Trade idea – SSO and UBT
Base system rules:
At the end of each month, put 50% of system capital into SSO and 50% into UBT. Rebalance monthly.
Base system results:
CAGR: 15.7%, Max Drawdown: -48.5%, Average Drawdown: -8.2%
Note: see bottom for notes about price estimates prior to inception.
Improved system rules:
If the 7 month rate of return on the base system is positive (on a monthly basis), follow the base system. If the 7 month rate of return is negative then put 50% of system capital into TLT and 50% into cash. Rebalance monthly
Improved system results:
CAGR: 16.8%, Max Drawdown: -21.0%, Average Drawdown: -4.8%
Notes: SSO and UBT are leveraged ETFs. SSO is a 2X tracker of the S&P 500 while UBT is a 2X tracker of long term government bonds. The 2X leverage only applies to a single day. Test period above is 8/1990 to 9/2012. I created estimates of SSO and UBT prior to their inception. My estimates are not perfect by any means. For UBT prior to inception, I use dividend adjusted values of TLT and VUSTX to estimate prices. Differences between my series and UBT are very small during the overlap period. For SSO prior to inception, I use non-dividend adjusted values of VFINX to estimate prices. The non-dividend adjusted values of VFINX created a much better match in the overlap period. My series is not as close of a match for SSO as the one for UBT, but still pretty close. For the improved system, using 6 or 8 months as a filter would have produced similar results.
Posted on October 9, 2012, in Other. Bookmark the permalink. 9 Comments.
As always, you continue to be creative.
I tried a version of this strategy on ETFreplay.com (the only difference is that the cash trigger is when either SSO or UBT go below their 5 month ma) and it works about the same as your version. An interesting tweak is to use GLD (or UGL) as a cash equivalent.
Alex, do you think the use of etfrelay.com really help improve your portfolio return? as you know backtesting isn’t project the futures.
thanks
Great work.. I’m testing a weekly version of the system, i will let you know..
http://nightlypatterns.wordpress.com
http://backtestingvix.wordpress.com/
Greeting aimdal. One thing you can also try out is a risk parity approach. Instead of half and half allocation, allocate based on recent volatilities of each product. This also greatly improved the base system I showed here.
Mike
Here’s another very simple system: buy VTI when it’s above its 100 day sma. Rebalance monthly. If VTI falls below its 100 day sma, buy TLT, unless TLT is also below its 100 day sma, in which case go to cash. Since 2007: CAGR: 15.4% Max DD: 18%.
Mike, I’m new to your site and find it really interesting. Regarding the “improved system”, and when you exit a 50% TLT/50% cash position, does that occur when the 7 month look back turns positive using only SSO/UBT monthly returns or are you using your actual return including the month or months while you are in the TLT/Cash positions? Thank you.
Hi Joel, welcome to the site. Only the 7 month performance of the base system (using only SSO/UBT) is used to determine the allocation.
What about using 3x leverage with TMF/UPRO?
Good question. I haven’t tried it but will add it to the list of things to do.