Divergence and Gambling Mathias Winther Madsen - - PowerPoint PPT Presentation

divergence and gambling
SMART_READER_LITE
LIVE PREVIEW

Divergence and Gambling Mathias Winther Madsen - - PowerPoint PPT Presentation

Divergence and Gambling Mathias Winther Madsen mathias.winther@gmail.com Institute for Logic, Language, and Computation University of Amsterdam March 2015 Divergence Definition If X is a random variable with probability densities p ( x ) ,


slide-1
SLIDE 1

Divergence and Gambling

Mathias Winther Madsen mathias.winther@gmail.com

Institute for Logic, Language, and Computation University of Amsterdam

March 2015

slide-2
SLIDE 2

Divergence

Definition

If X is a random variable with probability densities p(x), then the surprisal associated with the value X = x is s(x) = log 1 p(x). The entropy is the average surprisal, H = E[p(X)].

slide-3
SLIDE 3

Divergence

x r a b p(x)

1/4 1/2 1/4

p-code 00 1 01 q(x)

1/2 1/4 1/4

q-code 10 11

rabarbararabaa ...

slide-4
SLIDE 4

Divergence · rabarbararabaarbabra

5 10 15 20 25 30 35 40 Observation Surprisal

slide-5
SLIDE 5

Divergence

Definition

The Kullback-Leibler divergence from p to q is D(p || q) = E

  • log

1 q(X)

  • − E
  • log

1 p(X)

  • =
  • x

p(x) log p(x) q(x), where E[ · ] is the expectation with respect to p.

Properties of the KL divergence

  • 1. Nonnegativity: D(p || q) ≥ 0;
  • 2. Coincidence: D(p || q) = 0 if and only if p = q;
  • 3. Asymmetry: Often, D(p || q) = D(q || p).

Solomon Kullback and Richard A. Leibler: “Information and Sufficiency,” Annals of Mathematical Statistics, 1951.

slide-6
SLIDE 6

Divergence

0.5 1 1 2 0.3 1 1 2 x 1 2 p(x) .5 .5 x 1 2 p(x) .3 .7

slide-7
SLIDE 7

Divergence

x 1 2 3 p(x)

1/4 1/4 1/2

slide-8
SLIDE 8

Divergence a b

.38 1 .62 p(abaab . . . ) = p(a) · p(b | a) · p(a | b) · p(a | a) · · · q(abaab . . . ) = q(a) · q(b) · q(a) · q(a) · q(b) · · ·

slide-9
SLIDE 9

Gambling

Problem

A horse race has three horses: x 1 2 3 p(x) .2 .3 .5

  • (x)

4 4 2 b(x) ? ? ? Which capital distribution b maximizes the expected rate of return, E[ R ] = E [ o(X) b(X) ]?

slide-10
SLIDE 10

Gambling

x 1 2 3 p(x) .2 .3 .5

  • (x)

4 4 2 b(x) 1 0 2 4 6 8 10 20 40

slide-11
SLIDE 11

Gambling

Daniel Bernoulli: “Specimen Theoriae Novae de Mensura Sortis,” Commentarii Academiae Scientiarum Imperialis Petropolitanae, 1738. John L. Kelly, Jr: “A New Interpretation of Information Rate,” Bell System Technical Journal, 1956.

slide-12
SLIDE 12

Gambling

Definition

The doubling rate is the logarithm of the rate of return, W(X) = log R(X) = log o(X)b(X).

Proportional Gambling

The doubling rate attains its maximum at b∗ = p, regardless of the odds. x 1 2 3 p(x) .2 .3 .5

  • (x)

4 4 2

slide-13
SLIDE 13

Gambling

200 400 600 800 1,000 10−1 100 101 102 103 104 105 Rounds Capital

slide-14
SLIDE 14

Gambling

Problem: Dependent Roulette

A bookmaker draws the 52 cards in a deck one by one: , , , , , , , . . . , , Before each draw, you can place bets on and . Is this a favorable game, and what is its doubling rate? Cover and Thomas (1991), Example 6.3.1.