Back to Community
How to improve the return of gold and silver pairs trading strategy?

Hello, I am very new to Quantopian and Algorithm Trading. So, I want to share what I have done and learn from all of you.

The Strategy
It is a mean reversion strategy that tries to capitalize the mean-reverting feature of the spread between gold and silver.

Run every day
First of all, I extracted 1 year (around, I don't know how to make it exactly a year) of the price of gold and silver. Then using 20 days moving windows, I calculate the slope and intercept from the linear regression between the two assets. I stored the calculated spreads (should have 365 -20 number of spread, approximately). Finally, I calculate the MEAN and SD of all the spreads.

Entry Run every day
I checked the entry after market started by using the schedule function. If the current spread is higher than 1.5SD + MEAN, I will sell. If the current spread is lower than MEAN - 1.5SD, I will buy. The max. open order is 5. That means the program will continue entering the position if conditions are met. The reason for doing this is that I cannot pick the exact top but I am sure that it will revert to the mean. So, I keep adding positions.

Exit Run every minute
If the current spread reverted to the mean, I will close all positions. However, it is difficult to catch the exactly equal. So, I put a range of 10% error.

I have attached the backtest for your reference.

One problem I got is that I don't know why sometimes, even the 5 opened positions reverted to the mean, I still have negative return. You can try that in the backtest started from 01/01/2017

Clone Algorithm
4
Loading...
Backtest from to with initial capital
Total Returns
--
Alpha
--
Beta
--
Sharpe
--
Sortino
--
Max Drawdown
--
Benchmark Returns
--
Volatility
--
Returns 1 Month 3 Month 6 Month 12 Month
Alpha 1 Month 3 Month 6 Month 12 Month
Beta 1 Month 3 Month 6 Month 12 Month
Sharpe 1 Month 3 Month 6 Month 12 Month
Sortino 1 Month 3 Month 6 Month 12 Month
Volatility 1 Month 3 Month 6 Month 12 Month
Max Drawdown 1 Month 3 Month 6 Month 12 Month
# Backtest ID: 5a2103b2989366443665d485
There was a runtime error.
3 responses

Max Lvrg: 2.15 2016-02-11 is one thing to address. Also take a look at some others that trade GLD:
https://www.google.com/search?q=gld+"def+initialize"+site:quantopian.com

Thanks for your reply!
Sorry. What do you mean by Max Lvrg: 2.15 2016-02-11?

On 2016-02-11, leverage hit a new high of 2.15. It's fine if intended, if accidental can ruin one's day. This might help. Also logging context.portfolio.cash can be useful.