This strategy buys one or more of [[DDM]] [[SSO]] [[QLD]] [[DXD]] [[SDS]] [[QID]] when the RSI(2) crosses below 10. The buy is made on the open the day following the signal. The essence of the system is that the ultralong and ultrashort ETFs put the trader in the Dow, S&P, and Nasdaq when they are overbought or oversold, just before a reversal.
The trade is closed when the RSI(2) crosses above 80. The sell is made on the open the day following the signal.
Only 3 positions can be entered per day, and the system allows a maximum of 3 positions at one time.
A stop of 15% was used.
Starting Capital: $100,000Â Â Â Â Â Ending Capital $140,545.59
Take a look at the actual trades here. Long/Short Index ETF RSI(2) Trades
Further testing needs to give consideration to another type of stop. Perhaps if an entry is made, and RSI(2) rises back above 10, but then crosses below 10 again, a stop is executed.
As this RSI(2) stategy relies on a pullback or reversal to achieve gains, subsequent testing may reveal that the addition of a moving average filter may avoid entries when the market embarks on a strong trend, with very few significant pullbacks. An example of this kind of trend is the sharp decline in early January and the recent sharp uptrend that started in March. In a strongly trending environment, the first RSI(2) trades that occur after the trend reversal result in drawdowns.
Wood,
Would you be going long DDM right about here?
Kurt, no, I wouldn’t be buying it right here. I’m already long it, but I bought when RSI(2)=6ish. Right now RSI(2) is near 40.
Woodshedder,
What time frame are using for your charts…5,10, or 15 min.?
Could you post an example?
Edit, never mind, I see after rereading your post you are on a daily time frame.
Mr. Natural, it is my goal to test the strategy using 1 hour bars. In fact, using hourly bars may allow for some optimization so that instead of buying when RSI(2) <10, it becomes <8.
Again, my learning curve on tradestation is STEEP, so, it will be awhile before the testing is completed.
Thanks for posting this. I ran your system using the FXI/FXP pair and it works there too. That always gives me a warm and fuzzy feeling.
Bob, not really my system, as there are several other bloggers out there running similar tests, but you’re welcome nonetheless.
What software did you run FXI/FXP?
Did you buy 2x the FXI to account for the FXP double inverse?
I included those on earlier tests, but have left them off recently. Thanks for reminding me.
Looking at your Google Docs spreadsheet, if you had a stop at 12% you’d get a better result, but that is a historical curve fit and may not hold true for future trades.