# Welcome back

## Find a book, put up your feet, stay awhile

Sign in with Facebook

Sorry, we are unable to log you in via Facebook at this time. Please try again later.

or

Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more

Download

Standard view

Full view

of .

Look up keyword

Like this

Share on social networks

4Activity

×

0 of .

Results for: No results containing your search query

P. 1

Column 24 Understanding the Kelly Criterion 2Ratings: (0)|Views: 18|Likes: 5

Published by TraderCat Solaris

See more

See less

https://www.scribd.com/doc/33566593/Column-24-Understanding-the-Kelly-Criterion-2

12/15/2011

text

original

Column overall title:A Mathematician on Wall StreetColumn 24

Understanding The Kelly Criterion – Part II

by Edward O. ThorpCopyright 2008During a recent interview in the

Wall Street Journal

(March 22-23,2008) Bill Gross and I discussed turbulence in the markets, hedge funds andrisk management. Bill considered the question of risk management after heread

Beat the Dealer

in 1966. That summer he was off to Las Vegas to beatblackjack. Just as I did some years earlier, he sized his bets in proportion tohis advantage, following the Kelly Criterion as described in

Beat the Dealer,

and ran his $200 bankroll up to $10,000 over the summer. Bill has gone frommanaging risk for his tiny bankroll to managing risk for Pacific InvestmentManagement Company’s (PIMCO) investment pool of almost $1 trillion. Hestill applies lessons he learned from the Kelly Criterion. As Bill said, “Here atPIMCO it doesn’t matter how much you have, whether it’s $200 or $1 trillion.… Professional blackjack is being played in this trading room from thestandpoint of risk management and that’s a big part of our success.” The Kelly Criterion applies to multiperiod investing and we can getsome insights by comparing it with Markowitz’s standard portfolio theory forsingle period investing.

Compound Growth and Mean–Variance Optimality

Nobel Prize winner Harry Markowitz introduced the idea of mean-variance optimal portfolios. This class is defined by the property that, amongthe set of admissible portfolios, no other portfolio has both higher meanreturn and lower variance. The set of such portfolios is known as the efficientfrontier. The concept is a cornerstone of modern portfolio theory, and themean and variance refer to one period arithmetic returns. In contrast theKelly Criterion is used to maximize the long term compound rate of growth, amultiperiod problem. It seems natural, then to ask the question: is there ananalog to the Markowitz efficient frontier for multiperiod growth rates, i.e. arethere portfolios such that no other portfolio has both a higher expectedgrowth rate and a lower variance in the growth rate? We’ll call the set of such portfolios the compound growth mean-variance efficient frontier.Let’s explore this in the simple setting of repeated independentidentically distributed return per unit invested, the random variables

{ }

: 1, ,

i

X i n

=

K

with

( )

0

i

E X

>

so the “game” is favorable, and where the non-negative fractions bet at each trial are specified in advance

{ }

: 1, , .

i

f i n

=

K

Tokeep the math simpler, we also assume that the

i

X

are not constant andhave a finite number of distinct values. After

n

trials the compound growth6/25/20101

rate per period is

{ }

( )

( )

1

1log 1

ni i ii

G f f X n

=

= +

∑

and the expected growth rate

{ }

( )

{ }

( )

( ) ( )

( )

1 1

1 1log 1 log 1 log 1 .

n ni i i i ii i

g f E G f E f X E f X E fX n n

= =

= = + = + ≤ +

∑ ∑

Thelast step follows from the (strict) concavity of the log function,

X

has thecommon distribution of the

,

i

X

we define

1

1

nii

f f n

=

=

∑

and we have equality if and only if

i

f f

=

for all

.

i

Therefore if some

i

f

differ from

f

we have

{ }

( )

{ }

( )

.

i

g f g f

<

This tells us that betting the same fixed fraction alwaysproduces a higher expected growth rate than betting a varying fraction withthe same average value. Note that whatever

f

turns out to be, it can alwaysbe written as

*

,

f cf

=

a fraction

c

of the Kelly fraction.Now consider the variance of

{ }

( )

.

i

G f

If

X

is a random variable with

( ) ( )

, 1 ,

P X a p P X q

= = = − =

and

0

a

>

then

( )

2

1ln 1 ln .1

af Var fX pq f

+ + = −

(Compare Thorp, 2007, section 3.1). Note: the change of variable

, 0,

f bh b

= >

shows the results apply to any two valued random variable. Wechose

1

b

=

for convenience.) A calculation shows that the second derivativewith respect to

f

is strictly positive for

0 1

f

< <

so

( )

ln 1

Var fX

+

is strictlyconvex in

.

f

It follows that

{ }

( )

( ) ( )

( )

1 1

1 1log 1 log 1 log 1

n ni i i ii i

Var G f Var f X Var f X Var fX n n

= =

= + = + ≥ +

∑ ∑

with equality if and only if

i

f f

=

for all

.

i

Since every admissible strategy istherefore “dominated” by a fractional Kelly strategy it follows that the mean-variance efficient frontier for compound growth is a subset of the fractionalKelly strategies. If we now examine the set of fractional Kelly strategies

{ }

{ }

*

f cf

=

we see that for

0 1,

c

≤ ≤

both the mean and the variance increaseas

c

increases but for

1,

c

≥

the mean decreases and the variance increasesas

c

increases. Consequently

{ }

*

f

dominates the strategies for which

1

c

>

and they are not part of the efficient frontier. No fractional Kelly strategy isdominated for

0 1.

c

≤ ≤

We have established

in this limited

setting: Theorem. For repeated independent trials of a two valued randomvariable, the mean-variance efficient frontier for compound growth

over afinite number of

trials

consists precisely of the fractional Kelly strategies

{ }

*

:0 1 .

cf c

≤ ≤

So, given any admissible strategy, there is a fractional Kelly strategywith

0 1

c

≤ ≤

which has a growth rate that is no lower and a variance of thegrowth rate that is no higher. The fractional Kelly strategies in this instanceare preferable in this sense to all the other admissible strategies, regardless6/25/20102

of any utility function upon which they may be based. This deals with yetanother objection to the fractional Kelly strategies, namely that there is avery wide spread in the distribution of wealth levels as the number of periodsincreases. In fact this eventually enormous dispersion is simply themagnifying effect of compound growth on small differences in growth rateand we have shown in the theorem that in the two outcome setting thisdispersion is minimized by the fractional Kelly strategies. Note that in thissimple setting, a utility function will choose a constant

i

f cf

=

which willeither be a fractional Kelly with

1

c

≤

in the efficient frontier or will be toorisky, with

1,

c

>

and not be in the efficient frontier.As a second example suppose we have a lognormal diffusion processwith instantaneous drift rate

m

and variance

2

s

where as before theadmissible strategies are to specify a set of fixed fractions

{ }

i

f

for each of

n

unit time periods,

1, , .

i n

=

K

Then for a given

f

and unit time period

( )

2 2

VarG f s f

=

as noted in (Thorp, 2007, eqn. (7.3)). Over

n

periods

{ }

( )

2 2 2 21 1

n ni ii i

VarG f s f s f

= =

= ≥

∑ ∑

with equality if and only if

i

f f

=

for all

.

i

Thisfollows from the strict convexity of the function

( )

2

.

h x x

=

So the theoremalso is true in this setting. I don’t currently know how generally the convexityof

( )

ln 1

Var fX

+

is true (I suspect rather broadly) but whenever it is, and wealso have

( )

ln 1

Var fX

+

increasing in

,

f

then the compound growth meanvariance efficient frontier is once again the set of fractional Kelly strategieswith

0 1.

c

≤ ≤

Samuelson’s Criticisms

The best known “opponent” of the Kelly Criterion is Nobel Prize winningeconomist Paul Samuelson, who has written numerous polemics, bothpublished and private, over the last 40 years. William Poundstone’s book

Fortune’s Formula

gives an extensive account with references. The gist of itseems to be:(1) Some authors once made the error of claiming, or seeming toclaim, that acting to maximize the expected growth rate (i.e. logarithmicutility) would approximately maximize the expected value of any othercontinuous concave utility (the “false corollary”).Response: Samuelson’s point was correct but, to me and others,obvious the first time I saw the false claim. However, the fact that somewriters made mistakes has no bearing on an objective evaluation of themerits of the criterion. So this is of no further relevance.(2) In private correspondence Samuelson has offered examples andcalculations in which he demonstrates, with a two valued

X

(“stock”) andthree utilities,

( )

1/ ,

H W W

= −

( )

log ,

K W W

=

and

( )

1 2

,

T W W

=

that if any onewho values his wealth with one of these utilities uses one of the other utilitiesto choose how much to invest then he will suffer a loss as measured with his6/25/20103

- Read and print without ads
- Download to keep your version
- Edit, email or read offline

© Copyright 2015 Scribd Inc.

Language

Choose the language in which you want to experience Scribd:

Sign in with Facebook

Sorry, we are unable to log you in via Facebook at this time. Please try again later.

or

Password Reset Email Sent

Join with Facebook

Sorry, we are unable to log you in via Facebook at this time. Please try again later.

or

By joining, you agree to our

read free for two weeks

Unlimited access to more than

one million books

one million books

Personalized recommendations

based on books you love

based on books you love

Syncing across all your devices

Join with Facebook

or Join with emailSorry, we are unable to log you in via Facebook at this time. Please try again later.

Already a member? Sign in.

By joining, you agree to our

to download

Unlimited access to more than

one million books

one million books

Personalized recommendations

based on books you love

based on books you love

Syncing across all your devices

Continue with Facebook

Sign inJoin with emailSorry, we are unable to log you in via Facebook at this time. Please try again later.

By joining, you agree to our

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

CANCEL

OK

You've been reading!

NO, THANKS

OK

scribd