Pairs trading strategy on bond futures
Hello everyone,
I'm new here. This is a simple pairs trading formula used to calculate the price difference between two assets (Equation 1).
Therefore, I backtested a simple pairs trading strategy on the German bond market, specifically on
Euro-bund/Bobl pair (Euro-bund = 10-year Treasury futures, Bobl = 5-year Treasury German bond futures). I have not tested the correlation between the two assets, but overall it is quite high (over 0.90 or 90%).
So, here are the backtests I did over the past five years (December 1, 2009 - October 17, 2014). In the data I used, there are some gaps (excluding data between February 28, 2013 and April 29, 2013 and
Except August 30, 2013 and February 10, 2013). The price used is the approximate price of Euro-Bund,
BOBL futures (EUREX) from December 1, 2009 to October 17, 2014.
Here is the formula for measuring the difference between two assets:
The following are the trading rules:
• No stop loss since we assume capital and margin calls are not limits,
• Whenever +1 crosses the 1 threshold upward and +1 crosses the 1 threshold downward, a position is opened.
-1, only one crossover (for example, the signal passes from -0.9 to +1.1 to open a position),
• Opened positions will be closed once the mean (0) crosses downwards for a +1 cross and upwards for a -1 cross • Parameters are i = 1, A is Bobl futures, B is Euro-Bund futures,
• The trading period for futures contracts ending in the current month is the last trading day of the month before the futures contract ends in the month following the end of the month (for example, for Bobl futures in December 2014, the trading period will be June 2, 2014 to August 29, 2014),
• To calculate the signal, the previous 20 data are taken from the same futures contract trading period.
It is possible to backtest more periods from Equation 1, but if I choose to return 1 period and the standard deviation over the past 20 periods, the cause of the daily pricing error is 1
20 represents the average volatility (in days) of returns over the previous trading month.
The result is as follows:
• Sharpe ratio = 0.25
• First withdrawal from November 29, 2010 to July 30, 2012 approximately 10.6%
• The second drawdown from June 24, 2013 to September 6, 2014 was approximately 3.3%
• 27 trades in total • 74.07% win rate
Here are the descriptive statistics:
in conclusion:
The results are very interesting and result from no parameter optimization due to lack of data (two gaps) and data mining or data snooping bias (which means over-optimizing model parameters based on transient noise in historical data).
The histograms of profits and losses show positive skewness, so this pairs trading strategy appears to be profitable. This can be confirmed by bootstrapping or Monte Carlo simulation, but only 27
The conditions presented in the sample are borderline. Furthermore, the two gaps in the data put the trading results at odds, as huge losses could occur.
More than 66% of trades are closed after 1 day or 2 consecutive trading days. The winning rate is 74.07%, and most of the losses are swallowed up by profits. This short exposure to the German bond futures market is interesting because it is possible to add some of the same currency pairs to a trading strategy on the same futures account. By doing so, profits may be greater and losses may be greater as well.
I'm new here. This is a simple pairs trading formula used to calculate the price difference between two assets (Equation 1).
Therefore, I backtested a simple pairs trading strategy on the German bond market, specifically on
Euro-bund/Bobl pair (Euro-bund = 10-year Treasury futures, Bobl = 5-year Treasury German bond futures). I have not tested the correlation between the two assets, but overall it is quite high (over 0.90 or 90%).
So, here are the backtests I did over the past five years (December 1, 2009 - October 17, 2014). In the data I used, there are some gaps (excluding data between February 28, 2013 and April 29, 2013 and
Except August 30, 2013 and February 10, 2013). The price used is the approximate price of Euro-Bund,
BOBL futures (EUREX) from December 1, 2009 to October 17, 2014.
Here is the formula for measuring the difference between two assets:
The following are the trading rules:
• No stop loss since we assume capital and margin calls are not limits,
• Whenever +1 crosses the 1 threshold upward and +1 crosses the 1 threshold downward, a position is opened.
-1, only one crossover (for example, the signal passes from -0.9 to +1.1 to open a position),
• Opened positions will be closed once the mean (0) crosses downwards for a +1 cross and upwards for a -1 cross • Parameters are i = 1, A is Bobl futures, B is Euro-Bund futures,
• The trading period for futures contracts ending in the current month is the last trading day of the month before the futures contract ends in the month following the end of the month (for example, for Bobl futures in December 2014, the trading period will be June 2, 2014 to August 29, 2014),
• To calculate the signal, the previous 20 data are taken from the same futures contract trading period.
It is possible to backtest more periods from Equation 1, but if I choose to return 1 period and the standard deviation over the past 20 periods, the cause of the daily pricing error is 1
20 represents the average volatility (in days) of returns over the previous trading month.
The result is as follows:
• Sharpe ratio = 0.25
• First withdrawal from November 29, 2010 to July 30, 2012 approximately 10.6%
• The second drawdown from June 24, 2013 to September 6, 2014 was approximately 3.3%
• 27 trades in total • 74.07% win rate
Here are the descriptive statistics:
in conclusion:
The results are very interesting and result from no parameter optimization due to lack of data (two gaps) and data mining or data snooping bias (which means over-optimizing model parameters based on transient noise in historical data).
The histograms of profits and losses show positive skewness, so this pairs trading strategy appears to be profitable. This can be confirmed by bootstrapping or Monte Carlo simulation, but only 27
The conditions presented in the sample are borderline. Furthermore, the two gaps in the data put the trading results at odds, as huge losses could occur.
More than 66% of trades are closed after 1 day or 2 consecutive trading days. The winning rate is 74.07%, and most of the losses are swallowed up by profits. This short exposure to the German bond futures market is interesting because it is possible to add some of the same currency pairs to a trading strategy on the same futures account. By doing so, profits may be greater and losses may be greater as well.
Attachment to original post (1)
📦 Summary of post attachments (1)
Below are all the files shared in the reply (1).
💡 Featured Recommendations
✍️ Latest by the author
- •
- •
- •
- •
- •
- •
📌 Popular topics
- •
- •
- •
- •
- •
- •
- •
- •
🔗 You May Be Interested In
- •
- •
- •
- •
- •
- •