Topic B5 — Discrete Distributions
Table of contents
- Topic B5 — Discrete Distributions
- Discrete vs Continuous RV
- Probability Mass Function vs. Probability Density Function
- Expected value of a discrete RV
- Variance of a discrete RV
- Properties of the expected value and the variance of two RV
- Expected value
- Var/covar
- Risk neutral vs. risk averse vs. risk lovers
- Portfolio risk and return
- Bernoulli trial
- Binomial distribution interpretation
- Properties of the binomial distribution
Discrete vs Continuous RV
Discrete vs Continuous
Probability Mass Function vs. Probability Density Function
- Relative freq
- Mass for discrete
- Density for continuous
- Discrete properties
- Prob of each value is btw 0 and 1
- Sum of prob = 1
- Cumul distrib fnct (CDF)
- “Proba that X is at least val x” P(X <= x)
Expected value of a discrete RV
Expected value converges to the pop mean
$ E(X) = \mu = \sum x_iP(X=x_i) $ where $X = x_i$ is the probability of hitting $x_i$
Variance of a discrete RV
$ Var(X) = \sum (x_i-\mu)^2P(X=x_i) $ just the variance multiplied by the probability.
SD $ \sqrt{Var(X)} $
Properties of the expected value and the variance of two RV
Expected value
- $ E(X+Y) = E((X)+E(Y)) = E(X) + E(Y) $
- Constant $ \alpha$ : $ E(\alpha) = \alpha, E(\alpha X) = \alpha E(\alpha) $
Var/covar
$ Covar(X+Y) = E(XY) - E(X)E(Y) $
$ E[(X-\mu_x)(Y-\mu_y)] $
or $ E(XY) - E(X)E(Y) $ where $ E(XY) $ is the joint probability
Prop
- $ Var(X+Y) = Var(X) + Var(Y) - 2Cov(X,Y) $
- $ Var(\alpha) = 0 $ and $Var(\alpha X) = \alpha^2 Var(X) $
Risk neutral vs. risk averse vs. risk lovers
- Averse: might decline a risky prospect even if $E(X) >0$
- Neutral: always accepts if $E(X) >0$
- Loving: might decline a risky prospect even if $E(X) <0$