/pairs-pipeline

Pairs trading strategy that includes a research pipeline for identifying and selecting pairs. Tests all possible pairs in a universe for cointegration using the Johansen test, then runs in-sample backtests on all cointegrating pairs, then runs an out-of-sample backtest on the 5 best performing pairs.

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

pairs-pipeline

Pairs trading strategy that includes a research pipeline for identifying and selecting pairs. Tests all possible pairs in a universe for cointegration using the Johansen test, then runs in-sample backtests on all cointegrating pairs, then runs an out-of-sample backtest on the 5 best performing pairs. Calculates daily hedge ratios using the Johansen test and times entries and exits using Bollinger Bands. Trading rules adapted from Ernie Chan's book Algorithmic Trading. Runs in Moonshot on the universe of US ETFs.

Clone in QuantRocket

CLI:

quantrocket codeload clone 'pairs-pipeline'

Python:

from quantrocket.codeload import clone
clone("pairs-pipeline")

Browse in GitHub

Start here: pairs_pipeline/Introduction.ipynb

Find more code in QuantRocket's Codeload Library