Computational Investing Course Notes

Portfolio Management and Market Mechanics (week 1)

Incentives of portfolio manager

two types of incentives

  1. Mutual fund, ETF etc.
    • incentive: expense ratio, usually less than 1% of the fund -> attract more investors
  2. hedge fund
    • incentive: two and twenty, 2% * fund + 20% * profits -> attract + growth

Metrics for assesing fund performance

Common metrics

  • Annual return
  • Risk
    1. Standard deviation of return
    2. Draw down: how much it went down (max), difference between that trailing high point and the current low point.
    3. Sharpe ratio: reward / risk (include both upward and downward deviation)
    4. Sortino ratio: reward / risk (exclude upward deviation)
  • Jensen's Alpha

Sharpe ratio

a very important metric on assessing performance


a combination of return / risk

\[ S = { E[R - R_f] \over \sqrt{var[R - R_f]}} \]

  • S: sharpe ratio
  • E: expected return of
  • R(f): risk free return, low risk, guaranteed return, need to be subtracted because it's guaranteed
  • var: variation

Simplified version: k * mean(daily return) / stdev(daily return)

  • stdev: standard deviation
  • k: sqrt(250) for daily returns, 250 the number of trading days in a year, 250 for daily, 12 for monthly
  • dividing stdev(daily return) by k to calculate the stdev of daily return

How Prices Move Up and Down

Computing structure inside a hedge fund

// TODO: fill this up after finishing week2

Company Worth and Capital Assets Pricing Model (Week 2)

What is a Company Worth?

multiple methods of estimating a compnay's value:

  • Market cap: # shares outstanding * price
  • Future dividends
  • Book value

Intrinsic value measured by future dividends

future income generated by the asset, and discounting it to the present value

\[ \sum_{i=1}^{\infty} dividend * gamma^i = {dividend * 1 \over (1 - gamma)} \]

gamma: discount rate (<1), how much the company is trustworthy?

Why/How does information/news affect price?

Key: copmany's profitability

E.g. * CEO effectiveness * Cost of raw materials

Fundamental analysis

  • Book value: how much a company is worth if it ceased operating today, sold all its assets and paid off all its debts
    • total assets - (intangible assets (patents, goodwill) + liabilities)
  • Future returns: "intrinsic" value or future revenue

the fundamental approach is to say let's see what the book value is, let's look at the value of future revenue. Put those two together, and that's what a company's worth.

Capital Assets Pricing Model

a model that describes the relationship between systematic risk and expected return for assets, particularly stocks

Read more: Capital Asset Pricing Model - CAPM Definition

General idea

Expected Return = Time value of money + Risk

\[ r_a = Time Value + Risk \] \[ = r_f + \beta_a * (Risk Premium) \] \[ = r_f + \beta_a * (r_m - r_f) \]

  • \(r_f\): Risk-Free rate, usually is a 10-year goveronment bond rate
  • \(\beta_a\): beta of the security, usually measured by stock volatility
  • \(r_m\): expected market return

CAPM assumptions

  • Return on stock has two components
    • Systematic (the market, usually mean S&P 500 in this class)
    • Residual (\(\alpha_a\)): expected value = 0, or is random \(r_a = \beta_a r_m + \alpha_a\)

however, hedge fund manager thinks \(\alpha_a\) is predictable.

What is beta?

A measure of the volatility, or systematic risk, of a security or a portfolio.

Think of beta as the tendency of a security's returns to respond to swings in the market.

Read more: Beta - Complete Guide To Investment Companies, Funds And REITs

  • beta > 1 indicates that the security's price will be more volatile than the market.
  • beta = 1 indicates that the security's price will move with the market.
  • 0 < beta < 1 means that the security will be less volatile than the market.

  • beta = 0 regardless of which way the market moves, the value of cash remains unchanged. It could mean that the stock either is a new issue or doesn't yet have a beta calculated for it. (E.g. In Yahoo finance)
  • beta < 0 indicate an inverse relation to the market - which is possible but highly unlikely.

Beta and Correlation are different!

Correlation is related to correlation coefficient

Excess returns, are returns greater than the marktet, can only be accomplished if the beta of your portolio is greater than one. Also keep in mind: greater beta, greater risk

Manipulating Data in Python and QSTK (week 3)


Efficient Markets Hypothesis and Event Studies, Portfolio Optimization and the Efficient Frontier (week 4)

3 versions of Efficient Markets Hypothesis

  • weak
  • semi-strong
  • strong

Investopedia wikipedia

Weak-form efficiency

Current prices reflect all past public available information

  • prohibits profit from technical analysis

Semi-strong-form efficiency

Weak + price instantly change to reflect new public information

  • prohibits profit from technical analysis and fundamental analysis

Strong-form efficiency

Semi-strong + prices instantly reflect even hidden or "insider" information

  • prohibits profit from insider information

Portfolio optimization

optimizae for portiolio's reward and risk

Mean Variance optimization

  • mean: return
  • variance: risk
  • optimization: find the balance between them

Model's input:

  • Expected return for each equity
  • Volitility(risk) for each equity
  • Target return
  • Covariance matrix: for each equity, how does that equity vary compared to the one

Output: Portfolio weights that minimize risk for target return

Investopedia: Coorelation Coefficient

Efficient frontier

The efficient frontier (or portfolio frontier) is a concept in modern portfolio theory introduced by Harry Markowitz in 1952.

It refers to investment portfolios which occupy the 'efficient' parts of the risk-return spectrum. Formally, it is the set of portfolios which satisfy the condition that no other portfolio exists with a higher expected return but with the same standard deviation of return.



Investopedia: efficientfrontier

Extended reading: Capital Market Line

the Capital Market Line(CML) is better than the efficient frontier because it considers the infusion of a risk-free asset in the market portfolio.

Covariance vs Coorelation [side topic]


Covariance is a measure of how much two random variables vary together. It’s similar to variance, but where variance tells you how a single variable varies, co variance tells you how two variables vary together.

\[ Cov(X, Y) = {\sum{(x_i - \bar{X}) (y_i - \bar{Y})} \over (n - 1)} \]

A large covariance can mean a strong relationship between variables. However, you can’t compare variances over data sets with different scales (like pounds and inches). A weak covariance in one data set may be a strong one in a different data set with different scales.


The problem with covariances is that they are hard to compare. The solution to this is to 'normalize' the covariance: you divide the covariance by something that represents the diversity and scale in both the covariates, and end up with a value that is assured to be between -1 and 1: the correlation.


Coorelation Coefficient

Correlation coefficients are used in statistics to measure how strong a relationship is between two variables. There are several types of correlation coefficient: Pearson’s correlation or Pearson correlation is a correlation coefficient commonly used in linear regression.

Pearson correlation

\[ r = { \sum_{i = 1}^n{( x_i - \bar{X}} ) \sum_{i = 1}^n{( y_i - \bar{Y}} ) \over \sqrt {\sum_{i = 1}^n{(x_i - \bar{X})^2}} \sqrt {\sum_{i = 1}^n{(y_i - \bar{Y})^2}} } \]

Alternative formulae:

\[ r = { \sum{x_i y_i} - n \bar{X} \bar{Y} \over{(n-1) S_x S_y} } \]

where $ S_x, S_y $ is standard deviation of x and y

Wiki/Correlation coefficient statisticshowto/Correlation coefficient

Advantages of the Correlation Coefficient

The Correlation Coefficient has several advantages over covariance for determining strengths of relationships:

  • Covariance can take on practically any number while a correlation is limited: -1 to +1.
  • Because of it’s numerical limitations, correlation is more useful for determining how strong the relationship is between the two variables.
  • Correlation does not have units. Covariance always has units
  • Correlation isn’t affected by changes in the center (i.e. mean) or scale of the variables

Digging into Data (week 5)


Fundamental law, CAPM for Portfolios (week 6)


check CAPM above

Information Feeds and Technical Analysis (week 7)

Technical Analysis

Three branches

  • Sentiment indicators
    • "emotions of inestors"
  • Flow of funds indicators
    • How much cash has been flowing "in" or "out"
    • What is the capacity of the market to buy or sell
  • Market structure indicators
    • Most indicators fall in this group


  • Simple Moving Average
  • Moving Aveage Convergence Divergence(MACD)
  • Bollinger Band