority being correct,
is
121
whence
it
by a given
the same as that of a decision ren
dered unanimously by a jury equal in number to the differ

ence between the majority and minority, and is therefore
independent of the total number of jurors. This, however,

only true on the supposition that the value of u is known
a priori ; for if u be not absolutely known, the weight of the
is
verdict depends on the ratio of the majority to the whole
This
ber of jurors.
for it will readily
is
in
accordance with
common
num
notions,
be admitted that a verdict given unani

much more weight
mously by a jury of 10 will be entitled to
than one pronounced by a jury consisting of a large num

ber, as 100, in which 55 are of one opinion, and 45 of the
opposite.
In this case, the opinion of the minority throws
great doubt on the correctness of the verdict.
It is to
be
observed, however, that the probability of a verdict being

given by a small majority becomes less and less as the num
ber of jurors
83.
When
the majority
jority
bility
is
increased.
the
is
number who
dissent from the opinion of
unknown, and we merely know that the
ma
exceeds the minority by at least i jurors, the proba

of the verdict being correct is found as follows. Sup
pose the verdict to be guilty.
On
the hypothesis that
it is
correct, the probability of the accused being found guilty
by h
x and not guilty by # jurors,

}
is
by the formula
in (80),
APPLICATION TO THE
122
Uuh
*iv
if we
Now,
0, 1, 2,... w,
give
where n=.^(h
z),
successively
and assume
the values
all
to denote the
when a?=0, U A its value when #=1, and so on

and also make W=the probability of the accused being
pronounced guilty by h n at least, we shall have
value of
In like manner,
guilty
if
denote the probability of a verdict
n jurors at
by h
not guilty,
accused
is
whence,
p and
least,
we
on the hypothesis that the
shall have,
q being as above the a priori probabilities
of the two hypotheses, the probability that the verdict guilty

is
correct,
84. It
when pronounced by h
is
n jurors at least, becomes
evident that no application can be
made of these
formulae without assigning arbitrary values to

less,
indeed, we have
With respect
from experience.
u and
data for determining their

to p,
mean
we may
/>,
un
values
assume, for
the sake of shewing the general consequences of the formu

lae, its
J,
value to be
for
it
cannot well be supposed
or that a person brought before a jury
is
more
less
than
likely to
be
innocent than guilty ; and if it much exceeds J and ap

proaches to unity, a verdict of guilty may be expected from
any jury, however constituted. When a mean value of u

cannot be determined from experience, the only way of ob
taining numerical results,
is
to suppose
to
have
all
possible
values within given limits, and to integrate the equations
between those
limits.
that a juror
more
right one,
is
As it seems unreasonable
we may assume
Suppose, then, that
to
suppose
likely to give a
that
u increases by
wrong verdict than a

u cannot be less than ^.
infinitely small
increments
DECISIONS OF JURIES
from
w=J
to tt=l,
and
let it
probability that a decision
infinite
the accused has
m jurors,
we must therefore have
Let
to the formulae in (51.)
Px =
to determine the
when
and not guilty by n.

number of hypotheses may be made re
specting the value of u, and
theses,
123
be proposed
correct
is
been pronounced guilty by
Here an
AND TRIBUNALS.
u=x
recourse
be one of those hypo
the probability on that hypothesis of the event
observed (that
is, of the accused being pronounced guilty

not
and
by m,
guilty by n jurors,) 33-*:= the probability of
the assumed hypothesis, and n
the mean probability of
the correctness of the verdict from
the formulae in (81)
and as
all
all
the hypotheses.
the hypotheses are supposed equally probable,
we have (45) ^rrP^-r-SP,. But between

limits
if,
By
we have
tP x =pft x m (l
therefore,
>x)
n dx
+ (1
we make j5=^j we
the proposed
jo)/jV
shall
(1
m
x) dx
have by reason of
and therefore
But (82) the proba

of
the accused being guilty, is
bility on this hypothesis
for the probability of the hypothesis.
ity
of the hypothesis,
multiplying this by the probabil-
wx we
,
obtain for the probability of
x m (\
n
n
m
x) --J^x (l
x) dx ;
and, therefore, for the probability of the verdict being cor
rect on all the hypotheses from x=^ to #=],
the
verdict
being correct
124
APPLICATION TO THE
Hence
m out
of
the probability that a verdict given
m-)-n=h jurors
is
Jc
wrong,
by a majority
is
(lx} n dx
which, on effecting the integrations by the formula in (51)
becomes
after reduction
~-~T-2-3
"
1-2
(h+l)h(hl) ...... (h
1-2-3
+ 2)
................
I
j
Assuming h (the number of jurors) =12, and making n

successively 0,
1, 2, 3, 4, 5,
the series gives
14
92
378
8192
8l92
8192
8192
for the respective probabilities
1093
2380
of the error of a verdict
when
pronounced unanimously by 12 jurors, by a majority of 11

In the
to 1, of 10 to 2, of 9 to 3, of 8 to 4, and of 7 to 5.
last
case the probability of the error
85.
From
these results
it
is
nearly
=f
appears that the chance of a
wrong which has been pronounced unanimous

twelve
ly by
jurors is very small ; but it is to be remarked,
that they have been deduced on the supposition that the
verdict being
unanimity proceeds from agreement in the same opinion, and

that the jurors are unbiassed by each other. In this country,
is compelled by law, the mean probability
of a correct verdict can scarcely be considered as greater than
where unanimity
that of a verdict pronounced
though in
by a simple majority;
most cases the verdict may be supposed
for,
to repre
sent the opinion of a larger majority than seven, it may

number than five,
happen, not unfrequently, that a smaller
or perseverance,
possessing greater energy
may persuade the
DECISIONS OF JURIES
AND TRIBUNALS.
others into a surrender of their judgment.
In
125
fact, unless
the presumption of the guilt of the accused be very great,

it
would scarcely be
unanimous verdict
possible, without concert, to procure

in
any
case.
It is also to
that the assumption of all values of u from Jto
an
be observed,
1
being equal
may lead to results widely different from the

The mean value of u, which depends on the general
ly probable,
truth.
intelligence of the class of persons from

lists
of jurors are
made
amongst
whom
the
up, can only be rightly determined
from data furnished by experience. One of the elements,

however, which require to be known for this purpose, is the
number of jurors who concur in, and dissent
The forced unanimity ofthe law renders

this
it
from, the verdict.
impossible to obtain
element from the records of the English courts
but in
France and Belgium, where the majority and minority are

known and recorded, the same obstacle does not exist, and
the " Comptes Generaux de FAdministration de la Justice
Criminelle" published by the French Government, have

enabled Poisson to deduce mean values of u and p for that
country, and consequently to obtain the necessary data for
one of the most interesting applications of the theory of

Probabilities. The general results were as follows
During
:
the six years from 1825 to 1830 inclusive, the system of cri
minal legislation in France underwent no change ; the jury
consisted of 12,
concur, though
and a simple majority was only required to

it
happened that the majority was the
when
least possible, the
On
Court had power to overrule the verdict.
comparing, according to the rules of the theory, the ver
dicts given
in the cases tried
during those six years,
it
before the criminal courts
was found that
for the
whole of
France, the probability (u) of a juror giving a correct

with respect to crimes
verdict was a little greater than
APPLICATION TO THE
126
and nearly equal
against the person,
crimes against property

of crime,
it
was found
to
to
\% with respect
to
without distinction of the species
be a very little below
The
J.
other
element, the probability (/>) of the guilt of the accused be

fore the trial, was found not much to exceed ^ (being be
tween 0.53 and 0.54) with respect to crimes against the

in respect of crimes
person, while it a little exceeded
against property.
Without
distinction of crime, its value
was very nearly 0.64.
On
86.
substituting these values of
u and
(namely
w=j, jo=.64, whence w=^, q=.36) in the formula

and making m=7, /z=5, and consequently
Hence
X.36)=:.07 nearly.
hundred trials it
pected, that in a
will
in (81),
may be ex
it
happen only seven
times that the accused will be pronounced guilty by the

If
smallest possible majority.
p+w q=:. 02027= ^

h
have uh
w=12, and
ra=0,
we
shall
nearly, for the probability of
b
an unanimous verdict of guilty, and u q-\-w p=:.Oll4 for
the probability of an unanimous verdict of not guilty.
ft
Making
have
the
same substitutions
in the formula in (82),
we
for the probability of a verdict guilty being correct,
from which 5 jurors out of 12

for the probability of its
dissent, zj^rr-j-f
and
ss<2
-=.^
being wrong.
Substituting the same values in the series represented by
W and W
from
w=0
in (83),
to
n=5,
and supposing n
there results
x 239122, whence
nearly.
This
is
Wp
W=
^--^
to
have
all
37
X7254,
126915984
values
W=
y^^ =
118
TT9
the probability that a verdict guilty, pro-
DECISIONS OF JURIES
AND TRIBUNALS.
127
nounced byamajorityof sevenagainst fiveatleast, is correct.

The probability of the same verdict being wrong, is there
fore
Yig
so that out of
know nothing
else,
119
verdicts, respecting
which we
than that seven at least of the jury con
curred in finding the accused guilty, we may expect one to

be wrong, or that one person out of 1 1 9 so condemned will
be innocent.
128
SOLUTION OF QUESTIONS
SECTION
VIII.
OF THE SOLUTION OF QUESTIONS INVOLVING LARGE
NUMBERS.
87.
The
compound events
from the combination of any number of

simple events, E,, E 2 E 3 &c. being (13) measured respec
which can
result
by the several terms of the developement of the mul

tinomial (p-j-^-f /-}-&.)*, the most probable of those com
tively
pound events
will
be that which corresponds to the term

Let us consider the
having the greatest numerical value.
E and F, the proba

of which are respectively p and q, and suppose the
number of occurrences to be 4. Neglecting the order of
occurrence, the different combinations, with tlieir respec
case of two simple contrary events
bilities
tive probabilities, are the following
EEEE, EEEF,
Now
it is
bilities
EEFF,
EFFF,
FFFF,
evident that the numerical values of these proba
depend on the ratio of p to <?, as well as on the co

by which they are multiplied, and that values may
efficient
be given
made
to
and
</,
such that any one of the terms
the greatest or the least in the series.
may be
If we suppose
p*= q, and consequently /?=J, <7=J (since p-\-q=l) the

probabilities of the different cases become respectively
INVOLVING LARGE NUMBERS.

whence
it
129
appears, that the most probable combination

2
that which corresponds to
6/?
2
<?
is
or in which each of the
simple events occurs twice, the probability of this combina

tion being
while that of either of the simple events oc
-^
curring four times in succession
When
number of
the
is
trials is 5,
only
-J^.
the probabilities of the
several cases are respectively
p
which,
lOpSq
5p*q,
Wp*q*,
5pq*,
5
,
when p=q, become
&>
so that there are
A
two
bable, namely, that in
&
JS
&,
*V>
different combinations equally pro
which
occurs three times and
twice, and that in which E occurs twice and F three times ;

and of the six possible combinations these two are the most
probable, having in their favour a

as great as the
two cases
in
number of chances twice
which one of the events occurs
only once, and the other four times, and ten times greater
than the two cases in which either of the simple events oc

curs in each of the five
From
when h
is
pound event
is
that
trials.
these two instances
it
may be
inferred in general,
an even number, the most probable com

that of which the probability is represented
by the middle term of the developement of (p
when h
+ q)
h
>
and
an odd number, there are tv/o compound

events equally probable, and more probable than any other,
namely, those corresponding to the terms which occupy the
that
is
middle of the
series,
supposing in both cases

A
p=q.
This
therefore in the
=(l l)Xi)
an even number, the general expression
supposition gives (p-J-^)
case in which h
is
for the greatest
term
is
h(h-l)(h-2)
1.2.3
(h-^h+l)
~~
(a)
130
and when h
odd, the general expression for either of the
is
two equal terms, which are greater than any of the other
terms,
is
A(A-I)(A-2) ...... {A-KA+1)+1]
1-2.3
When p
88.
KA+I)
......
and q are unequal, the greatest term of the
h
expansion of (p-\-q) will not occupy the middle of the
series, but its place may be found by comparing two con
Let
secutive terms.
The
h=m+n.
general term of the
series then becomes
1.2.3 ......... h
1.2.3 ......
and the term immediately preceding
1
is
.2.3 ......... h
1.2.3 ...... (m+l)xl.2.3 ......

Dividing the
first "of
these
by the second, we get
quotient (m-\-\)q--np, which, therefore,
for the
the ratio of two
is
consecutive terms taken at any part of the series.

ratio
be greater than
as the dividend
has been
is
1,
If this
the term which has been taken
greater than the preceding one which
taken as the divisor
and
it
is
evident that the
terms must go on increasing, from the beginning of the se

But
ries, so long as the ratio in question is greater than 1
.
if the ratio
be
less
than
1,
the preceding term
the succeeding, and the terms will
become
is
greater than
less
they are nearer the end of the series. Let (m -|-
and
less as
1
q-r-np
and m-\-n=h, we have n=(h-\-\)q,

and consequently the ratio of any term to the next preced
then, since p-\-q=\,
is
ing
greater or less than
greater than (A-f- l)q.
ber therefore if (h
;
Now n
according as n
is
is
1
)q be a whole number, take ?i= (h
and the two terms of the
less or
necessarily a whole
series given
num
)</,
by the expansion
131
of (p-i-qY, in which the exponents of q are n

1 and n, will
be equal to each other, and each greater than any other
term of the series. But if (A+ 1 )q be not a whole number,
let
x be the nearest whole number
(h +1)7
and make
less
than
then the greatest term
;
(h
)</,
n=(h -f- )q
of the developement will be that in which the exponent of
is 71.
Since n=i(h -f.
p=
n+x =
7-
Now
)<?-#, we have
m-f-1
?"
by hypothesis
is
q(n
-j-
x) -~(h -f 1) whence
,
q
and therefore i
A-fl
are large numbers,

since
we
less
than
1,
= -m n+x
therefore if
have, very nearly, q :pzzn
m+n^h, mhp,
m and
:
m;
or,
It follows therefore, that
nzzhq.
the greatest term of the developement of the binomial

A
tna * i n which the exponents
ofp and 5 are to each

more
nearly in that ratio
p
than are any other two numbers whose sum is h. In other
words, the most probable combination of two simple events,
4- ?)
*s
other in the ratio of
and F,
in
to q, or
any number of
ber of occurrences of
is
trials, is
to the
that in
which the
num
in the ratio of their respective probabilities.
89. In the
same manner
it
may be shewn,
that
when
more than two simple events, of which one must

every trial, the most probable result of any number
there are
occur in
of
trials is
that combination in which the
number of repe
of each simple event is in proportion to its probabi

in a single trial.
Thus, the probabilities of the simple
titions
lity
events being respectively^,
pound event
is
that
q, r,
whose probability
term of the expansion of (p

its
argument
90.
hp
,
&c. the most probable
hg
,
hr
,
is
expressed by^that
+ g+/ +&c.)
f
com
A
,
which has
for
&c.
Having determined the form of the greatest term of
SOLUTION OF QUESTIONS.
132
the series,
to
its
we have next
numerical value
method of approximating
to find a
for its coefficient containing the pro
duct of the natural numbers from
to
inclusive, its di
becomes impracticable even when h is only

a moderately large number. The theorem which gives the
rect calculation
approximate value of
this
product
is
known by
the
name
of
Stirling s Theorem, having been discovered by that mathe
As
matician.
we
its
investigation
shall not stop to give
Treatise on Differences
it
is
as follows
1
and Integral
is unit,
is
1
Calculus, p. 568.
*=*^V2^(l+
the
or the
comes very
is
The theorem
i + HI? +
&C
>
number of which the Napierean logarithm

number 271828, and TT the ratio of the cir
cumference of a
When
John Herschel,
Elementary Treatise on the
Sir
Let x be any number, then
2 3
where
a matter of pure analysis,
and Series, by
in the translation of Lacroix s
Differential
is
here, but refer the reader to the
circle to the diameter, or 3-14159.
a large number, the term divided by 12# be
small,
and the
series within the brackets
considered as equal to unity.
may be
In this case, then, the for
mula becomes
1.2. 3
which gives a
x=*<r-**j2^
approximation in most cases. If,

for example, #=1000, the result will be within a 12000th
part of the truth.
Now,
let
sufficient
E and F be two events
one or the other must happen
in
their respective probabilities,
and
of such a nature that the
every
trial
let
p and
q be
the probability that in
Stirling s investigation of the theorem, or rather of its equivalent,
to find the
sum of
the logarithms of a series of numbers in arithme

given in his Methodus Differentialis, p. 135.
tical progression, is
w + ft=A trials, E will happen m times and F

1
When
times
then
.2.3 ......... 6
m n
-2.3 ...... n P *
and A are large numbers, the value of this

may be computed from the above formula, which
w,
coefficient
ra
by (12) we have
133
n,
gives
.2.3 ...... A=
1.2. 3 ...... m=
1.2. 3 ...... 7z=rc
whence
This expression represents any term of the series (p-f- 9)*.

greatest term, which corresponds to the most probable
The
result, is
(88) that in which
the ratio of
to q, or
m and
greatest term therefore be denoted
P
we
n are
to
each other
when m=:hp, and n=hq.

by
that
is
be the chance of the most probable result of h

shall have
in
Let the
to say, let
trials,
and
P =V(h+2nmn),
or P =<v/(l-j-27r^).
formula shews that the absolute probability of that
combination which has the greatest number of chances in
This
its
last
favour becomes less and less as the
increased
for the fraction
which the probability
is
1 -f-/*,
to
number of
trials is
the square root of
proportional, diminishes as
is
in
creased.
91.
As an example, suppose
times in succession.
a shilling to be tossed 100
In this case
p=q=, hp=50, hq50,
and the most probable result of the trials is 50 times head

and 50 times tail. We have then ^=100, #z=50, and
4/(h-t-2irmri)
1 -j- \f(5Q-rr)
for the
measure of the probabi-
134
lity that
found =.07979
is
although 50 heads and 50
100
its
way exactly. On
whence it appears,
the event will happen in this
culation, this
trials
tails is
cal
that
a more probable result of
than any other combination which can be named,
absolute probability
The probability
is
measured by a very small
fraction.
of the contrary event, or that there will not
be thrown 5 heads and 5
tails
07979
exactly, is 1
= .92021,
so that the odds against the event are about 92 to 8, or
Had
to 2.
the
number of trials been 1000,
of 500 times head and 500 times
tail
23
the probability
exactly, though
more
likely to occur than
any other combination, would have been
found 1 -r- V(500?r)
that
is
to say,
V 10 times, or rather more
than 3 times less than in the former case.
In general,
when
the chances in favour of the simple events are equal, the pro
bability of the combination
than any other,

the number of
92.
is
which
more
is
likely to
happen
inversely proportional to the square root of
trials.
The formulae
in (90) enable us also to
determine the
term of the developement of (p-\-q)*

to any other term of the series, and consequently the rela
ratio of the greatest
tion of the probabilities of the different
Let
: :
q,
whence m=hp and
note the probability that in h
and
the event
let
events.
P* de
will occur
x) times, and the event F (n

x) times, the probabi
of the simple events E and F being respectively p and
(m
__
lities
q.
trials
compound
nhq,
By
(13)
we have
1
3 ......... h
1.2.3 ...... (m
a?)
x 1.2. 3 ...... (n
which by (90) becomes
_
~~
.^
w ie nce,
}
substituting m-^-h for p, and n-^-

for ^,
and leaving out the
tor
and denominator, we
P,
factors
common
135
numera
to the
find,
^/Y
Now log (m
#)
m +*
*;=:(
log
(m
a?)t= log
therefore log
(m
x)~
m + x~*
m-\-x
J)log(/w
a?)
and
--m 2m
y
sc
&c.
5
2
=
x^
(x +
+ &C J
2^2
2
3
whence, neglecting terms divided by m m &c., m being
supposed to be a large number in comparison with x,
,
-*=(
therefore,
m+x
on passing to numbers,
.fl
2m
snce
or
2m
=
2m
In like manner, by changing
we
m into
/z,
and
a?
into
get
Multiplying the
first
of these two expressions by
and the second by w n +-r we have

,
136
whence, substituting these values

glecting the quantity divided
\ _^L._^_
it h
Px = v 2n =
e 1m
\lTrmnJ
The term of the

value of P*
is
greatest term
series
(p+q)
that
which
and
it
is
P*, and ne
in that of
by mn,
It
/ ( -
*r
***
2mn.
\27rmnJ
which corresponds
to this
places to the right of the
has been shewn, (90), that the great
term has
for its expression V(h-r-2Trmn) ; therefore the

term
being denoted, as before, by P and the term
greatest
which comes after it x places by P^, we have
est
that
is
to say, the probability the event
times and
of
its
fail
n times
to
e~
m -f
trials, is to
will
+ x)
times in
Since the numbers

expression ehx^rZmn^
have been the same
happen
the probability
x) times and failing (n
happening (m
the ratio of
in
m and n
jt
if,
js
enter symmetrically into the
ev icl en t that the result would
instead of seeking the ratio of the
greatest term to that which succeeds
it
by x places, we had
sought the ratio of the greatest term to that which precedes
by x places. Hence if the most probable result of m -f- n

trials be that E will happen m times and fail n times, the
x times and fail n + x
probability that it will happen m
it
times
is
the same as the probability that
times and
The
fail
it
will
happen
-\-
times.
following example will suffice to shew the applica
thrown 6000 times, required

the probability that the number of aces turned up will be
exactly 960 ?
tion of the formula
die
is
Here p, the chance of throwing ace, is J, q=$

7i=6000; whence m=hp= 1000, and n=hg=50QQ.
have
first
to find
and
We
the chance of the most probable result,

or of
1000 aces.
By
we
get log
=8.14050, whence
e 7u2 f2mn
P =.0138.
"
The
2
performing the operation indicated by the logarithmic
tables,
37
P =\/(h--27rmn) whence,
P =V3 V(5000x 3.14159).
(90),
substituting the above values,
On
of
calculation
=hx 2 +2mn. We
have
is
as follows:
x=l 000
Assume
960=40.
log
40= 1.60206
2
3.20412
log h
6000=3.77815
log
2
log 7** =6.98227
log 2mn= log
10,000,000=7
2
=9.98227
.43429=9-63778
*
log
log e=.43429, log

2
log
t*
x-43429)^9.62005
(*
x .43429=* 2
log
e~
.41692
addlogP =8.14050
log
therefore
P =.0053,
actly.
When
h,
is
against this
53, or nearly 188 to
93.
which
;E
The odds
P x =:7.72358
the chance of 960 aces ex
event are therefore 9947 to
1.
m, and n are large numbers, and x is small,

-h*^ 2mn differs little from
unity, and it de
the exponential e
creases slowly as
rison of
m and
e- l =
n.
increases, so long as
Suppose
m~n
so that if
is
it
compa
becomes
we assume m=100,
10th term before or after the greatest would
3d part of the greatest.
small in
and x^z^/m^
still
the
exceed the
But when x becomes greater than
138
A/n the exponential, and consequently also the terms

which are multiplied by it, begin to diminish with great
-v/m or
rapidity,
and the diminution
is
more rapid
as
increases.
Ifw=/z=100, and #=50, then the exponent &r 2 ~2ww=25,

so that fr***-&mn 9
25
a quantity which
is
altogether
We may therefore conclude generally that when
insensible.
= l_i_g
a large number, the principal terms of the developement

Q+gO* are those which are near the greatest term, and
is
of
h may be taken so large that the terms towards the

beginning or end of the series may at length become smaller
that
than any assignable quantity.

94. From the proposition which has
strated
ticular
it
now been demon
follows, that although the probability of that par
compound event which has the greatest number of

its favour is very small when the number of trials
chances in
is
great, yet
on account of the rapid diminution of the terms
towards the beginning andendof the

paratively small
greatest,
series,
the
sum of a com
number of terms taken on both sides of the
may be very much
terms of the series
greater than
the remaining
all
and, consequently, there will be a very
great probability that the
compound event
sented by one or other of those terms.
will
be repre
This consideration
leads us to one of the most important questions in the theory,
namely, to determine the probability that in a large number

of trials, h, an event E, which must either happen or fail in
each
trial,
trial is
and of which the chance of happening in any

/ times, and not
p, will happen not less than hp
oftener than
hp+l times
or,
making hp=zm, hq=n,
to de
termine the probability that the number of occurrences of

E will be included between the limits nf=zl.
Let x be any number between

probability that
will
occur
(m
and
/.
Then (92)
x) times and
fail
the
(n + x)
i*K_
yw
U/fc-/
A^/7tf
=.P e- " ^ 27W "(where P =V(h-s-27rmn). Now if

this expression we make a; successively equal to each of the
times isP x
in
numbersO,
of
p2
1, 2,.../,
we shall have
happening
and the sum of these
E
m
139
TW,
1,
the respective probabilities
2,
times in h
trials
happens not oftener than

I times.
The same suppositions with respect to x
give the probabilities of E happening m,

times, the
be the probability that

and
not seldomer than
times,
probabilities will
-J- 1
-f-
2, .
. .
will
-f-
sum of which will be the probability that E happens
not seldomer than
m times,
and not oftener than m-f-/ times.
Adding, therefore, those two sums, and deducting
the pro-
bility which corresponds to x=0, on account of its being in

cluded in each sum, and therefore having been counted
twice, the result will be the sum of the terms of the binomial
(p+q)
comprised between, and including, the two terms

m ~l
first has for a factor jo* , and the last p
,
of which the
and
will therefore express the probability that the
occurrences of
this probability
sum of
all
the values of
Px
writing for
P^ and
h
let
Let
SP* represent the
obtained by substituting suc
cessively 0, 1, 2, 3,..,/ for x,
whence,
m=zl.
will fall within the limits
be denoted by R, and
number of
we then have R=2SP^

,
their values,
\ _*f!
27rmn
95, In order to find an approximate value of this expres
sion
we must have
recourse to a formula first given by Euler

sums of the kind denoted by S into definite
(for which see Lacroix, Traite du Calcul Differen-
for converting
integrals
tiel et
Integral, torn.
Differences, p. 513).
x,
the formula
is
iii.
p. 136, or
Assuming u
as follows
Herschel
Treatise on
to denote a function of
140
/*
bw = /udx +
u
2
t/
,
On making
*/m
&c. 4- constant.
k-
u=Px P er-***42mn
e -hx2~2mn. therefore, if
than
du
+- ^5-3
2 b ax
we g ntj
we suppose
a?
to
.._.
mn
ax
be not greater
or </n, this differential coefficient
is
of the order
may be easily shewn by substituting hp for m, and

and may be rejected, since h is supposed to be a
ri),
l-^-h, (as
for
hq
very large number.
SPX = P
The above
e-hx ^ 2mn dx
and on supposing
#=0
therefore the constant
equation therefore becomes
P e-h * ^ 2mn + constant
this gives
is
equal to
0=
|P
^P
+ constant,
and we have
Assume t=x\/(h-t-2mn), whence dt=dx*/(h--%mri)

substitute these in the
reason of
above equation, and
The
and x=l.
the limiting value of t, that
is,
obtain
must be taken between the
When #=0 we
when x=l, then t=li/(h-+- 2mn)

therefore
P we
R=2SPa.
integral in this expression
#=0
becomes by
=</(h--27rmn),
whence, from the equation
limits
it
have also
=0, and
therefore assuming r for
making T=/t/(A-*-2m), and
l=T+/(2mn--h), there
results
o
for the probability that the
fall
number of occurrences of E
between m=^r^(2mn-^-h)
or,
replacing
will
and n by
INVOLVING LABGE NUMBEKS.

hp and
of
to
hq, the probability that the
one of those
number of occurrences
hpzZT^(V,hpq), or be equal
limits.
This expression for
has been found by neglecting
quantities of the order of smallness
-t-h ;
the greater the value of h the more nearly
accuracy, but
96.
141
it is
only rigorously true
The integral^"
dt
is
it
consequently
approaches to
when h
is infinite.
computed as follows.
Develop
tz
ing the exponential e~ in a series of the ascending powers of
2
,
and integrating the successive terms between the

and t=r, we find
limits
=0
a series which converges rapidly
when
In the contrary case, however, or

unity, the series
by a
is
divergent, and
is
when
it is
less
is
than unity.
greater than
necessary to proceed
Let the factor e~^ be multiplied

we have then
method.
different
and divided by
t;
r tr**dt=r^<r
and on integrating by parts
Repeating the same process on the last integral, and so

on with the last after each succeeding integration, the fol
lowing series
When t=
comes
we have
is
obtained,
infinity the right
hand
whence between the
side of this equation
limits
=r
and
be
infinity,
142
a series which converges very rapidly when r is greater

than unity. Now the value of a definite integral between
and
the
infinity is obviously equal to its
first is
and
taken between
r and infinity
that
is
T,
two
parts, of
which
and the second between
to say,
oo
/oo
e^dt
well
is
known
to
have
for its expression
therefore
r
may be computed from
so that the integral
above
series, according as r
The
integral
mathematics.
is less
either of the
or greater than
1.
is of great
importance in the higher
occurs in the investigation of the path of
fe^dt
It
a ray of light through the atmosphere, and of the law of

the diffusion of heat in the interior of solid bodies, as well
as in the determination of the degree of reliance that
may
be placed on the results of astronomical observations, and

generally in most of the more difficult and important ap
plications of the
values from
by Kramp,
theory of probabilities.
=0 to
at the
=3,
for intervals
each
table of
its
= 01, was given
end of his Analyse des Refractions Astro-
/OD
tz
dt=^-r
is
ascribed by Gauss (Theoria
>
Motus Corporum Ccekstium,
et*=z,
the integral
the limits
2=0
is
p.
212,) to
transformed into
Laplace.
{J
rfzflog -
On making
J
between
and z=l, the value of which =^\/*> nad b een

See Le-
Memoirs.
given by Euler, long before, in the Petersburg
301.
gendre, Exercises du Calcul Integral, torn. i. p.
143
nomiques, Strasburg, 1799-
In the Berliner Astronomisches
Jahrbuchfor 1834, there
also a table of its values
is
tQ t2 (for the same intervals) multiplied by

to
with their
first
and second
cilitating interpolation.
from
2-^-/y/7r,
differences, for the purpose of fa
This
which appears
last table,
have been derived from that of Kramp, and which
is
to
imme
diately applicable in the calculation of the probability R,
we have extended
As
sent article.
to
t=3, and given
at the
the function which
is
end of the pre
thus tabulated will
occur frequently in what follows, we shall

venience in printing, denote it by 0, that
in future, for
is
to say,
we
con
shall
assume
the two forms being equivalent in consequence of the above

equation.
97. Some very important conclusions follow immediately

from the formula in (95). The quantity R denotes the pro
bability that in a very great number of trials 7^, the event
E, of which the a priori probability in any
occur not seldomer than
Tip
hpJf-l times, or that the
number of
times,
trial is p, will
and not oftener than
its
occurrences will be
included between the limits hpizl, or at least be equal to
one of those
limits.
Hence
also denotes the probability
that the ratio of the occurrences of
whole number
to the
be included within the limits pi=.l--h. We

have assumed T=l\/(h+2mn) ; but m=kp and n=hq ;
of
trials, will
whence l=r\/(2hpq), and con

sequently l--hTV(%pq--h). Now, if we suppose r to be
therefore T=l-r-*/(2hpq),
constant, so that the probability expressed
the same, then
and q being given,
by
/ is
R may remain
proportional to
144
the square
of A, and consequently the greater the
root
number of trials the
smaller will
be
in proportion to that
number of trials be 1000, and we

Thus,
have a given probability R that the number of occurrences
number.
of
if
the
not differ more than 10 from the
will
the most probable of
take 100,000
that the
10
all
(that
we
trials,
shall
number which
is
from lOOOjs), then if we

have the same probability R
is,
number of occurrences of E will not differ more than
V 100= 100
from the most probable number. But

in 1 000 is 1 - 1 00th of the whole, whilst a dif
a difference of 1
ference of 100 in 100,000

the ratio of
to
1-1 000th of the whole,
is
h becomes smaller and
of the occurrences of
to the
and thus
smaller, or the ratio
whole number of
proaches nearer and nearer to p, as the
trials
number of
ap
trials is
and the experiments may be repeated until the

difference between p and p-^=.l-~h, in respect of a given
increased
probability
be
less
If,
which may be as great as we
on the other hand, we suppose
then r
trials.
is
But
as r increases, 0,
to the table, that
have
l-s-h to
be constant,
proportional to the square root of the
it is
it
may be
only necessary to
0= 9999779)
number of
and consequently R, approaches
nearer and nearer to unity, (and
to
please, shall
than any assignable quantity.
seen, by referring
have rr=3 in order
whence the number of
trials
A may
always be increased until we obtain a probability approach

ing as nearly to certainty as we please, that the number of
occurrences of
(hp-==.l)
or,
will
which
is
be comprised within the given limits

the same thing, that the ratio of the
to the
whole number of
trials,
from p, the probability of E in a single trial?

more than a given quantity l-r-h which may be less than
shall not differ
any assigned
fraction.
This
is
the celebrated theorem
145
which was demonstrated by James Bernoulli
in the
Ars
Conjcctandi.
98.
cal
The
application of the preceding results to numeri
examples,
rendered extremely easy by means of the

From the formula in (95) we
is
table of the values of 9.
have the probability
that the occurrences of
in
trials will fall
between
limits 7</?rz/, the relation
within the
and r being given by
the equation lT*/(2hpq).

If, therefore, we suppose / to
be given, r becomes known, and the corresponding value
is found from the table ; and,
of
be as
conversely, if
sumed,
is
limits / are
we may
given by the table, whence the corresponding

With respect to the quantity P e~ rZ
deduced.
observe that
it
denotes the probability that the
E will be hp -{-/or hp I
and
is therefore
precisely (92),
always a very small fraction
h
when is a large number (90). It may be regarded as a
correction of 0, which in most cases might be omitted with
number of occurrences of the event
out sensibly affecting the result

large, or
it
/ is
into account.
directly as in the
in
example
avoided by increasing
but when h
number of
(92)
ing hp^=i(x-\-l), and
but this labour
tions
it
may be
within
the
be the probability
be included within
let
the probability of the limits be
let
and
corresponding values of the integral.

P.,.
not very
E will
Thus,
arrivals of
the limits lip^=.x, and
giving
is
it
so as to include
r,
limits of the integral 0.
that the
becomes necessary to take

In such cases its value may be computed
a small number,
be respectively the
We have then,
the same signification as in (92), the two equa
146
and the difference
of these two probabilities
is
ob
viously the double of the probability that the result of the
be either (hp -|- x -f- 1 ) times E, or (hp x 1)

But the chance of either of these events
trials will
times E, exactly.
being Pî,
is
large,
we have
therefore
R -^-R=2P.r+
P* and P^+i are very
to each other, (their difference

tities omitted)
is
0^2?^,,
(0
_j-0)
so that if
corresponding to
But
as the interval
small, half their
of
we
or
/+ 1,
2P X
P x =(>
e). Substi
we get R=J
e-f-P^.,
sum will
half their
and
we have
give R.
in the table is
will not differ sensibly
corresponding to /+^>
equal to R, and
quan
take from the table the values of
/and
between
sum
Now, when
0, and also 2P^ +1
tuting this value of P^ in the equation
and
i.
and very nearly equal
in fact of the order of
R=
hence R
and consequently
small,
whence
always
from the value
this value
of
is
the following rule for determining
the limits corresponding to a given probability, or vice

versa :
When
the limits are assumed, find r from the equation
l-\-jjr\/(2hpq)
then the value of
in the table, corre
sponding to r is the probability that in h trials the number

of occurrences of the event E, the chance of which in a sin
gle
trial is
p, will
lie
Conversely, when
lue of t in the table,
within the limits hp-==.l both inclusive.
assumed, find the corresponding va

by means of which the limit / will be
is
/+ ^=rV(2/ipq). It is obvious, that

and the probability
be both assumed, then
h may be determined from the same equation,
r 99. We will now give some examples of the application
given by the equation

if
the limit
of the preceding formulae.
Suppose
pq=^ and A=200, and
assign the limits within which there
let
is
it
be proposed
a probability
to
=J

that the
Now,
it is
easily
r^/ 1 00= 1 Or*
for
e=l we have
=4-769, and J=4-269.
/+
a shilling 200 times,

that head will
found from the table that
4769, whence
In this case
will fall.
the equation 1+^ =T</(2hpq) becomes /-f
147
On tossing
more than an even wager

turn up net seldomer than 95 times, and not
it is
therefore
oftener than 105 times.
Suppose p=q=i, /*=3600,and

exceed the
/4-i
1800r:30.
limits
let it
be proposed to assign
not
will
In this case the equation
= T-v/(2/^) becomes 30-5 = rv/(2x 900) = 30rv/2,

= 30-5-f-30-v/2 = -7l89; and the table gives
whence
=-6907= ff
nearly.
Hence
in tossing
a shilling 3600
times, the odds are 28 to 13 that head will not turn
301770
times.
Neglecting the second term of
and taking simply /rrlOr, the table gives

is
up
+ 30=1830 times, nor seldomer than 1800
oftener than 1800
e= 6827,
(95)
which
the solution given by Demoivre, p. 245.
Suppose />=|> y = $> an d let

trials must be made
how many
to one that the
it
be proposed to determine
it
may be one
in order that
will
not
differ
more than 10 from the most probable number.

For
we have r= 4769 therefore the equation
Jl+l=rV(Zhpq) becomes 10-5=-4769vXlO&-f-36), whence
;
On computing this formula h is found

Say 1746, .J of which is 291 ; and it follows that
if a die be thrown 1746 times it is an even wager that the
/*=3-6(10-5-:--4769)
t=t:l745-2.
number of aces
will fall
between 291=^10, that
is,
be
tween 281 and 301, or be equal to one of those numbers.

In (92) we found the probability to be -0053, that in
6000 throws of a
die the
number of aces
will
be exactly
148
now be proposed to
6000 throws the number
Let
960.
that in
it
960 and 1040,
/>=,
and
>
<?
that
is,
assign the probability 0,

lie
between
between 1000qp40. Here
^6000,
/=40
of aces will
the equation of the limits therefore
becomes 40-5=7V(10000-5-6), whence T=-405 v 6=-992,

corresponding to which the table gives 0=-8394.
The following question is discussed by Nicolas
Ber
Montmort s Analyse des Jeux de

Hazard, and is noticed by Demoivre and Laplace. From
the observations of the births of both sexes in London dur
noulli in the
Appendix
to
ing 82 years (from 1629 to 171 1) it was found that the aver
age number of children annually born in London, was about
14,000, and the ratio of the
number of males
males, was nearly as 18 to 17, the average

births being 7200,
in
which
to that of fe
number of male
and of female births 6800.
In the year
the greatest difference from this ratio took place,
the actual numbers were 7037 males and 6963 females, so

that the difference from the average
amounted
to
163.
Assuming, then, the comparative facility of male and female

births to
be as 18
to 17, required the probability that out of
14000 children born, the number of males shall not be greater

than 73M3, nor
Let 14000
than 7037.
less
This question
is
dice,
black, be thrown
evidently equivalent to the following
each having 35 faces, 18 white and 17

;
what
is
number
of white faces turned up, will be comprised between the

limits
7200dtl63.
We
have therefore
=1400,
j=J,
<7=i|. /=163, and the formula l+%=T</( 2hpq) becomes

1-955.
l63,5=V(2x 14000X 18 X !7)-=-35, whence
T=
The corresponding value of

which
is
is
found from the table= 9943,
number of white
faces shall
149
The odds
not be greater than 7363, nor less than 7037.

in favour of the event are therefore
175 to
100.
9943
to 57, or
about
1.
We
now proceed
to consider the case in
which the
probabilities of the simple events are not known, a priori,
but inferred from the results of experience. It was shewn

in (52) that the probability n of an event happening in
f
times, and failing n times in h trials, (h
-f n ), when
=m
it
has been observed to happen
^previous
trials, is
Now, when m,
n,
this
expressed by
larger,
and
fail
n times
in
equation
are large numbers, an approxi
ri,
mate value of n, more accurate
numbers become
times,
is
in
proportion as those
obtained from Stirling
theorem
(90), which for any number x gives \_x~\=x?e~ \/(2iTx}.

Applying the theorem therefore to the expressions within
the brackets in the above equation, and assuming
-
l*
we
obtain, in consequence
ofm-\-n=k,
(+
Let m =0m, n ~Gn, and consequently h =.6h
then taking
*
,
for
/"
out sensible error, since h

ber,) the equation
^h
h
is
,.,
m-}-n=h,
(which
mav be done with-
by supposition a
becomes
m(\
or, since
+ 6) m + mfn
^-w =A
/
,
large
num
150
r^TiTT
the same substitutions in the expression denoted
Making
by K, we
K= 1-^(1 + 0);
get, after reduction,
_ (m\
m
(
Vr+0U7
The
n now
value of
found,
whence,
\n
\f)
the probability that in a
is
future series of trials the ratio of the occurrences of
those of
will
be the same as
in the
preceding
are supposed to have been very numerous.
of
and
F had been
in
m\m
T)
m
n
future
-\-n
to
If the chances
given a priori equal to m-r-h and
n-i-h respectively, the probability of
times
trials,
which
times E, and n
would have been P==
trials
\ n/
(T)
between the probability P of

that combination of simple events which has the greatest
n ~-h
z=n-i-h), the relation
number of chances
in its favour,
known a
simple events are
the same combination
when
when
priori,
the chances of the simple events
presumed from previous
are only
the chances of the
and the probability of
trials,
is
expressed by
this equation,
101.
When
is
comes a very small

have then n P
.
with
h,
exceeds
is
1.
less
in comparison of h, 6 be
and may be neglected, and we
But when h f is a number comparable
very small
fraction,
than
The
trials
and
known
may have been there
chance of drawing a white
by WH-A;
it
reason of this
tents of the urn are not
the
diminishes rapidly
is
obvious.
a.priori,
is
when
If the con
however numerous
only a presumption that the
ball in a single trial is
measured
whereas, in the case of the ratio of the balls
being
151
previously known, the measure of the probability
As an
tain.
manner
instance of the
in
is
cer
which the proba
bility of an assigned series of future events diminishes,

when the probabilities of the simple events are inferred
from experience,
us suppose h
let
consequently n=P
-r-
in (91) that if a ball
-v/2=-7071
be drawn
at
=h, whence 0=1, and
XP
Now it
was shewn
random 100 times from
an urn which contains an equal number of black and white

that the result will be 50 white
balls, the probability P
,
balls,
and 50 black,
precisely,
we can
it
-07979-
It follows there
and
numbers of the two
sorts
contains from having observed that in 100
trials
only judge
of balls
is
of the urn be unknown,
fore, that if the contents
of the relative
there have been drawn 50 white balls and 50 black, the

probability
of that combination in 100 future
comes -07979
102.
The
X -707 1
trials,
be
05642.
result obtained in
termine the probability that

E in h* future trials, will not
(100) enables us to de
the number of occurrences of

differ in
excess or defect from
by more than a certain given

has been shewn (95) that in the case of the
the most probable number,
number
It
/.
and q of the simple events being given a

we determine r from the equation l=
probabilities
priori, if
the formula
gives the probability
limits hp-=*=.rV(2hpq)
that the ratio of
h).
1
is
that
or,
m will be comprised within the

dividing
by
h,
the probability
be comprised within the limits

Conversely ^whenp and q are not known,
to
will
This inference, though admitted by both Laplace and Poisson,

In a paper published in the Transactions of
not strictly correct.
the
Cambridge Philosophical Society,
(vol. vi. part iii.)
Mr.
De Mor-
152
but the event
has been observed to happen
triAls,
-f V(h-lTTmn)e-
gives the probability
R that p
is
m_^_T
These
r-
comprised within the limits
~h
approach more nearly to each other as h in

and when A is a large number, the ratios m--h,
limits
creases
maybe assumed, without sensible error, as
n-i-h
times in h
/2mn
~Ti~h*l
of
then
and
in
computing the probable
the chances
result of a future
A trials, provided, however, that h (though abso

a
When this
lutely
large number) be small relatively to h.
condition is not fulfilled, the assumption of m--h and n-~h
series of
E and F, might lead to consi

but an approximation to the limits corre
sponding to a given value of R may be obtained from the
a priori chances of
as the
derable error
following considerations
Suppose a large number h of events to have been observ

times E
ed, and that the result of the observation gave
and n times F.
Let a new
series of
new
that in this
series
trials
be made, and
the real chance of
is
gan has shewn by a direct analysis that
hf
in the case of
p and q not
known a priori, but made
equal to the observed ratios m-S-h,

w-f-A, the presumption of the true value of p lying within the limits
being
stated in the text
but
is
not, as there inferred,
---4n----- e
The
-f-
last correction to
is
smaller
6hpq \/ TT
than the former, and being divided by A, is of the order of quantities
It is right to state
that have been rejected in the approximations.
that the
method of simplifying the calculation of R in the direct case,

between limits corresponding to l+\ instead
integral
by taking the
of
/,
is
noticed, for the first time so far as
Morgan
in the
same paper.
we
are aware, by
Mr. De
and q of
the
we have then a given
kfpz=rV(%Jt pq),
observed in the
n~h, and the
Now,
first
probability
that
p and q
for
153
substitute the ratios
set of experiments,
limits corresponding to
namely, m-^-h and
become
which, therefore, are the true limits on the hypothesis that

the chance of
chance
in
a single
trial is
m~h.
But
as this
not certain, but only presumed, the limits require

to be extended in order that
may preserve the same va
is
Confining our attention to 0, the
lue.
expression for
=m -}-n
sent approximation), let h

is
in
the
and m! n =.m
:
sum of the terms of the binomial
which the exponent of

is
exponent
the chance of
term of the
first
R (the second may be disregarded in
is
hf
the pre
:
n,
then
(p-}-q)
from that
4- / to that in
which the
Now, when p and q are given apriori,

times E and n times F in h trials is P ;
/.
and when p and q are only presumed from the results of

previous trials, the chance of the same combination is n ;
and (100)
n is less than P
in the ratio of
to ^/(l -{-6).
In
manner, the chance of each of the other combinations

will be less in the
of E and F included in the integral
like
case of p and q presumed, than in the case of p and q given,

in the
same
ratio of 1 to */(\
+6).
But
it
has been seen
(93) that when h is a large number, the terms of the deh

velopement of (p -f- q) which are nearest the greatest term,
diminish at
first
very slowly
and, further, that only a small
number of terms on each
side of the greatest are required
to be taken, since I
than
is
less
+/m
therefore, without sensible error,

al to the
or */ri (95)
assume
number of terms included
to
in the
we may
be proportion
summation, or
154
not be changed if we include in the

summation a number of terms greater in proportion as the
value of each individual term is less.
Hence it follows that
that the value of
will
the limits must be increased in the ratio of
and the value of

bility that
the
y^ !-}-#)
corresponding to r will give the
number of events E,
in
future
/*
to 1,
proba
trials, will
be included between
The
103.
ample of the application of the
number h of individuals taken
m are
served that
may be proposed
following question
alive at the
at the
age A,
age A + a;
as an ex
Out of a given
last formula.
it
has been ob
required the pro
bability that out of h other individuals taken at the
the
number who
survive at the age
m :=/,
the ratio of in to
ed between
solve this question,
equation
/=
value of
From
p.
a
-f-
will
same age
be includ
h 1 being the same
as
M to h.
that of
To
555,
we have
^ \/(2h mn(l-\-6))
to find r
from the
and the corresponding
in the table, will give the required probability.
the table given in the article
it
MORTALITY,
vol. xv.
appears that out of 5642 individuals taken at the
age 30, the number surviving at the age 50, according to

the Carlisle Table, is 4397.
Taking those numbers as an
example,
suming
we have /*=5642, ^=4397, ^=1245; and
also 7^
= 5642,
whence
0=1
the equation of the limits becomes
proposed to determine
from the condition
7.
that of
Hence
it
0= J.
Let
it
be
In this
we have consequently
appears, that if
5642 individuals taken
as
+ 6)=^/2,
I=T X 62-30,
case the table gives T=*4769, and
fc=29
and \/(\
at the
it
has been observed
age of 30, 1245 die

before reaching the age of 50,
it is
155
an even wager that out of
5642 other individuals also taken at the age of 30, and sub
jected to the same chances of mortality, the
before reaching the age of 50 will
that
is,
lie
number who
die
between 1245r:30,
between 1215 and 1275.
104.
The
following experiment recorded by Buffon, in
Arithmetique Morale, affords an example of the ap

plication of the preceding formulae to the determination
his
of the probable existence of a physical cause from the

results of
a large number of observations.
piece of
money was tossed 4040 times successively, and the result

was head 2048 times, and tail 1992 times. Supposing the
piece to have been perfectly symmetrical, the most pro
bable result would have been the same
and
Let
tails.
it
now be proposed
number of heads
to assign the probability
by the experiment that the piece was not symme

and
that its form or physical structure was such as
trical,
to render head an event, a priori, more probable than tail.
afforded
In
this case
we have
=4040, ?w=2048, rc=1992; and by (102)
the probability
tion) that p,
between the
=.50693,
the
(or 0, neglecting the correc
unknown chance
limits
of head,
-^^.
r
=r X
~
and^Y?^
/2mn
.011124
is
Now
;
comprised
m
-
2048
therefore if
we
fi
assume r x011124=.00693, we
shall
have the probability
comprised between the limits .50693rr.00693,

p
that is, between two limits of which the least is .5, or onethat
is
This assumption gives T=.00693~.011124=.623 ;

is found from the table
and the corresponding value of
half.
=,62170.
is
Now
if
lie
between the above
evidently greater than ^
limits, its
but the probability of
its
value
lOO
between those
limits
greater than
-J
is
not the whole probability that
for there
is
a chance of
greatest limit, in which case also
The
than l.
probability that
the assumed limits
quently J
greater than
is
and
limit,
Hence
.37830=. 189 15.
greater than that of
.5,
tail
not
if it is
an equal chance of
or less than the least ;
exceeding the greatest limit
its
that
value will be greater
limits, there is
being greater]than the greatest
its
is
not comprised between
.62170 =.37830
is 1
comprised between these
the probability of
its
is
exceeding the
its
is
conse
the whole probability
or that the chance of head
is
.62 170 -f .189 15 =.81
085
is
and
the odds are therefore 81 to 19, or rather more than 4 to 1

that the piece
105.
The
was not perfectly symmetrical.
formulas which have been demonstrated in the
present section are immediately applicable to the determina

tion of the probable limits of the gain or loss
arise
which may
from undertaking a great number of risks with a given
expectation in respect of each.
The
following question has
A is
interested in a great
important practical applications.
number of similar enterprises, in each of which E or F must
When E happens he receives the sum

; required the
happens he pays the sum
that
his
or
loss
shall
be
probability
gain
comprised within
necessarily happen.
a,
and when
given limits
Let
be the chance of the event E, q that of F, and k

Suppose E happens m times,
be received will be ma, and the
the
number of enterprises.
and
F n
sum
be paid will be w/3, and therefore his gain will be

Let m=/ip, n=hq, then m times E and n times
the most probable result, and in this case the gain
ma
is
ma
times
the
sum
to
to
n&.
w.,3
becomes Ji(pa
then (98)
is
q&).
Find
T from/-j-
number
TV(2//p,y),
of occur-

rences of
will lie
I
happens hp
between the
corresponding benefit
/(a-f-/3)
and
the benefit
whence
is
is
if
is
(hp
F hq + l
times,
and
if
times, the
(hq-\.l)p=h(pa
l)a
happens hp -f
But
limits f}p+.l.
and consequently
times,
157
^,3)
F hq
times,
(hp^-l}a(hql)^=/t(p:iq3)-{-l(a + ^);
the probability that his gain, that
is,
the
diffe
rence between what he receives and what he pays, will be in

cluded within the limits h(pa <jr,3)rp:/(a-f-/3) both inclusive.
106.
The
following conclusions follow immediately from
this solution.
(1). If pa
tical
be greater than
advantage (however
be repeated a
sufficient
qfr so that
each
small) in
number of
has a
the risk
risk,
times, or
mathema
may
h may be taken
a sufficiently high number, to give a probability as nearly

equal to certainty as
we
please, that
any given sum, however great.

(2). Let there be two players
A s gain
A and
shall
exceed
B, whose chances
of gaining a game
p and q, and let /3 be
the sum staked upon each game by A, and a the sum staked
are respectively
by B, then pa
is
the mathematical expectation of
spect of a single game, and
qft that of
and
in re
if
pa be
greater than q$ (however small the difference) the game

may be repeated so often as to give rise to a probability
approaching as nearly to certainty as we please, that A s

gain shall become equal to the whole of B s capital, and,
consequently, that
will
be ruined.
of the
(3). If the mathematical expectations
two players
;3zrO, and the most probable individual

result of a large number of games, is that the gains and losses
on either side shall be the same. But if /be supposed constant,
be equal, then pa
then T
is
inversely proportional to
game may be repeated
until 0,
Vh- and consequently the

158
SOLUTION OF QUESTIONS.
gain or loss
shall
/(a+)
become
shall
as small as
be comprised within given
we
Hence
please.
bability that the gain or loss shall not
given
may be rendered as
limits,
follows that although the play

equality,
may be continued
it
be comprised within
we please ; and it
great as
may be on terms of
perfect
until a probability shall
obtained, approaching as nearly to certainty as

that one of the two players shall be ruined.
(4).
limits,
0, the pro
we
be
please,
The number
afford a given
shall lose the
of games which must be played, to

amount of probability that one of the parties
whole of
depends on the magni
his fortune,
tude of the stakes (a-{-/3); but whether the stakes be large

or small, the final result is the same.
When the stakes
number of games must be played.

As an example of this class of problems, we may
are small, a greater

107.
take the following question
and
with equal chances of winning, and stake
in
play
on
how many games must they undertake to play

that it may be two to one that one of them shall
each game
in order
engage
five sovereigns
100 sovereigns?
lose at least
Here>=^, q=^, a=5, /3=5, and Z(a-j-/3)=100, whence

The equation l+=r\/(2hpq) therefore becomes
/=10.
10-5=iV(-i-2), whence
odds being 2 to
h2 x (10-5)
table gives
by
is
-f-T
Now,
the
against the limits of the gain or loss
not exceeding 100, the probability
ceeding 100
of the limits not ex
^=-33333, corresponding
interpolation r=r -30458
in the above equation we find
A=r2376 8; so that
are played, the odds are 2 to
to
which the
substituting which
if 2377
games
that one of the players shall
10 games more than half that number,

and, consequently, that the other shall have gained at least 10
less than half, or that one of them shall have gained at least
have gained
at least

20 games more than the other, and consequently have gain
ed at least 100 sovereigns.
It is to
be carefully observed that this question supposes

A and B not to be balanced until
the account between
2377 games have been played. If the condition of the play

had been that it should cease as soon as A or B should have
100 sovereigns, the question would have been of an
entirely different kind, and a much smaller number of games
lost
would have given the same probability of an equal
The
108.
loss.
question just alluded to belongs to a class of
problems connected with the Duration of Play of extreme

difficulty, and which have given rise to some of the most
>
abstruse
and refined researches
in the
order to give an idea of the subject,
modern
we may
analysis. In
take the
fol
lowing question, which has been frequently consideced.

A and B, whose chances of winning a game are respec
A has m counters, and
tively p and q, play on these terms
:
n counters
has
and when
to B,
and the play
is
when
loses a game he gives a counter

game he gives a counter to A,
when one of them has lost all his
to cease
counters.
What
may go on
for ever, shall
shall
games
To
ters,
loses a
is
the probability that the play, which
be finished before more than k
have been played.
take a simple case, suppose each to have three coun
and
let
the probability be required that the play shall
be concluded with or before the ninth game. As the play

cannot end with less than three games, let the binomial
5
(p-\-q)
be developed, and the terms
P s + 3/> 2 7 + 3/^ 2 + q *
give the respective probabilities of
The
all
term
the cases which can

the probability of
arise in three
games.
the three games, the last term
gaining
all
first
is
is
the proba-
160
bility
of
terms
is
the probability that neither will win
or the chance that a
fourth
Bs
sum of the remaining two
gaining them, and the
game be
fourth will
played,
As
is
all the
games,
be played. Now, if the
chance of winning it, and
but these chances
will only exist in respect

q
of the fourth game, provided the play be not concluded with
chance
3p-q + 3pq*.
by p + q, the product
the previous one, the probability of which

Multiplying, therefore, 3p-q
+ 3pq
is
of the different ways in

gives the respective probabilities
which the four games may be gained by A and B, except
in which the play would have terminated
ing the two ways
with the third game. But the play cannot end in any of
these ways
for,
example, if B
gains three, the play cannot ter
taking the
counter before
gains a
first
term
for
gain back that counter, and three others

In fact, it is
five games must be played.
that
so
besides,
of
an
odd
number of
is
no
obvious that there
way
gaining
minate until
counters in an even
last
number of games,
The
or vice versa.
the chance of the 5th

product therefore expresses
game
2
reason ofp gf=l it is equal to 3p q

being played ; and by
2
of the 4th being played, as it obviously
chance
the
4- 3pq
>
ou^ht to be, since the play cannot terminate with the 4th.
be played, p is A s chance of gain
Again, if the 5th game
B
s chance of gaining it; multiplying there
and
it,
q
ing
fore the last product

result,
by p + q, the
different terms of the
namely,
of all the cases which can

give the respective probabilities
first
term is the probability of
The
5th
the
arise by
game.
A gaining 4 games and B gaining 1,
and the
last
term
is
the

probability of
gaining 4 and
161
These
gaining 1.
terms therefore are the probabilities of the play ending in
favour of A and B respectively with the 5th game, and the
sum of the
other two terms
is
the probability that the play
will not terminate with the 5th
another
By
game being
game, or the chance of
played.
pursuing the same reasoning
it
will
be evident that
on rejecting the two extreme terms of the above product,

and multiplying the remainder byjo-fg, there will result
the probabilities of the different ways in which six games
may be
played without the one player gaining
all
the
But as the play cannot termimate

with the 6th game, multiply again by p-\-q> and the result
counters of the other.
will indicate the probability
arise out of the 7th
of the different cases that can
game.
Rejecting the two extreme
terms, which give the respective probabilities of the play
being concluded in favour of A or B, and multiplying the

remaining two first by p -\-q to obtain the different proba
bilities in respect of the 8th game, and again by p-\-q, as the
play cannot terminate with the 8th,
27
of which the
of
and
we have
the product
jy + 81 jy + 81 />y + 27/>y,
first
and
winning
last
terms give the respective chances
at the 9th
game, and the sum of the
other two terms the probability that the play will not be
concluded by the 9th.

If
we now
collect the terms
respective probabilities
game, or sooner, we
which have been
set aside
and denote by a and b the

of A and B gaining at the 9th
in the successive products,
shall
have
162
a=p + 3p*q + 9/>Y + 27 jo

where the law of the
It
easy to
is
series
see
is
that
evident.
this
may be
process
whatever be the number of counters which
of games to which the play

as follows
which
is
of the two numbers
and
winning n games
reject the first
mn.
The
limited.
is
not less than the other.
n,
have
m and
Raise
h,
general rule
n, let
p -\-qto
m be
that
the power
term (which gives the chance of

and also the last if
in succession),
Multiply the remainder
(1i
n) times in succession
rejecting at each multiplication the
by (p-\-q\
last
applied
and
commencement, and whatever be the number,
at the
is
term of the product when
it
first
or
gives a combination which
would terminate the play in favour of A or B ; the sum

of the terms rejected from the left-hand side of the dif
ferent products gives the probability in favour of A,
the
sum of
and
the terms rejected from the right-hand side the
As
probability in favour of B.
the coefficients of the suc
cessive products are obviously formed
by adding the
coeffici
ent of the corresponding term in the preceding product to

that of the term immediately before
be written down
tion; but
bers,
it
it is
the products
it,
may
once without the trouble of multiplica

evident that when m, n, and h are large num
at
would be quite impracticable
to
sum
the series
formed of the rejected terms by the ordinary methods.

From the manner in which the series are derived, they are
called recurring series
a general theory of which was
first
given by Demoivre in his Doctrine of Chances, and forms

the most remarkable portion of that work.
109-
The
general problem
finite differences as follows
is
reduced
to
an equation of
Let yXit represent
As
expec-
INVOLVING- LARGE NUMBERS.

tation
when x games have been
counters to win, or
the next
game
has
played, and he has
counters in his hand.
his expectation in respect of that event

if
it
is
become
isjo; therefore
py x +\, ti-
On
next game his expectation

and the chance of losing it is q ;
therefore his expectation in respect of that event
Hence, according
still
A gain
loses the
become yx + \,t+i->
will
If
the value of his expectation will
y,+i,f_i, and the chance of his gaining
the other hand,
163
to the principles laid
down
is
qyx +i, +i
t
in (32),
a linear equation of finite differences, with three independ

ent variables.
It is therefore on the integration of an equa
tion of this kind that the
problem of the duration of play
ultimately depends, but the subject
cated a nature to admit of

in this place.
We
which forms the
of Laplace.
1
the
of
much
too compli
being satisfactorily explained

ourselves, with
to
the treatise on generating func-*
first
part of the Theorie Analytiquz
Lagrange, in vol.
first
is
must therefore content
referring the reader

tions,
its.
who shewed
of the Memoirs of the Society of Turin, was

that the investigation of the general term of
i.
a recurring series depends on the integration of a linear equation of

finite differences.
In vols. vi. and vii. of the Memoires presentes a
f Academic des Sciences of Paris, Laplace proposed a general method
for the summation of recurring series by the integration of such equa
and in the latter volume gives a number of examples of their

use in the more complicated questions in the theory of chances,
The subject
amongst which is the problem enunciated in ( 108).
tions,
was afterwards resumed by Lagrange in the volume of the Berlin

Memoirs for 1775, where he has given a more direct method than that
of Laplace, for the integration of the class of equations in question,
and also applied it to the solution of the principal problems proposed,
works of Montmort and Demoivre.

general solution of
the problem in the text is given by Ampere in a Tract entitled Con
siderations sur la Theorie Mathematique du Jeu> (Lyons, 1802.)
in the
RESULTS OF DISCORDANT OBSERVATIONS,
164
SECTION
IX.
OF THE MOST PROBABLE MEAN RESULTS OF NUMEROUS
DISCORDANT OBSERVATIONS, AND THE LIMITS OF PRO

BABLE ERROR.
In the preceding section
110.
class of questions
stant causes,
which apply
and supposed
necessarily happen or
fail
to
in
we have
considered a
depending on con
be of such a nature that they
to events
each experiment, and have
given formulae by which approximate results can be ob

tained
when
the numbers involved are so large that they
cannot be conveniently treated, or cannot be treated at
by the ordinary methods of calculation.

a more
difficult
We
all,
come now
problem, namely, to investigate
to
the pro
bable result of a large number of observations which have

reference not to the simple occurrence or failure of a cer
tain event, but to the
magnitude of a thing, susceptible,
within certain limits, of a very great or an infinite
num
ber of different values, equally or unequally probable, the

chance of any particular value being also supposed to vary
each experiment. On account of its immediate applica

determination of the most probable values of
astronomical and physical elements from the results of ob
in
tion to the
servation, this
is,
perhaps, in reference to practical utility,
the most important question in the theory.

111. Let A represent a thing of any sort
(as a line, or
lo
AND LIMITS OF PROBABLE ERROR.
an angle, or a function of any quantity) which may have

every possible value within given limits, or which may
be constant
in itself,
but of such a nature that
its
real
mag
nitude can only be observed within certain limits of accu
and suppose a great number of observations to be

made. The object is, in the first place, to assign the pro
racy,
sum of the observed
bability that the

limits,
values shall
given
of
to be known a priori
the law of the chances
is
unknown,
also the limits within
to determine
which there
is
true,
but unknown value of A,
when
from the
mean
value of
a given
amount
of probability that the difference between such
and the
within
and, in the second place,
observations themselves the most probable
A, and
fall
supposing the chances of the different values
mean value,
be contained.
shall
112. Let a and b be the limits of the possible values of
A, x a value of
that the
will
be
and
lib.
sum of
between a and
the values of
s exactly, s
Assume
/3,
and
the probability
given by h observations
being a given quantity between ha

A to be equidifferent, and
ae=,
a,
and
the values of
multiples of a certain constant
where
b,
are whole
o-
e,
and make
Pf=l), o-e=S, ie^X,
numbers (which may be
posi
whole number proportional

to x, and varying between the limits e=a, and ^=/3, and
If
which, therefore, may be positive or negative, or zero.
tive or negative),
and
i is
also a
A are supposed to be equally proba

chance of obtaining any given one of them, as x,
the different values of

ble, the
in a single trial
is
by the number of possible

and if we assume an
a-f-1)
unit divided
values, or equal to l-i-(/3
indeterminate quantity w, then (20) the

nations which give the
sum of h
number of combi
values of
equal to
the coefficient of that term of the multinomial
o-e is
lt>6
(or of the developement of
which the exponent of w

bability
is
actly
from
(Sw )*
rra to z:=/3) in
and consequently the pro

that the sum of the values of A will be s ex
is
a-
that coefficient divided
by
a -|-l)\
(/3
113. If the chances of the different values of

equal,
and
also vary in each trial, let p\
of the observed value of

probability of
being x
its
in the third,
is
only a single
and if
its
sum of
as
ways
that of
or
/3),
and
in as
o-e
p% the
being
when
there
and we have
(i
P=p
may
to
and
ts
2/? 2 ?#
served values of
ways
which the exponent of iv
if
the
equal to
in as
may be
and consequently
which the exponent of w
sum
ere
is
sum of
A=3, then
equal
the ob
different
many
the coefficient of the term of
is
the developement of the product
the
is
as the equation i-\-i -\-i"=.<r admits of different so
lutions,
in
and
of the two series represented
w)
in
In like manner,
ere.
give
different
many
that term of the product (arranged ac
cording to the powers of
bySpjW*
being any two
consequently, according to the theory of combinations,
in the first
possible to satisfy the equation z-f-z =o-;
the coefficient o
un
its
the two observations
the two values equal to
it is
be the value of A
to
e e,
value in the second,
numbers between a and

the
are
first trial,
p
Now when h= 1,
then
trial,
If 7i=2, then, assuming

trial,
being x in the
in the second,
and so on.
be the probability
number of
is
observations
equal to
the coefficient of
the sums
"2,
including
in the
all
ere.
2p 3 w
Generally, when
li
Sp^v"
being
s,
or
ere,
the probability
is h,
of the observed values of

ff *
11
Sp^w
of the
exactly,
developement of the product
values of
from i= a
to
is
AND LIMITS or PROBABLE ERROR.

6 ""
lt)7
(e being the base of the Napierean

above product be denoted by X.
the
and
let
logarithms),
We shall then have
Assume
iu-=.e
Now since P
is
ment of
product which contains the factor e
we
this
the coefficient of the term of the develope-
P
a
we
conceive the developement effected
series in
which
all
shall
the terms are of the same form.
tiplying both sides of the equation
Now by a well known theorem

r ~ ) ^|
269), e<
= cos
by
if
have
&c.
ff
art.
*~~1
Mul
*~ we
1
get
in trigonometry, ( AL.GEBRA,
<r)^+V
(<r
substituting therefore this value,
lsin((/-o-)fj
and multiplying by
dd,
the equation becomes
Xe-*t V
The
dently
(TT
I=
i"^=P^+P
factor
[cos (r
_-0*+\/^T
which multiplies
sin (r
} <#+
*)
&c
in this equation will evi
become zero when integrated from
6=
being the semicircumference to radius
1),
to
6= -j-rr,
the positive
and negative elements of the integral being equal, and con

The same thing also
sequently destroying each other.
takes place with respect to the following terms, which are
all
of the same form.
limits,
Integrating therefore between those
and observing that /e?#=27r, we

t
114. This value of

that the
sum of
and
Let
p.
and
let Gt
P denotes
the values of
find
the infinitely small chance
A in
trials will
be
s exactly.
be two integer numbers between ha and h$,

s will be comprised
denote the probability that
168
between the two
limits ju
cluded between ha and kb)
and
(these limits being in
j/e,
then Q, will be found by sub
stituting successively y, /*-f
1, ^-j-2,...!/
sum of
value of P, and taking the
all
for
above
in the
a-
the resulting terms.
This substitution gives the following series multiplied by
under the sign of integration
On multiplying the series now found by

(=2/^7
sin 10),
and the
all
the terms of the product, excepting
destroy each other, and the

the terms becomes simply
the
first
therefore on
making the
multiplication
we
last,
now
substitution,
Q the
obtain for the value of
of
and performing the
and dividing by 2*J
indicated,
sum
sinj#,
equation
115. In order to simplify the expression for Q, let the
number of
possible values of
conceived to be
infinite,
comes
infinitely small,
finite,
p,
and
tutions also
and
v infinitely
be made
p=^
in
within the given limits be
which case the constant
be
therefore, since the limits are
Let the following
great,
substi
8,
vc=. ^-J-5,
6=fZ,
6 being positive in order that v may be greater than /t,

agreeably to what has already been assumed. On substitut
ing these expressions in the above equation, the limits of
the
new
variable z will be
cr
infinity; for e
having been
supposed infinitely small, z must become infinitely great
when 6=7r.
and
v-j-
Now
become
since p
sensibl
and
/*
v are infinitely great,
and
i/,
whence we have
/*
Again, by reason of 6=ez,
IsinSz.
d6=edz
and
sin
^6=2dz--z.
we have
must be a very
being infinitely small, 6
small arc, therefore ^6
d6+
163
may be taken for sin ^0, whence

By means of these transformations
the expression for Q, becomes

i/
YZ v
\t>
AC
o; n
am
*~
o^
and denotes the probability that the sum of the h values of

will lie between i^îS.
116. It
is
now
denoted by X.
necessary to assign a value to the product

number of possible values of
Since the
between a and b has been supposed infinite, the chance of

obtaining any given one of them, as x, in a single trial, is
infinitely small.
this
Assuming
of X) and to vary in the different
by
<
na?
in respect of the nth
chance to be a function
be represented
In order to preserve
trials, let it
trial.
continuity in the values of A, this must be understood as

signifying that
value of
$ nxdx
is
the infinitely small chance that the
given by the nth observation will
x and x + dx.
The
function
<J> n
x, therefore,
lie
between
represents the
law of the facility of the different values of A. It is posi

tive for all values of x between a and b, and vanishes for
all
values of x less than a or greater than b
integral
unity
xdx taken from
fcf) n
a?
and
number n may
portant to remark, that whatever

to
xb
is
b,
the
sum of
all
always equal to
then,
xdx=pM whence
we have
the sums
(f>
A be
the probabilities in respect
of each observation must be unity or certainty.

assumption,
im
be, the
for since every observation gives a value of
tween a and
<p h
it is
xdx=p
Zp n e
li
(f)
From
this
xdx=p 2
(11 3) are changed
17(7
into definite integrals
and therefore, since 6=ez, x-=if?
and consequently i0=xz, we obtain
for the value of
X,
xb.
the limits of the integrals being x=a and

zx ~~ 1
=r cos zx-}- \/
1 sin zx, each of
By reason of e *
may be expressed in terms of the cosine and

The nth, for instance, becomes f(j> Hx cos zx.dx
these integrals
sine of zx.
-f
lf(j) nx sin zx.dx.
the integrals
(from
x=a
zx.dx = B
fan x cos
Now since /</><&?= 1
$ x can have only positive values, each of

whence we may assume
is less than 1
and
to x=z b),
cos rn
f<p n
sin zx.dx
= Rn sin rn
En
being a positive quantity, and r n an angle having

ways a real value. This gives
pz
whence
1, 2,
"
*J-\$
xdx=iK n (co$rn---^~\ sinrn)=R ne^
substituting
3...^,
and
successively
for
for the sake of brevity
the
TZ
al
V-
1
5
numbers
making
r^r, ............ +r w
1
we
get
XsYe*^""
and the expression
Q becomes
for
>+C
0= w y/ co
117.
two
The
others,
integral in this last expression
Now, on
equivalent to
namely
/Ycos(y-^) sin 8 z
sented by
is
+ */
/Y sin (y-tyz) sin 8
2;
attending to the nature of the quantities repre

it will be manifest that according as z is
Y and y,
r
positive or negative, n ,
negative, while
is
and consequently y
is
positive in all cases, since
positive or
Rn
is
always

positive.
Hence
ments of the
cos (y
first
is
^z)
17 1
always positive, and the ele
of the above integrals having thus the same
value and the same sign for the same value of 2, whether z
be positive or negative, the value of the integral from

oo
is the double of its value from
to oo
On the
-J-
to
other
hand,
sin (y
y and z have both the same
since
sign,
z is positive
tyz) is positive or negative according as
or negative, and the elements of the integral into which
z and
enters being equal for
#=
The
oo vanishes.
-j-
it
but having contrary
each other, and the integral from
signs, destroy
to
-J-z,
#=
oo to
expression for Q, is therefore
transformed into
00
Y
118.
fa
cos (y
TJSZ)
sin 8z.
The formula newfound cannot
in general
be integrat
ed by any of the known methods, but in the present case the

quantities denoted by Y and y are such that an approxi
mate value of
may be
more nearly equal

is
observations,
obtained, which will always be
to the true value as h, the
increased.
On
two quantities represented by
number of
adding the squares of the
Rn
Rn
cos rn and
sin rn
we
get
R B 2 =(/0 nar
If
2=0,
When
is
x,
this
cos zx.dx)
becomes
(fcf> n
sin
zx.dxf.
R n =/</> n#er, whence by (1 16), R n =
z has a real value, then
it
may be shewn
that
1 .
RM
x f be any value of A different from

then as x can only vary from as a to b, we have obviously,
less
than
for let
f
and/ftX B>mzx .dx =sjfy nx sin zx.dx,
and the above equation may be put under this form,
R & n =f<p nx cos zx.dx-f$ nx cos zx dx +f$ nx siii zx.dx-fQ nx sin zx .dx,
f$ nx
cos zx .dx =fip n x cos zx.dx,
whence
cos
z(xx )dxdx
f
.
172
the case in which
Now, excepting
is
integral
(f(f) n xdx)
less
is,
2
,
than unity.
equal to unity
n is
other values of
it
follows that
or
its
We
double
this
z,
>
Since, then,
it
when z=Q, and
and since
Y must
is
has been shewn that

less
than unity for
a quantity of the order
diminish with great rapidity
all
R nA
when
z,
equal #-f-e differs sensibly from 0, and even for very small
values of z becomes insensible
is
2=0,
than ff$ nx$ nx dxdx or less than

always
and consequently R,, is less than f<J> n xdx, that
less
may
therefore assume
when h
Y=e~~*
is
a large number.
an expression which
when 6=0, and diminishes rapidly

and becomes zero when 6 is infinite.
as 6
equal to unity
creased,
is
in
119- For the sake of abridging let us assume
k n =fx<f> n xdx, k
=fx
2
<}>
nxdx,
k"=fx*4>jcdx, &c.
x being always from

From known formulae we have
(the integrals in respect of
#= b).
cos
x-zza to
zx=
t
TC
substituting these series for cos zx and sin zx in the inte

grals f<f) nx cos zx.dx audf(p n x sin zx.dx,
&c., for the values they
sent, then,
n>
k" n
to repre
from (116) we have
R n cos rn = 1
and
and also k n k
have now been assumed
k "n
&c.
be seen presently that all the terms involving

higher powers of z than the cube may be neglected. Add
ing together the squares of these two equations, we get
2
it
JR n s=l
will
z\k
k\)
+z 4/
&c.; whence
f being independent of z. On dividing the second by the

there results tan rn =zk n
first,
z
lz k" n +z*kji
tan r n
whence by reason of r n
1*3 ("
3k n k
r n =zk n
the values of
R m =l
2
{z c
Y=2
z^
Now,
^=1
to
Y=
(/
icB )
let
2
,
z^hf
z* ffn
&c.
......
+ s 4/
&c.)=
and consequently
In like manner, since
&c.}
^=2z^ B--S 5 ^ +
values of cn ^ n gnt from
therefore
include
the
mean
all
fl
values of those quantities,
be denoted by c, k, g, that
-c n =hc, 2/ n =M, 2^
7l
and make
*)\+ &c.
+ &c.) But we have also assumed
hence log
the sums
w=/i
therefore,
&c.} (by reason of the formula
;K)
y=r^ +r 2 +r 3 ... +rA =Srn

&c.
z 2 cB
J(K n
Y=e
=2 \z c
R =2log (1
log
2
1
+ 2k
YsR^R^xRa
hypothesis (116)
log R B =R B
(117)
&c.; r n =zk n
+*4/
z*(f
tt
&c.
and rn become respectively
Now, by
Rn
* cB
therefore log
tan 3 r n -f- &c.,
we make
therefore,
If,
173
f
is
to say, let
=%,
2
2
hf"=2(f ^C B ), and we have 6 =z /ic
By reverting the series the value of z is found
also
Jf &c.
in terms of 6;
*=
namely
But the second term of
-^
this series,
^^
&c.
being divided by h*/ft 9
and h being by supposition a large number, is very small in

comparison of the first, and may be neglected as insen
sible.
All the succeeding terms of the series are divided
by higher powers of A, and may therefore be rejected a for

tiori.
Confining the approximation, therefore, to terms of
the order
its
\--/h, and
powers enters as a
rejecting
divisor,
all
those into which h or
we have z=6--i/(hc), and
likewise dz-~-z=dd-^-0.
From (116) we have
also
y=2rn =zZk n
z*2g n +
&c.,
174
therefore in consequence of the above transformations,
z*hg-\- &c.
zliR
found
and on substituting
terms of
in
consequently?/
tyz
gb* -i-c\/(hc), and
y=k&^(h--c)
6,
= (hk
In order to deduce from
this
y=
for z its value just
^tf-j-yÂc)
an expression
g6 c-r-\/(hc).
for cos
(ij
\/AZ),
u and v denote any two arcs, then by trigonometry,

cos (u
t?)=cos u cos v-{- sin u sin v. Suppose v to be small,
let
and
let its cosine
and
cos (u
y)=cos u
--1
be developed
sine
stituted in this equation
will
it
cos
u+
in series
and sub
become
&c. -{-v sin u
1 -v
--
fc
sin
u+
&c.
^)0-i-v (^ c ) v=yd -+-c</(hc),

and_rejecting as before terms of the order 1 -r-h, we have
whence, making
If
we now
found
we
u=(hk
substitute the values of Y, z, dz, cos
in the last three paragraphs in the value of
^r)
obtain the following expression in which the largest
terms omitted are of the order

is
(jj
Q (1 17)
more
2
accurate as
_*
is
c
1 -r h,
and which therefore
a higher number, viz.

,,
sn
A/ (he)
1
20.
As no
been made with respect to

is a mean between /*eand j/e,
restriction has yet
the value of ^, excepting that
it
and therefore included between ha and hb (115), let us

now assume ^l/=hk. This gives cos (hk ^)=1, and sin
;
and the equation becomes

2
- //"<
o
77^7
sin
V(hc)

which
lues of
is
sum
175
of the observed va
between hk=+=..
will fall
The last step in this investigation is to reduce the

to a known form, which may^be accom
integral now found
u be a new variable, then by means
Let
follows
plished as
121.
of the trigonometrical formula cos
But
^-fwfl^/ZT
"
^Hi
v=d
therefore,
u\e
^~^-\-\e~
(whence dv
^~\
assunie
= d6), then
When 0=0,
then v =:
therefore, if the integral in respect of 6 be
is infinite
taken from
6=0
to 6
i^^/
oo,
1,
and when 6
the integral in respect of v must
\u ^HTl to v =: oo.
suppose u to be negative, we shall have
be taken from v ==
If we now
manner
this case
is infinite,
f^
^ J^de =
being from v
%?-**fe~**
+~2 M
y/
dv, the
in like
limits in
Hence
to v ==
1
on the right-hand side of

M
this equation, the first being taken from v
1
a */to infinity,
AM /
to infinity, and the second from v
-j\
But the sum of the two
integrals
obviously the double of j*e~* dv from v=0 to

and we have therefore
(96) equal to \^TT ;
is
Let both
or
cos
sides of the equation
integrated from
v== x,
w=0
ing that/^cos (u6)dd
to
be multiplied by du, and
=-=- \/(hc)=u
sin (z^)-j-0,
we
shall
then observ
have
176
Comparing
equation with
this
that
Q=(l -i- v/7ry*o

becomes when
V-* w2 G?w.
Now,
let
u=u =8~\/(hc)
/v
_>2
for the probability that

will
A,
d^
s,
or,
the
we
(110),
then iw 2
we
2
,
du=2dt,
have, finally,
Q=l
sum of the observed
be comprised between the limits
values of
and ^-}-&,
that is, between Mrp2r-v/(^c); or, that the arithmetical
of
the
all
observations,
namely
expression
now found
The
122.
find
u=2t, and let r be what*
8-r-\/(hc)=2T, or d=2r^/(hc), and
Q=
in
will
s-f-A,
for
lie
mean
between
that which in
is
(96) was denoted by 0, and of which the table gives the

values corresponding to the different values of
neral result of the investigation
is,
r.
therefore, that
The ge
whatever
be the nature of the function $ nx which represents the law

of the
facility
of the different values of A,
if
a large
num
ber of observations be made, the sum of the values of A,

divided by the
number of observations, approaches continu

(which is the true mean
ally to a certain special quantity k
value of
that
A)
as the
number of observations
is
increased,
and
by multiplying the number of observations, a probabi
may always be obtained, approaching as nearly to cer

tainty as we please, that the difference between the arith
metical mean or average of the observations and the true
mean value of A, will be comprised within limits which may
be made as small as we please.
The analysis employed in the preceding articles, (113
lity
to 121), for the purpose of establishing this very
tant result, belongs to Poisson,
same form
in the
and
is
given
in
impor
nearly the
Recherches sur la Probabilite des Juge-
AXD LIMITS OF PROBABLE ERROR.

ments, chap,
and
iv.,
in the
We
des Terns for 1832.
177
Additions to the Connaissance
have preferred
it
the method
to
followed by Laplace in the Theorie Analytique, as being
somewhat simpler and

1
also
more
general.
23. In order that the limits 2r\/(hc)
may be
real, it is
necessary that the special quantity c be positive, a condition
which has hitherto been assumed.
Now,
since
c=Zcn
/<,
=(k
obvious that c will be positive if c n

k 2 n ) be po
n
sitive. On writing for k n and k n their values
(119) we have
it is
x=a
the limits of the integrals being always from
But
evident that no change will be
it is
made
to
x=b.
in the values
of these definite integrals (the limits continuing the same),

by substituting in them any other of the possible values of
We have therefore jx ^ x dx ==fx(f) xdx, and since

dx =l, the above equation may be other
cases J
f
A, as x
in all
(f>*x
wise written
whence 2cn =flfy nx<p nx (x 2

or
2c n =ff<p n xcf) nx (x 2
Adding together the two
a quantity which
zero so long as
124.
The
values of
is
x x)dxdx f
last equations,
necessarily positive,
x can have
there results
and can never be
different values.
special quantity k to
which the average of the

is connected with the
continually approaches,
centre of gravity of the area of a curve
by the following
rela
Let x and y be the co-ordinates of a curve, of which the

then the element of the area is n xrJx.
equation is y=<p nx
tion.
But (116)
value of
jr
-f-
dx
<p n
xdx
is
<$>
the infinitely small probability that the
in the nth observation will lie
between x and
therefore the element of the area of the curve
178
represents this probability, and the curve itself represents

the law of the probability of the different values of
spect of the nth
trial.
ordinates are x and

probability of
in re
In like manner, the curve whose co
( 1 -r-^)2ip n .r,
represents the law of the mean
in respect of the
whole
series of observa
be the absciss of the centre of gravity of

curve
whose
co-ordinates
are x and y, the well known
any
Now,
tions.
if xi
formula of mechanics gives x^fyxdx-i-fydx; therefore,

applying this formula to the curve of the mean probability,
and making the whole area (fydx from

the absciss of the centre of gravity
But
cial
x=a
to
#=&)== 1,
Xi=(\-^-h)^fx^ n xdx.
denoted by k (1 19} hence the spe
quantity to which the average of a large number of ob
is
this is-the quantity
servations indefinitely approaches
is
the absciss of the cen
tre of gravity of the area of the curve
which represents the
law of the mean chances of A.

125. It has been assumed in the foregoing analysis that
is
susceptible of an infinite
continuously from a to
b.
number of values,
The
results,
increasing
however, are easily
adapted to those cases in which the number of possible va

A is finite. Suppose A to be a thing susceptible of
lues of
only X different values, represented by a v
and
let
the chances of these values, which
2,
as
may be
# /?
different
be respectively y lf y 2 y.
yx
in respect of the wth trial.
Now, suppose q> nx to be a dis
continuous function, which vanishes for all values of x, of
in the different trials,
which the difference from one or other of the above values

of
exceeds an
infinitely small quantity
series of
<z<=b:c,
X partial
the
x=a
to
x=b,
sum of which will be
the values of
then the whole
be made up of a
xdx
taken
between
the limits
integrals Jty n
integral J$ xdx from

n
will
unity, since
one or other of
must necessarily be given by the
trial.
But

the integral f<p n xdx between the limits r/,:re
Now
between a.rtze
the difference x
cannot exceed
r/ t
=Z,so
ya
t
for those \im\ts f<p n

i
a must be
t
we may
trial
xdx=y
infinitely small, since it
therefore substitute a for x,

f
and
x 2 under the sign of integration, when the limits are
for
a?
whence
the expres
is
chance that the value of A given in the nth
sion of the
will lie
On
we ha.vej $q>, xdx=ia
that for those limits
writing for i all the different
and observing that the X
numbers
J<p H Jcdx==
1, 2,
3 ...... X,
formed make
partial integrals thus
xa
to x=b, and that

up the whole integral Jx<p n xdx from
sum is k H we have, in respect of the nth trial,
therefore their
2
In like manner, for k n -=zfx
2
xdx (from a
=7ii +yBfl +y
so that the
two
a
3
......
-f
to
we have
>),
/A.
2
;
f
special quantities k and k become
the sums 2 extending to

trials,
all
the h values of n, or to
the chances denoted by y 15 y
e.
all
the
being supposed
to vary in the different trials.
126.
When
the chances of the different values of
are
equal and constant, then y<= 1 ^-X, and the above values of h
and k become
so that k
is
the arithmetical
mean of the
possible values cf
f
A, and k the mean of the squares of those values. On this
f
hypothesis, therefore, k and k may be computed a priori,
and consequently the

is
a given probability
limits
determined within which there
that the average of h observations
will fall, thelimits

being /fcnp2r v /(c-r- A,) where
c=J(
).
180
When
the chances of the different values of
equal, but constant in the different

k -=.k n and we have
trials,
then
are
A=# n
un
and
In this case the special quantity k to which the average of the

observed values continually approaches, is the sum of the
possible values, each multiplied into
bility
and k
the
is
its
respective proba
sum of the products of
the squares of
those values into their respective probabilities.

127.
Resuming the
we
in (121),
when
the
consideration of the general formula
now
give an example of its application

which represents the law of facility
values of A is supposed to be known a
shall
function
of the different
priori.
Of all
law of
may be made
the hypotheses which
the simplest
facility,
chances of
all
respecting the
that which supposes the
is
the possible values of the thing observed to
be equal, and to remain constant during the series of trials.

This supposes <>#= <>#=: a constant; whence ftpxdx, be
tween the
limits
between those
<px= l-t-(b
x=a
and #=&, becomes (b
a).
the special quantities k and k

kzs.k n
x=a
and k
But
a) Qx.
we have also f$xdx-=. 1 therefore

From this value of <px it is easy to deduce
limits
the present hypothesis
therefore, the limits of the integral being
and x=zb, we have
A= IxQxdxiz If*xdx =
a), whence k^=^(b -j- a)
On
2
.
8-
a)
2(b
Inlike manner k =fx*qixdx
/xdxa=.i(6 -|-a4-a
j
b
a2
b2
whence c=\(k
Hence b
121
k~)
we have
181
given by h
the probability
that the average value of
sum of the
observations, or the
number,
will lie
values of
divided by their
between
128. This formula
Of the
tion.
may be applied to the following ques

comets which have been observed since the
year 240 of our era, the parabolic elements of
38 have
been computed, and the mean inclination of their orbits to the

Now, supposing every possi
ecliptic is found to be 48 55
.
ble inclination of an orbit to be equally probable, let the

probability be
demanded
orbits will not differ
clinations)
more than 5
mean
that the
from 45
(the
inclination of
mean
138
of the possible in
in excess or defect.
In this case the limits of the possible values of the phe
nomenon
are
We have
and 90.
=0, =90,
therefore
/j=138, and the above limits of the error of the average,

In order that
become 45 rpr x 90-7--v/(6 X 138).
the limits
the
may
equation
not exceed 5,
T
90
-r-
we have
V(6 X
138)
from
to determine T
=5,
which gives
whence T=l-6 very nearly. The tabular value

corresponding to r=l-6 is 97635, or nearly fj and
T=^V23
of
the odds are therefore 41 to
that on the supposition of all
inclinations being equally probable, the
138 comets would
and 50.
40
puted
falls
fall
mean
between 45r^r:o
The mean
inclination of
that
is,
within those limits (being 48
45
therefore a very great probability that whatever
the nature of the
between
of the inclinations actually
com
there
is
may be
unknown causes which determine the
positions of the cometary orbits,
it is
not such as to render
different inclinations unequally probable.
If the question
it is
had been
as probable that the
to assign the limits within
mean of
which
the inclinations will
fall
182
as not,
we
should have had
0^,
-476936, and the
the table) T
and consequently (from

would have been
limits
45:zp90 X-476936-rV(6x 138), which is found on cal

On the supposition, therefore,
culation to be 45;qpl -5.
that
inclinations are equally probable,
all
mean of 138
that the
and 46J, or
On
129.
inclinations will
at least not
fall
exceed those
is
it
limits.
the same hypothesis of an equal probability of
all
possible values, if
0,
we have
we suppose the mean value of A to be

b, and $x becomes l-r-2, whence
then
(127) become
the limits corresponding to a given value of
0=+22 T b
Let 0=J, whence T=-476936, and
V(6h).
With
suppose 7i=600.
Or+z 0166 nearly
that
is
these values the limits

to say,
it is
the average of 600 observations will

true
mean
of a or
b,
130.
value of
what
A to
creases from
<p#:=/3
to
A more than the sixteen-thousandth
x)
hypothesis, suppose the chance of a
then
<px will
in
be found as follows
=0 we
have then by the hypothesis

whence <px=i(a #)/3-r-, and conse;
a,
cpientlyftpxdz^px
becomes J/3.
part
decrease uniformly as the magnitude
zt a
when
<px:$(n
become
an even wager that

not differ from the
the greatest possible difference.
is
As a second
given value of
Let
one to one
between 43^ and
/3# -7-2#,
#=0 to x
which, from
x=
But f$xdx from
to
x=-{-a
-f-#>
is 1
limits being supposed impossible),

(errors beyond those
therefore from
-J/Sa
1 -J-tf.
l, or/3
value of c
x=Q to x=. + a,f<xdx=^, and consequently
is
Hence $x(a
#)-7-
from which the
easily deduced, that of k being
0, as in
the
former case.
131. Although the function Qx which represents the law
of facility of the different values of

its
form
may be assigned
if
A is in general
we assume
that
it
unknown,
is
subject

certain
to
from the nature of the
which,
conditions,
183
must be very nearly, if not absolutely true, in most

1st, That the chance of an error dimi
practical cases:
thing,
nishes as the magnitude of the error increases, and for er
beyond a certain limit vanishes altogether ; and, 2d,

that positive and negative errors, of equal magnitude, are
rors
The
equally probable.
last
condition
equivalent to the
is
assumption that the average of the observed values

true
mean
For
value.
simplification,
we suppose
is
the
the chance
of an error of a given magnitude to remain constant in
all
the
trials.
of values of A, the sum

and make m=s~-h, then
the arithmetical mean or average, which by hypothesis
132. Let x,
of which
m is
is
be a
x", &c.
series
and the number
is s,
h,
phenomenon A.
w=A", &c., so that A, A
m=&
errors of
ror
is
x"
x x
<p(x
A", &c. are the
the most probable single er

and the probability of obtaining an error of a given
}
magnitude A
that
m = A,
Let x
the true value of the
x", &c.
in
Now,
any observation
is
of obtaining a given value
m)= <pA
so that <pA
is
obviously the same as
of
therefore <px=
the probability of a single error
being exactly A. In like manner, the probability of an error

ror rzA is <pA7 ; and if we take P to denote the probability
of a given system of errors, A,
A", &c., then the errors
being supposed independent of each other,

<pA", &c.
<pA
P=pA
Let
this
(7)
system be assumed to be the most probable result
of the observations, then

tial
we have
co-efficient zero.
is
maximum, and
its
differen
Taking the logarithms of both sides
of the equation, differentiating, and making d log. (pArr^
and c?P-f-c?A=0, we obtain
84
an equation which
This
is
tem of
may be
otherwise written
the conditional equation of the most probable sys

But the hypothesis of the average being
errors.
the true value, furnishes this other equation,
w) + (a?
0=(ar
or,
which
is
the same,
+ (ar"
i)
&)
-{-A"-f-, &c.
0=A-{-A
&c.
and on com
with the above conditional equation, it is evi

paring
dent that they can only be both true simultaneously on the
this
<Z>
supposition of
that
is
A--A
is
A
A7
=,
&c.
Hence
it
follows
independent of any particular value of A, or

which we shall call K. We have then
equal to a constant,
A
The
A _d.log. 0A
=K
AdA
integral of this expression
which, making the last constant

bers, gives
<pA=He*
A2
we suppose
A
Assume
Â=He
H and
yA*.
increases,
p or
t ]ie
s|KA
log. <pA
log.
-{-
now
only remains to deter
K.
With
it is
y,
respect to K, as
to
be
obvious that
0,
-fa
and that
K must
be
and the formula becomes
determination of
H we have
equation y*<pAe?A= 1, the limits of the integral being
of x.
const.,
H, and passing to num
the most probable value of
<pA diminishes as
negative.
It
mine the two constants
is
the
a f and
where a =^(b a), a and b being the limiting values

But it is to be observed, that as all values of A ex
zi a are supposed to be impossible, or

be so improbable that it is unnecessary to take
account of them, the value of the integral /<pAc/A from
ceeding the limits
at least to
A=
limits
to
from
A=-f-fl will not be altered by extending the

We have therefore
infinity to -f- infinity.
pA=He?
A=-rVy,
Let
l.
? A2 =He-* 2
we have/e-
on substituting which
^rr-v/Tr (96),
Whence,
H=V(y-s-ir).
1
The
33.
may be
Let
AB
is
be
make MB^=a
was shewn
finally,
and
in the last
oo to
+00
(H-sVy) -/"=!, and

Al1
pA=V(y-Mr>T-*
.
line.
now found
Let
aNb be
the ordinate corresponding to the
its axis,
MN
and
its
MP=A,
and
greatest ordi
M, draw PQ, an
andpq indefinitely
MA=_a
find
Suppose the origin to be placed at
ordinate at any point P,
as
by means of a curve
a curve of which <pA
nate.
we
general properties of the function
illustrated
absciss A.
d=dt+*/7
then
equation, and observing that from
185
near to
PQ, and
PQ=?A
then
element of the area, repre

sents the chance of an error lying between A and A-j-e?A, that
but less than M;;. Now, if <p A
is,ofan error greater than
in (124), PQ<7/>, the
MP
y ^ tne function
2
VCv-r"
6
")
71
will not
A therefore <pA =#(

ing A into
symmetrical on both sides of MN, as
;
be changed by chang
A), and the curve is
it
obviously ought to be
making MP =MP, then

of
errors
and
positive
negative
equal magnitude being equally
2
probable, we must have P Cy^PQ.
Again, since e *^
according to the hypothesis
diminishes rapidly as
for on
increases, the curve at a short dis
MN must approach
very near to its axis AB ;

but as the function only vanishes when A is infinite, the curve
will not meet the axis at any finite distance from MN.
This
tance from
curve, therefore, can only represent approximately the law of
186
facility,
inasmuch as
it is
tain limit are impossible
supposed that errors beyond a cer

but on account of the rapid dimi
nution of the ordinate at a short distance from
MN, the
MB,
chance of an error exceeding a small value of A, as
becomes
Hence
insensible.
the limits of the integrals in
respect of A
may be extended without
values from
Azra
134. It
is
now
to
A=r=op
sensibly altering their
necessary to find the special quantities
k,
and observing that as the

Substituting
law ofthe chances is here supposed to remain constant, we have
k andc.
,
k =k n
2
=fA <pd.
k=k n
k
for x,
become
the formulae in (119)
k=f<p Ac/A,
* A2 ,
Hence on making $=-*/(y-+--n)e
we
have
When A becomes infinite, this

A
oo to A= -J- GO, k=0.
becomes
This
is
quence of the symmetry of the curve,

necessarily in the straight line
is
vity
With respect
k f we
to
an obvious conse
for the centre of gra
MN.
may proceed
7/rr/A (/)AÂ=:-v/(y-r-7r)/ A
2
therefore from
0,
A2 fZA.
e--y
thus.
We
have
But from the
principles of the differential calculus,
Ae~ y^=e
i
therefore, integrating
Y An~dA
1y
* A VA,
and transposing,
Now, from A
tion
which
is
ootoA=
ey^-d\=\/(TT-^-y) (from
7A
2
),
00,
the term of this equa
not under the sign of integration vanishes, and
therefore/A
consequently k
=l
*?
* A V.A
1y.
(96), on substituting
(1
-f-
^y)^/(^
-f-
for
7); and

2
we assumed c=^(k
In (119)
c=k
k )
187
therefore in the pre
whence c=l-f-4y, or y=l-r-4c.

135. The expressions which have now been found
sent case
for
the function which represents the probability of an error,

and the limits corresponding to an assigned degree of pro
bability, are given in
terms of the indeterminate constant y
(or c), which depends on the nature of the observation, and

therefore, where instruments are requisite, on the goodness
of the instrument and the skill of the observer.
This con
stant
It
is
called
by Laplace the modulus of the law of facility.

; but if we assume
cannot, in general, be assigned a priori
that positive
and negative departures from the mean are

which is the most plausible hypothesis the
alike probable,
nature of the thing admits

in respect
of,
an approximation to
of observations of a given kind,
its
value,
maybe deduced
with great probability from the results of a large series of

observations of the
same kind already made.
ceed to give the analysis by which this
lowing the
is
We now pro
accomplished,
method ofPoisson. The approximation
is
fol
carried
only to quantities of the order l-^-y/A; terms having A for

a divisor are neglected on account of their smallness, h
being supposed a large number.

136. In the expression for Q
and consequently
in
(119)> suppose ^-=5,
5=25, and write

the equation then becomes
2
50
f
e
- r
^ 5=0,
\p--f-
also
z for
e~
cos (hkz
sin
and
Q is the probability that
given by
all
If therefore,
bz) sin bz
(hkzSz) sin
s,
the
the observations, will
we
sum of the
lie
values of
between
and
23.
of
suppose 5 to be variable, the differential
188
taken with respect to S, will express the

sum of the values being 25
this expression
infinitely small chance of the
exactly.
v) sin v
cos (u
v) cos
-f-
v=* cos (2v
v) cos
sin (u
v=
sin
u),
(2v
u),
shall find
r* _
2dd
Let
d& be denoted by qdt,
these values,
_/>*
the corresponding value of
let
we
on substituting
shall have,
and replacing z by B-^-V(Jic), qdt

27fe
cos (2t6)d6
/*
/o
__ S 2
e
sin
TrC*/(tlC)J
The two
formula
integrals in this equation are found from the
in (121).
?
and
zdB
be a variable quantity, and assume 28=hk+2t^(hc),
whence <2&=<#V(Ae), and

ao
u and v
if
arcs, the trigonometrical formulae give
v) sin v-\- cos (u
sin (u
we
and observing that
Differentiating,
denote any two
if this
last
Writing 2t for
w, that
formula gives
cos
equation be differentiated in respect of
/,
three times in succession, the result will be
whence,
if
we make
V=
TT-\ (%t
^
2t z
\ we
shall
have
2c^/(hc)
qdt=
(1 -=-v/7r)
(lV)<r*dt,
a quantity containing only uneven powers of (,

and of the order l-^-yV*, so that when multiplied by another
where
is
of the same order, the product will be of the order
1-f-A,

and
will therefore
189
be rejected in the present approxima
This value of qdt is the probability that s will be pre

2t^/(hc) or it is the infinitely small pro
cisely 28 or hk
of
the equation
bability
tion.
137. In order to apply this result to the determination
of the probable limits in terms of observations actually

made, it is necessary to remark that the analysis by means
of which
has been obtained
grounded on the very

general supposition that the thing to be measured may be
any function whatever of the quantity observed for the in
it
is
chance of a particular value of the function
finitely small
is
evidently the same as that of the corresponding value of
the quantity, and
is
consequently
be a function of x, and
when
is
let
<$> n
xdx.
Let
K, C, T, be what k,
X therefore
c, t
become
substituted for x, the above equation then be
comes
the symbol 2 including
all
bability of this equation
is
as that
which
is
the h values of
and the pro
an expression of the same form
represented by qdt.
made with respect

introduce the hypothesis that positive and
negative departures from the mean of equal magnitude are
138. Hitherto no restriction has been
to <ja
we now
equally probable, and consequently that the curVe repre
senting the law of facility
is symmetrical, but shall

sup
pose the chances of a particular value, or a particular error,
to vary in the different trials.
Let the origin be transferred

and let x k
to the centre of gravity, the absciss of which =.k,
k= A c. We have then by (132) <#=$ A,

2
2
A d A the integrafx(f>xdx=f&. A d A fx $xdx=f&
A, x
</)
tion in respect of
being from
</>
oo to -f- oo.
The special
100
quantities k
= 0,
and k then become 7e=(l-j-/i)syA0 n Ac? A
A =(l-i-/Os/A 2
The
object
now
is
<
A</A, whence
to eliminate
and determine
n A,
<
c in
terms of the observations.
=# be the observed value of A in the wth ob

=A is the true error of the observa
then \ n
139. Let A M
servation,
tion.
=A
2
,
Let the function denoted by X in (137) be (X n

and the corresponding value of K (since in this
K=(l-5-A) sfXfaxdx) becomes K=(l -f-A)

this
Comparing
in the equation (137),
and assuming
2
T and C when X=(A n

=2/ic-f 2*V(^0 whence
values of
k)
c=(l-h2A)2(X B
a quantity of the order
(U being
lity
of this equation
where
t
is
and
we
-f-\/^)
cf
case,
n A6?A.
<
we have
therefore on substituting these values of
K=2c;
2/A
with the value of c found above,
k)
and
be the
to
get 2 (A n
&)
(1)
an d the probabi
is
a function containing only uneven powers of
and of the order 1-t-Vh.
X=#=A n
In the equation (137) suppose

c" be the corresponding values of
K=k,
and
let
/"and
and C, then since on
the equation becomes 2X n
this
supposition
(U"
of this
being of the order l-t-^7*); and the probability
/"U",
equation
where
i"
and
140.
V
is
(2)
is
r/
,
like
and V, contains only uneven powers of
of the order l-s-\/h.
The two
equations (1) and (2)
may be
regarded

as
two
now
having the respective probabilities

and therefore the probability of their
distinct events,
assigned to them,
being true simultaneously
and
probabilities,
V V"
which
is
191
the product of their respective
is
accordingly (neglecting the product

a quantity divided by A),
is
^W.
V">-<*<r-<"
q q"dt dt"=(\+Tr)(\
Let the value of k given by equation (2) be substituted
in (1), and the expression now given will accordingly be
the probability of the resulting equation, namely,
c=
i 2(X "~T
Let wz=(l-f-A)2X n then

,
U ) 2 -<"U".
.+<
is
the average or arithmetical
mean of the observed values, and X n

of the observation. The last equation
(l^-27i) 2 (X
which
is
m+t U y
fJ
of the order
c=(l
is
the
then become c=:
or,
rejecting
<VU
1 -4-^,
-s-2A):B {(A n
*=(1^)Z(X B
/i
the reputed error
*"U";
m)
For the sake of abridging

so that
m
will
+ 2(A
let
mXU }
t"\J".
us also assume
m)\ =(! -4-A)i(A n _mX,
mean of the
squares of the errors, or
mean
square of the errors, and the equation becomes
c-\^v\j
the probability of which
or
is
t"U",
(3)
q q dt dt".
141. Now, by (121), we have the probability

that s-^-h
SA u -i-A z=.m the arithmetical mean of the observed va
lues of
A,
and
k-f=2rV(c^-h). Sub
above value of c, and observing
stituting in those limits the
that
(^+*U --*"U )^= VCW + NÛ *"U")+ &c.,

that U and U being of the order \~ ^/h when di
/
//
7/
vided again by
Jh
are to be rejected, the limits
^2rV(i/^),
or
become
fc+rrV^-f-A),
and the probability of these being the true
limits
is
mul-
192
tiplied into the probability of the equation
*"U"
and
is
(1
The
142.
el/A-fj/U
therefore (140)
+7r)(l
expression
now obtained
\")<r**<r**dedt
is
the infinitely small
probability of the limits Azfzr x/(2/i-4-^) of the average m,

in respect of the particular value of
which we have de
for
s,
duced the equation (3). But for every value of s between

the limits
and 2S, there will be an equation correspond
have the whole proba
bility of those limits, the integral of the expression must
be found for all values of t and t". From the nature of
ing to (3)
therefore, in order to
the expressions
e~~*"*
and e~ trri
as well as the consideration
beyond a certain magnitude, though possible,

are wholly improbable, it is evident that the integration may
that errors
be extended without sensible error from

since the functions
of
and
t",
and
V"
oo to -J- oo
and
contain only uneven powers
the terms into which they enter, disappear in
the integrations between those limits. (SeeLacroix, Calcul

Diff. et Integral, torn.
to
iii.
p. 506).
= + oo we have (96)/e~
W= V
Now, from
r
and/e~
=.
oo
W= V *
therefore
The
result of the preceding analysis
is
therefore that on
the hypothesis of positive and negative errors of equal
mag
nitude being equally probable, and on rejecting terms di
vided by h (the number of the observations

so great as to render such terms
may be always
insensible), we may sub
stitute \\i for c in the limits of the error to
be apprehended,
without sensibly altering the probability, and consequently

there
is
the probability
0=
Je
^dt that the true

mean value k of the phenomenon
193
A will lie between the limits
or
.^-
which contain only quantities given by observation.

On this hypothesis we have also (138) c=(l_ 2h)
2
2/A n Ae?A, or, supposing the law of facility to remain con
<
2
$Ae?A, therefore /*==/A
$Ac?A; that is to say, the mean of the squares of the actual
errors may be taken for the sum of the products of the
stant during the trials,
c=4/A
squares of the possible errors multiplied by their respective

It is
probabilities.
important to remark that as the obser
become more numerous, the quantity p, the mean

of the squares of the errors, converges more and more to a
vations
constant quantity, and finally becomes independent of the
number of observations.
1
By
The
43.
limits
now found may be
m=i (l-r-A)SA =the
hypothesis,
ra
the observed values, and p.=(
otherwise expressed.
mean of
arithmetical
-4-^)2(X n
of the squares of the reputed errors.
m)
Now
the
mean
2
m)
2
XB
2X,,m-j-w and (l-^h)^2\ nm=2m(l-^-k)^\ n =:2m ;
2
2
w that is to say, the mean of
therefore /m=(l-7-A)SX n
(X n
the squares of the observations minus the square of the
Hence
mean.
bility
the limits, corresponding to a given proba
0, of the difference between the average of
observations and the true value, are expressed
by
all
the
either of
these formulae
rzr^/ {(2 X mean square of errors-4-Aj,

2
square of obs.
(mean of obs.) }-f-V^
h being the number of observations, and the relation be
dtrViSx mean
tween
ing, the
and
first
T being given
by the
of these formulas
is
table.
Generally speak
the most convenient for
calculation.
194
144. Let
be the
limit of the error to
be feared
in tak
ing the average of the observations as the true result, then
Now when
l=Tj(2n -t-h), and r=/ x/(A-j-2 /i.)

stant, that
for a given probability 0, the
is,
is
con
determination
be more exact in proportion as / is a smaller number,

and the precision will therefore be proportional to *J (h-^-2p.).
Hence ^/(h-i-2p) is called by Gauss the measure of the
precision of the determination.
Suppose two series of ob
will
servations to have been
made for the determination of an
ment, the comparative accuracy of the results
two
things, the
will
ele
depend on
number of observations in each series, and the
amount of the squares of the errors in each. If the num

is the same in both series, the
precision
ber of observations
of each result will be inversely as the square root of the

squares of the errors, and the presumption of
sum of the
accuracy
is
in favour of that result with respect to
the
sum of the
On
the other hand,
same
both
squares of the errors

if
the
mean
is
less
which
than in the other.
square of the errors
is
the
then the observations are alike good

in both, and their relative values of the two results are di
in
series,
number of observations
rectly as the square roots of the
each
series.
Hence,
twice as good as another,

the
in
one determination may be

must be founded on four times
in order that
it
number of equally good
observations.
tions are very important, in
These considera
comparing tables of mean va
lues of whatever kind, for example, of the probabilities of

life at
the different ages, and in estimating risks which de
pend upon them.

145. Astronomers employ the terms, weight, probable
error, and mean error, of a result, to denote certain func
tions of
f*,
the
mean square of
the errors.
The
square of
the quantity which measures the precision of the result,
is
AND
LIMITS OF PROBABLE ERROR.
called thetveight of the determination.
we have
by w,
or the weight
195
Denoting the weight
therefore
is
equal to the square of the
servations divided by twice the
sum of
number of ob
the squares of the
Substituting this in the expression of the limits,
errors.
we have l^r^-Jw, and T=l*/w
that
is
to say, for a given
probability 0, the limits of the error to be apprehended in
taking the average as the true result are reciprocally pro

portional to the square root of the weight.
When
obser
vations of different kinds, or results deduced from observa

tion, are
compared with each
other, their relative weights
(supposing the number of observations the same) are inversely

as p,
and are expressed numerically by taking the weight
of a certain series of observations as the unit of weight.

146.
The probable
error of the determination
is
that
For 0= J we
which corresponds to the probability
5.
have r=-476936; whence rV2=-674489, and the formula
mdb3r x/(2/i-f-/) becomes

probable error
147.
The mean
mzi 674489*J(n-i-h)
whence
=-674489 V (/*-*- h }-
error of the result of a large
number
may be deduced from the general formula

That formula gives qdt(\-^- JTT)
in (136) as follows.
Vê-^dt for the probability that the sum of the ob
(1
Di
served values will be 2S
2t*/(hc) exactly.
of observations
=M+
viding the
sum by
h,
qdt
J(c-t-h ).
/>
also the probability that the
all
the observations will be exactly
Now, on
the hypothesis that positive and
average value given by
k+2t
is
negative departures from the
mean
are equally probable,
be trans
and supposing the origin of the co-ordinates
to
ferred to the centre of gravity of the curve of
mean proba-
196
we have
bility,
=0, and qdt=(l-s-Jvr)(l
V)*r-< <&
the
is
chance of the average error being 2 A/(c -4- ft)

Multiplying therefore this error into the chance
infinitely small
exactly.
of
its
of
all
t=0
taking place, and integrating the product from
t=.
to
we
oo,
shall
have the mean
mean
error, or
risk
the possible average errors affected with the positive
Now, observing
sign.
V represents a quantity
that
divid
ed by fjh, and therefore when multiplied by 2 A/(c-j-A)

becomes of the order 1 -f-^, and may consequently be re
jected, the product of the average error 2.t>J(c-+-h) into
probability
^fd.e-
is
2*J(c-i-7rh)
t2
ê-**,
Xte
which from
*dt;
tfrrO to
its
and since fte-^dtzz

oo becomes
simply^,
t=
the integral of the above product from
t=0
Substituting for cits value
to
=oo
is
=J/i, this re-
^(c-s-nh).
(142)
sultbecomes ^/(p^-2?rh) whence on computing ^/(l -r-27r)
;
we
obtain
mean
This
error of series =.398942^(^1-:- fi).
mean
the
is
tive errors alone, or
error or
errors are equally likely, the

is
But
as positive
mean
the same quantity,
respect of errors of both kinds

is
risk in respect of posi
on the supposition that negative errors
are not taken into account.
tive errors
mean
whence the mean
is
nega
error in
.797884 J^^fi).
usually called the average error.
from the probable error
and negative
error in respect of
This
The mean error differs
in this respect, that
it
depends on the
magnitude of individual errors, as well as on the proportion

in which errors of different magnitudes occur.
The proba
ble error
148.
is
independent of the magnitude.
When
rors) has
the quantity p (the mean square of the er

been found from a series of observations, the
precision, weight, probable error,
and mean
error, of a
com-
197
ing observation of the same kind are found by supposing

in the above expressions, and are respectively
k~\
precision
weight
nl-r-2/z
probable error
=.674489 V/*
mean
=.39 8942 Jp.
error
The preceding
149-
formula? give the limits of the error
be feared in determining the value of a quantity from a

series of observations, when the thing to be determined is
to
which the observations are immediately made. We

have now to apply the formulae to the cases in which the
that on
is
quantity sought
several others,
not observed
itself,
but
a function of
is
which are separately determined by obser
The following problem is important

Let u be a given function of a number of unknown quan-
vation.
titities, x,
x", &c.
it is
required to assign the limits of
determination of U, and the
the probable error in the
weight of the
result,
when
values of
or,
x", found from
observations independent of each other, and respectively af

fected with the probable errors
Jp.,
gvV>
*/ /-""?
& c< ($ ==
.674489) are adopted instead of the true but unknown

values of those quantities.
Let u=f(x, x x", &c.) be the given function, X, X X",

x x", &c. and make X x=e,
,
&c. observed values of #,
=e
X"
x"=e", &c.
so that
errors of observation, supposed to
squares
maybe rejected. Make
&c., then
ing x-\*e,
equation
du
x" Jê"
u=f(x, x
for
a?,
x"> &c.),
a?
e", &c. are the
du
=a,
a", are given quantities

l
-\-e
e,
be so small that
-7-,
du
,
their
7-^7
and on substitut
x", respectively, in the
and supposing u
to
become
198
when
the substitutions are made, so that
we
corresponding error of u,
Taylor
is
the
have, on expanding u by
theorem,
a e"-f &c.
E=ae+aV-f
in respect of a single observation of each of the quantities.
Taking the square of both
Now since
positive
and negative
errors are supposed equal
ly probable, the sums of the products ee
their
mean
values,
we have
sides of the equation,
become each
SE 2 =a 2 Ee 2 + a /2 2e/2 H-a
:Se
ee", e
e",
&c. or
therefore
/2
&c.
Taking the mean value of each of these sums, and observing

the mean value of Se 2 is independent of the num
p.
that
ber of observations (142), and assuming

value of
sE we
,
get
M=aV+a Y-i-"V +
all
&c.
This equation contains the solution of the problem, for

the functions of the error are given in terms of M.
The
probable error
150. Let
w,
M to be the mean
itf,
is
-674489 V
be the weight of the determination, and
w", &c. the weights corresponding
to
p.,
p.
//
&c.
then by the definition of weight, w is reciprocally propor

tioned to p., and
to
and we have by substitution,
;
W M
If the weights are supposed
W-~ a
Suppose the errors
plied
e,
f2
+a + a"
is
equal, this
becomes
+8tc.
be respectively multi
ef e", &c. to
by numbers proportional
weights, (which
all
to the square roots of the
equivalent to supposing
all
the observa-
AND LIMITS
tions to
199
have the same degree of precision measured by
*J(nio)),
then the value of
But w being reciprocally as
=1,
PROBABLE ERROR.
OF
M becomes
/a,
we have pw=p u/=p w",
therefore
w= ..
*2
*
I
s*
"
rt
&c.
OF THE
200
METHOD
SECTION
X.
OF THE METHOD OF LEAST SQUARES.

151. In the determination of astronomical and physical
elements from the data of observation, the thing which is

actually observed is for the most x part not the element
which
is
sought to be determined, but a known function of

be a given function of X deter
Thus, if
that element.
mined by the equation V=F(X), the quantity observed
may be
a value of V, whilst the element sought to be de
termined
V with
known
is
X.
If the observation could give the value of
absolute accuracy, then
but as
all
X would also be
absolutely
observations are affected with certain
errors of greater or less amount,
owing
to the imperfections
of instruments or of sense, or the ever varying circum

stances under which they are made, an exact value of
cannot be found from any single observation ; and in order

to obtain the utmost precision, it is necessary to employ a
great
number of observations, repeated under every variety
of circumstance by which the result can be supposed to be

affected.
152.
The observed
quantity V, instead of being a func
tion of a single element X,
may be
a function of several
example,
may be the posi
tion of a planet, in which case it is a function of the six
elements of the orbit, for the determination of which the
elements X, Y, Z, &c.
for
OF LEAST SQUARES.
observation
is
201
made. Each observation gives rise to an equa
tion of this form,
V=F(X,
Y, Z, &c.)
therefore
when
the
number of equations is just equal to the number of un

known quantities, the problem is determinate and suppos
;
be an algebraic function, the values of X, Y, Z,

ing
&c. may be found by the ordinary methods of elimination.
If the number of equations is less than the number of un
to
known
much
problem is indeterminate ; but if

to be more than determinate, inas
quantities, the
may be said
it
greater,
as the
equations
number of ways, each
may be combined
in
an
infinite
combination giving a diffe

becomes a ques
tion of the utmost importance to the perfection of the
distinct
rent value of the elements.
It therefore
sciences of observation, to assign the particular combination
which gives the most advantageous
results, or values
of X,
Y, Z, &c. affected with the smallest probable errors.

153.
As approximate
values of the elements are in
all
cases either already known, or can be easily found, the cb

ject of accumulating observations
Let
approximate values.
observed,
V be
is
the correction of the
the true value of the thing
an approximate value, however found, X the

X an approximate value,
true value of the element sought,
V
V =F(X
corresponding to
V=F(X),
in
so that
)
we have
any observation be L, and
is
the true but
make
v=VL, l=V
then v
the two equations
the observed value of
also, let
unknown
L,
error of the observation,
reputed error, that is to say, the difference between

the computed value of the function and the result of the
and
/ its
observation.
Now
if
we assume x
to represent the true
correction of the approximate element, so that

then, on substituting
X +#
for
X^^
in the function F,
we
OF THE METHOD
202
get
V=F(X
-{-#)
whence, expanding the function by
theorem, and rejecting terms multiplied byx 2 and

higher powers of x, because x is a very small quantity
Taylor
T.TH.g.
Let us now denote the
known quantity, by a ;
equation becomes
of the observation
differential coefficient,
that
#=/-{-#;
is
which
V V =v
then, observing that
is
I,
the
to say, the true error
is
a linear function of the correction of
the element.
154. In like manner,
when
X, Y, Z, &c.; on making
there are several elements,
dV
dV
-^ =
^y
dV
>
~d%
&c. a single observation furnishes the equation
and a
v
v=l^.ax-\-bi/--cz+ &c.,
whose errors are respectively
series of observations,
&c. gives a system of linear equations equal in

number to the number of observations ; namely,
v=l -f-## by c2-f- &c.
v,
v",
+ +
&c.
(1)
&c.
&c.
and the object

errors v, v
f
,
is
to give such values to #, y> z, &c. that the
v", &c. in respect of the whole of the observa
be the least possible. The equations being

of each other, if their number is just
independent
supposed
to
that
of
unknown
the
quantities, the errors v, v v",
equal
tions,
shall
&c. can be
made
all
zero
but
as
if,
is
usually the case,
more equations than unknown quantities, it is

impossible by any means whatever to annihilate the whole
there are
of them, and therefore
all
that can
find the system of values of x,
y->
z,
be accomplished
is
to
&c. which most nearly,
203
OF LEAST SQUARES.
and with the greatest probability,
equations.
the whole of the
satisfies
If the observations are not
equally good, the
all
by a number
equations are supposed to be each multiplied
presumed weight of
proportional to the square root of the
the observation on which
it
depends, in order that they
may
have the same degree of precision.
all
155.
As
the question
to find the
is
most probable values
thing necessary is to express each

of these elements in terms of the observations.
Suppose
of x, y,
z,
&c. the
first
k y K, k", &c. to be a system of indeterminate quantities,

independent of x y, z9 &c. and let the first of the above
conditional equations be multiplied
the third by k", and so on

k,
the second by k
make
k", &c. be determined so as to
of a? equal to unit, and those ofy,

is
by
then adding the products,
z,
if
the coefficient
&c. each equal to
that
to say, so as to satisfy the equations
ka
+ k a + k"a"+ &c. =1
f
kb+ k b +k"b"
kc
+k +
c
k"c"
-f-
&c.
&c
=0
=0
(2)
&c.
we
shall
K is
x=K + kv+Ki/+k"i/ +
then have
a quantity independent of
found
=K,
with an error
v,
v", &c.
&c. where
Hence x
=Aw-J-^V+A V 4&c. and

/
weight of the determination, by the formula in (150),
The wei &ht
&c
is
the
of the Determination
is
is
2
f2
2
-{- &c.
consequently greater in proportion as k -\-k -\-k
of
inde
all
of
the
and
hence
is smaller ;
possible systems
k , k", &c. which satisfy the equa

system which gives the most probable value
or the most advantageous result, is that for which
terminate coefficients,
tions (1), the
of x>
is
an absolute minimum.
OF THE METHOD
204
156-
We
have now to
in terms of
find,
known
values of the indeterminate coefficients
ties,
which
satisfy the condition
A,
of the minimum.
quanti
k f , k", &c.
For the sake
of abridging, let us denote the aggregate of the products

a
a a"+ &c, by S(aa), that of ab+a b +a"b u -f
aa
+a +
&c. by S(ab), and so on, and also assume

4. &c.
&c.
v"+
(3)
=ct;-fcV -fcV -f&c.
On
substituting in these equations the values of
v,
vf t
v",
&c. given by the equations (1), there results
+ ;2:S(ac) + &c.
xS(ab) +y$(bb) + z$(bc) + &c.
<2&)
=S(c/)
+ x$(ac) +y$(bc)+zS(ccj+
(4)
&c.
a system of equations equal in number to the number of

elements x, y, z, &c. and from which, consequently, those
elements would be determined absolutely

tions
were perfectly exact, that
the observa
the errrors
if
is,
if
&c. were individually zero, and consequently
On
were each zero.
tem, the value of x

quantities
by a
is
rj,
linear equation of the following

i
/ g,
v, i/,
77,
&c. from the
given in terms of
ar=A+ /J+^ + Af+

where
z,
eliminating y,
v",
&c.,
last sys
and known
form
&c.
(5)
&c. are co-efficients independent of x, y>

&c. and also of ^, 77,
&c.
If
Sec.
h,
we now
substitute in equation (5) the values
and
also
a=:fa+gb+hc+
&c.
given by equations (3),
H-^ +^-t-
==/
a
we
shall
=/ a -f-^ //
/
f,
&c.
(6)
/i
have by addition
=A-f-av4-aV+aV
whence
77,
assume
+ V + &C.,
/
z,
it
appears that
a, a
a",
&c.
&c. are a system of multi-
OF LEAST SQUARES.
pliers
(1)
205
by which y> z, &c. are eliminated from equations

they must therefore satisfy the equations (2), whence
atf-faV-f-a V-f&C.
=1
&C.
=0
&C.
=0.
(7)
we
Subtracting these from the equations (2)
obtain
&c
&C.
&C.,
on multiplying which respectively by

the products,
0=( k
This equation
from which
we
f, g, k,
get by reason of the equations (6),
a)a
+ (&
X + (k"a")a + &C.
f
may be put under the form k
it is
and adding
evident that k 2
minimum when A=r a, k
=a
-f-
+ k 2 -\-k"
-\-
2
-f-
k" 2 -f~ & c
&c. will be a
Hence
k =a", &c.
it- fol
lows that the most probable value of x which can be de
duced from the equations (1),

weight of the determination
is
is 1 -4-
x=A
(aa -f
and by (150) the
aV + aV -f &c.)=
This quantity S (aa) is equal tof the co-efficient of | in

for on multiplying the first of equations
;
the equation (5)

(7),
by^
the second
the products,
we
by
aa-f aV
157.
g,
and the
third
by
h,
and adding
obtain by reason of equations (6),
The method
+ a"a"+
&C.
explained in the
f.
two
last paragraphs
of determining the most advantageous combination of a
system of linear equations, of the form of those in (154), is

given by Gauss in his Theoria Combinationis Observationum
erroribus minimis obnoxice, (Gottingen, 1823).
tical rule to
near value
which
it
leads
V of a function
is
as follows
The prac
Having given a
of several elements, X, Y, Z,
206
OF THE METHOD
&c. and also a series L, I/, L", &c. of observed values of
V, make
v",
L)Vw*=,
(V
(VL )Jw =v
(V L")^*
From these equa
&c. and form the equations in (1).
tions (4) are easily
deduced
and from
elimination, are found the values of
these, again,
by
y, z, &c. the correc
a?,
approximate elements X, Y, Z, &c., in equa

form (5), which, for the sake of symmetry, may
tions of the
tions of the
be thus written
C.
&C.
&C.
then the most probable values of #,y, 2, &c. are respectively

A, B, C, &c.; the weights of the determinations respectively
-
r-,
7^-rr, 7
(aa)
r-,
&c.
and the probable
errors of the se-
(yy)
(ffi)
veral determinations
are p^(oa),
p^(/3j3), p*/(yy), &c.,
where p=-476936.
158.
The
values of
immediately, by
a?,
y, 2, &c.
supposing the
errors of observation to be a
now deduced
sum of
are obtained
the squares of the
minimum.
Thus, forming
the squares of the equations (1), and making Qsrt^-f-t/

t/ s-j- &c.,
-fi
the differentiation ofii in respect of each of the
variables #, y, z, &c. produces the quantities denoted in
(156),
by
T),
& &c.
that
is
to say,
it
-= 2 - =2,,^ =2
do,
do.
|)
therefore if
Q be
gives
dQ,
a minimum,
?,
f,
& C.
^,
become
severally
zero, and the equations (4) give by elimination, x=A,

3/=B, z=C, where A, B, and C denote the same quantities
as above.
Now
from equation (5) the general value of x
and the most probable value being
#=A,
it
is
follows that
OF :LEAST SQUARES.
207
the most probable values of the corrections x, y,
by making the differential coefficients of

that
z,
are found
equal to zero,
/2
by making v -\-v +t/ -f- &c. an absolute mini

Hence this method of combining equations of con
is,
mum.
dition
is
called the
from the preceding
method of least squares ; and it follows

analysis, that it gives the most probable
values of the corrections, or the most advantageous results.
159.
As an example,
let us
known element X, of which

mate
value,
and L,
suppose there
is
known
to
is
only one un
be an approxi
L", &c. are observed values, the
L/,
weights of which are respectively proportional to w, #/, w",

&c. and that
it is
determination.
let
required to determine the most probable
X from the observations, and also the weight of the
value of
x be the
Make(X
L)-v/w=r#,
(X
X=X
correction of X, so that
L)^/M?=:/, and
x.
On
sub
x L)-v/^=v,
L) </w=.v, we have (X
XA/W. Each observation gives a similar equa
equations (1) in (154) consequently become
stituting this in (X
or
vljw
tion,
and the
&c.
by the coefficient of its own x, we

xS (w), whence consequently x=S(lw)-^-
therefore, multiplying each
have=S(&?)
S(0)
|-r-S(w),thatistosay,the most probable value of a?
is
&c.
&c.
and the weight of the determination

/
reciprocal
Since
of.w-f-/+/ -f.
we have
X L=#
I,
this proposition
proportional to the
also
hence
is
&c.
&c.
If a series of values of an element
OP THE METHOD
208
are found from observations which have not
all
the same
degree of precision, the most probable value of the element
found by multiplying each observation by a number pro

portional to its weight, and dividing the sum of the products
is
by the sum of the weights

the result
is
unit divided
If the weights be
vations
be
then
h,
all
and the comparative weight of
by the sum of all the weights.

and the number of the obser
equal,
X=(L-f I/-J-L"-}-
&c.)-r-A
that is to
say, the average of a series of equally good observations
The average may,
gives the most probable value.

fore
there
be considered as a particular case of the method of
least squares.
160.
To
illustrate the
method of proceeding when there
are several elements to be corrected from the observations,
we
shall take the following
numerical example from Gauss
Suppose there are three elements, and
(Theoria Motus).
that three observations, of equal weight, have given the
2y 5z=5, 4 x -\-y-\-4z~21
equations x y-}-2z=3, 3x
and that a fourth observation, of which the relative weight
is
one-fourth, or
has given
its
precision one-half, of that of the others
2#-{- 6^+6^=28.
this last equation to the
others, for
The
first
step
is
to
reduce
same standard of weight with the
which purpose
then becomes
it
must be multiplied by ^ it
Now, as x, y> and z can
;
x -{- 3y -f- 3z= 1 4.
not be determined so as to satisfy four independent equa

tions, we suppose each observation, or equation, to be affect
ed with an error
v,
and accordingly obtain the following
system of equations, corresponding to equations (1),
vf
5 -f 3x -f
v"=
t/"=
2y5
viz.
OP LEAST SQUARES.
209
from which the most probable values of x, y, and z are to

Let each equation be multiplied by the co
be deduced.
efficient
the
fourth
of
of
first
first
own
its
by
1,
1
by
taken with
,v,
the results added together give the value
88 -j- 27 x -\- 6y. In like manner, let the

namely,
be multiplied by
1, the second by 2, the third by 1,
3,
the
Lastly, let the equations

coefficients of z,
;
proper sign, namely,
==
and the fourth by
its
3, the third by 4, and the
the second by
we have then
we
77.
to
the following equations corresponding to
=88
17=
these
will give
and the sum of the products made equal
the equations (4)
From
sum of the products
be multiplied respectively by the
H- 27x
fy +
70 + 6* + 15y+z
get by elimination
737y =26 1
71 2 +
5477
397982=76242 + 12-4i7 + 1473&

whence (157) A, B, C, the most probable values of x,
z,
y,
are respectively
470 B=
A = 49154=2.^TAT*
19899
and the
2617
relative weights w,
19899
=-m9--
=3-ii
C=
551 r
,
ieP,
=
=M 6 W,737
=~54
,~
>
whence the probable
76242
==1 916
39798
are respectively
13 6 >-
39798
"147 3=
27
>
errors (-476936-=- Jiv) are respectively
096, -129, -092.
The method
my
is
indebted for
by Legendre,
nation
of least squares, to which modern astrono
les
much of its
in his Nouvelles
precision,
was
first
proposed
Methodes pour la Determi
Orbites des Cometes, (Paris, 1806,) merely as
OF THE METHOD
210
means of avoiding inconvenience and uncertainty arising

from the want of a uniform and determinate method of
combining numerous equations of condition, and without
The same method,
reference to the theory of probability.
however, had previously been discovered by Gauss, and a

demonstration of it, deduced from the general theory of
chances, was given by
It
may be shewn
him
in his
bination gives values of the
unknown
with the smallest probable errors

that
Theoria Motus, (1809
in various ways, that this
but
method of com
quantities affected
it is
be observed,
to
the demonstrations are subordinate to the hypo
all
and negative errors of equal magni

tude are equally probable, or that the average of a large
number of results gives the most probable value, and con
thesis, that positive
sequently that the function which represents the probabi

lity
of an error has the form assigned to
The
it
in (132).
limits of this article will not permit us to enter into
further details respecting the applications of the

least
squares.
errors of results
On
method of
the general theory of the probable
deduced from observation, and the most
advantageous methods of combining equations of condition,

the reader
bilites
may
consult the Tfieorie Analytique des
of Laplace
the Theoria
Motus of Gauss
Proba-
the The
oria Combinationis Observationum, and the Supplementum

Theories Combinations, &c. (Gottingen, 1828) of the same
author
the Recherches sur la Probabilite des Jugements,
with the
two Memoirs of Poisson
des Terns for 1827 and 1832
by Mr.
in the
Connaissance
and three masterly papers,
Ivory, in the Philosophical
Magazine
for 1825.
In
the volumes of the Berliner Astronomisches Jahrbuch for
1833, 1834, and 1835,

at great length,
M. Encke has
treated the subject
and given a number of formulae calculated
OF LEAST SQUARES.
to facilitate the labours of the
refer, in conclusion,
computer.
to a very recent
211
We
may
and remarkable
also
dis
quisition on the theory of probable errors, by the celebrated

astronomer Bessel, forming Nos. 358 and 359 of Schuma
cher
Astronomische Abhandlungen, Altona, October 1838.
TABLE
OF THE
VALUES OF THE INTEGRAL

2
for intervals
each
= 01, from r=0

and second
to
r=3, with
differences.
their first
TABLE, CONTINUED.
213
214
TABLE, CONTINUED.
TABLE, CONTINUED.
215
ERRATA.
Page 40,
note, for
Essai sur
les
Probabilites, read
Theorie
Analytique des Probabilites.

43, for adopt, read adapt.
In the note, read
with.
49, line 13, read beginning
Traite Elememuire.
89, line 9, for
mm,
read
m+m
7
RETURN TO
DAY USE
WHICH BORROWED
pM
f OY.
# 7* M *i ,^cc
*
This publication is due on the LAST

and HOUR stamped below.
DATE
^
.SI
RB
17-50m-7 65
(F5759slO)4188
General Library
University of California
Berkeley
4--
609735
UNIVERSITY OF CALIFORNIA LIBRARY

Galloway (A Treatise On Probability - 7ed Encyclopaedia Britannica) (1839) PDF

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Galloway (A Treatise On Probability - 7ed Encyclopaedia Britannica) (1839) PDF

Transféré par

Droits d'auteur :

Formats disponibles

GIFT OP

THE ARTICLE UNDER THAT HEAD IN THE SEVENTH EDITION OF

THE ENCYCLOPEDIA BRITANNICA.

SECRETARY OF THE ROYAL ASTRONOMICAL SOCIETY.

BOOKSELLERS TO HER MAJESTY FOR SCOTLAND.

JACK, Printers, Niddry

GENERAL PRINCIPLES OF THE THEORY OF PROBABILITY.

Chance and Proba

1-4. Definition of the terms

formulae for simultaneous and successive events.

OF THE PROBABILITY OF EVENTS DEPENDING ON A REPETITION OF

COMPOUNDED OF ANY NUMBER OF SIMPLE EVENTS, THE

CHANCES IN RESPECT OF WHICH ARE KNOWN A PRIORI AND CON

11, 12, Probabilities of the different

Application of the binomial

Artifice for abbreviating the calculation.

required to render the probability of an event

equal to an assigned fraction.

OF THE PHOB ABILITY OF EVENTS DEPENDING ON A REPETITION OF

COMPOUNDED OF ANY NUMBER OF SIMPLE EVENTS, THE

CHANCES IN RESPECT OF WHICH ARE KNOWN A

THE DIFFERENT TRIALS

26. Expression for the probability of the different pos

sible results in particular cases.

Extension to the general

OF MATHEMATICAL AND MORAL EXPECTATION

31. Definition of the term

to express the value of a moral

sequences of the formulae.

33. Mathematical expectation not

Application of the theory of Ber

noulli to the subject of insurances.

OF THE PROBABILITY OF FUTURE

the formulae to physical or moral events.

49. Case in which

are not equally probable.

the different causes

OF BENEFITS DEPENDING ON THE PROBABLE DURATION OF

Principles on which the probability of life

nuities on joint lives.

annuity for terms of years.

the survivor of any

Methods of computing the values of assurances on

the value of an annuity on a

OF THE APPLICATION OF THE THEORY OF PROBABILITY TO TESTIMONY,

AND TO THE DECISIONS OF JURIES AND TRIBUNALS

64, 65. Expression for the probability of an event at

which the character of the witness

pression of the probability of an attested event, regard being had

events attested by several witnesses.

71. Formulae for the case of

Successive testimony, or tradition.

Application to the verdicts of juries.

acquittal and condemnation under different hypotheses.

Probability of a verdict being correct

Numerical expression for the error of a verdict when

arbitrary values are given to the constants. 85,86.

Values of the con

deduced from the records of the criminal courts in France.

OF THE SOLUTIONS OF QUESTIONS INVOLVING LARGE NUMBERS. P. 128.

the most probable event in a large

the rapid diminution of the terms

of the binomial towards the beginning arid end of the series