using QUANT for TRADING

Nitesh Khandelwal

QuantInsti

June 1, 2013

• Using quantitative techniques to build the trading model and execution. Statistical methods and mathematical computations are extensively used while creating the trading model as well as the during the implementation.

Quantitative Trading

• Build a statistical arbitrage trading model using quant with a statistical approach• Statistical tool kit for the strategy

• Strategy building on Excel

• Basic demonstration on R

Agenda

• Cointegration

• Dickey Fuller test

• Stationarity

• Granger Causality

Statistical Tool kit for the strategy

• Two time series are cointegrated if they have a common stochastic drift*. Typically you can determine this by checking if: For two individually non stationary time series, there exists a linear combination of the two time series that is stationary. Example: Drunk man and his dog.

*Stochastic Drift: Change of the average value of a stochastic process. Example: Stock prices

Cointegration

• A stationary time series is one whose statistical properties such as mean, variance, autocorrelation, etc. are all constant over time. Most statistical forecasting methods are based on the assumption that the time series can be rendered approximately stationary (i.e., "stationarized") through the use of mathematical transformations.

• A stationarized series is relatively easy to predict: you simply predict that its statistical properties will be the same in the future as they have been in the past! The predictions for the stationarized series can then be "untransformed," by reversing whatever mathematical transformations were previously used, to obtain predictions for the original series.

Stationarity

• It test for the unit root in an autoregressive model.

yt = ρ yt-1 + ut

• If ρ=1, then a unit root is present and the series is non stationary

Stationarity Test: Dickey Fuller test

• Cointegration gives a better estimate for short term predictions.

• Spurious Correlation. Example: Ice cream sales versus drowning casualties in the lake.

• Empirical Findings

Why Cointegration over Correlation

• Granger causality comes handy for quoting in strategies with multiple legs for execution.

• Time series ”A” is said to Granger-cause time series “B” if it can be shown using statistical tests on past values of ”A” & “B”, that they give statistically significant information about future values of “B”

• Example: Nifty vs. USDINR

Quant for Execution

Thank You!

To Learn Automated Trading

Email: contact@quantinsti.com

Connect With Us:

SINGAPORE

11 Collyer Quay,

#10-10, The Arcade,

Singapore - 049317

Phone: +65-6221-3654

INDIA

A-309, Boomerang,

Chandivali Farm Road, Powai,

Mumbai - 400 072

Phone: +91-022-61691400