A Backtest A Day

Everyday I will backtest a different quant strategy for the markets!

Types of Strategies

• Sector Rotation
• Precious Metals
• Agriculture Commodities

Best Long Only vs Long/Short Backtested Strategies

Best Sector Rotation Strategies

Best Long Only

Best Long Short

Best Precious Metal Strategies

Best Long Only

Best Long Short

Best Agriculture Commodity Strategies

Best Long Only

Best Long Short

Methodology

The main metric I will be using to measure a strategy is the sharpe ratio. What is the sharpe ratio?

Sharpe Ratio

The sharpe ratio is the average return of a strategy divided by its risk. Meaning if a strategy has high returns, and low risk, then it will have a high sharpe ratio. Why this metric and not total profit? Because total profit only keeps track of returns of a strategy, not risk, while sharpe ratio is calculated with both, allowing you to sleep better at night, instead of wondering if you are going to make 100% or lose 100% tomorrow.

How to use strategy yourself

1. Either download the notebook or copy the code into a Python script
2. Remove the end date parameter from yf.download()
3. Print the last row of weighted_signal using print(weighted_signal.iloc[-1]) (I will update the notebooks to already have this)
4. Go long the positive weights and (if the strategy is L/S,) short the negative weights

Other

First I will the evaluate the sharpe of each lookback of a strategy in a 5-fold cross validation to determine the best* parameters, then I will plot the cumulative return of the lookback with the highest sharpe. These backtests DO NOT take fees/slippage into account, in my opinion this isnt a problem for the long-only strategies, but needs to be considered for long/short strategies.

Abbreviations

• TS = Time Series
• L\S = Long Short
• XS = Cross Sectional
• Vol = Volatility
• MOM = Momentum
• STR = Short-term Reversal

Credits

Code was provided by @quant_arb on Twitter, I added the cross-validation to try to better pick parameters.