Mean Reversion Strategies in Python (Course Review)

/
Mean Reversion Strategies in Python (Course Review)

In this post I will be reviewing the course “Mean Reversion Strategies in Python by Dr. E.P. Chan” (<—affiliate link, use discount code LONGSHORT for an additional 5% off at checkout). In mean reversion strategies, the course author focuses on three categories of mean reversion systems. These are pairs trading, index arbitrage, and cross-sectional long-short strategies. 

Read More

BACKTESTING THE IMPLIED VOLATILITY STRATEGY WITH QUANTOPIAN (5/20/16)

/
BACKTESTING THE IMPLIED VOLATILITY STRATEGY WITH QUANTOPIAN (5/20/16)

We are now 19 weeks into testing this strategy and the results are strong. Since the last published backtest, the volatility of the strategy has declined from 6% to 5%, beta has declined from 0.20 to 0.19. Alpha has increased from 19% to 25%. The Sortino and Information Ratios have increased from 6.41, and 1.63 to 9.44 and 4.68 respectively.  On an absolute return basis the strategy is crushing the SPY benchmark, 10.3% to 0.7%. 

Read More

BACKTESTING THE IMPLIED VOLATILITY STRATEGY WITH QUANTOPIAN (4/09/16)

/
BACKTESTING THE IMPLIED VOLATILITY STRATEGY WITH QUANTOPIAN (4/09/16)

To see the origin of this series click here. 

To summarize, the strategy calculates a SKEW measure using ATM calls and OTM puts for a collection of ETF symbols. It then sorts the symbols into quintiles based on the SKEW factor.

Using daily close/close log return calculations for this strategy has shown exceptional performance as can be seen here. However, translating a successful daily strategy with no transaction costs and perfect trading fills into a robust strategy that can execute and perform well after incorporating the real structure of market trading is a difficult task. In some cases the strategy cannot survive this translation.

Read More

BACKTESTING THE IMPLIED VOLATILITY STRATEGY WITH QUANTOPIAN (2/27/16)

/

FOR A DEEPER DIVE INTO ETF PERFORMANCE AND RELATIVE VALUE SUBSCRIBE TO THE ETF INTERNAL ANALYTICS PACKAGE HERE

To see the origin of this series click here. 

To summarize, the strategy calculates a SKEW measure using ATM calls and OTM puts for a collection of ETF symbols. It then sorts the symbols into quintiles based on the SKEW factor.

Using daily close/close log return calculations for this strategy has shown exceptional performance as can be seen here. However, translating a successful daily strategy with no transaction costs and perfect trading fills into a robust strategy that can execute and perform well after incorporating the real structure of market trading is a difficult task. In some cases the strategy cannot survive this translation.

In order to test the viability of this strategy I used the Quantopian platform, which allows event-based point-in-time simulated trading on real market data. It also allows us to model transaction costs and slippage which can have large impacts depending on the strategy. 

The following backtest is a variation on the original strategy proposed in the series. This strategy does the following:

  • Calculate the SKEW factor using options data and implied volatility for selected ETFs.
  • Sort the ETFs according to the SKEW factor and divide into quintiles.
  • Go long the ETFs in top quintile while shorting the ETFs in the bottom quintile.
  • The strategy is equal weight and market-neutral.
  • Holding period is one week.
  • Trades are initiated on the first day of the week 20 minutes prior to the market close.
  • Rebalancing/liquidation occurs on the first day of the trading week 5 minutes after the market open.
  • The portfolio size is $100,000 USD. 

The strategy continues to impress with an annualized Sharpe ratio exceeding 5, single digit volatility, and extremely low sensitivity to the broad market as evidenced by the sub 20 beta. If I had access to point in time options data I would like to test this strategy during long trending periods to see if it is robust in multiple market environments.