Thursday, May 29, 2008

RSI 2 and the Qs

I don't want to overlook the indicative power of the RSI2 for the Qs. Here's the daily study of the Qs looking back 5 years based on the same TS code as the Tuesday IWM RSI post. The settings are optimized to 2,28,80,28,80 possibly reflecting the variance in beta between the IWM and the Qs.
This is a fairly robust system and with an overall performance > 80%, it may be worth taking marginal swing trade positions (IMHO). With 242 trades over the past 64 months, that's about 1 trade a week-- a time frame I find almost comfortable. The number of longs versus shorts maintains a good balance and the win/loss ratio on the long side is particularly attractive. With an average of 2 bars in the winning trades and 8 bars in the losing trades, a stop loss/trailing stop should only improve results and reduce overall exposure time. Ditto for the short side. The intraday drawdown on both the longs and shorts needs to scaled back for more risk adverse traders, and again, stops can help achieve those ends.
Further backtesting (with stops) is suggested and additional confirming signals may also improve performance. Any conditional changes to the system would likely reduce trade frequency.
I utilize the RSI2 daily bars as a fractal momentum indicator to skew my intraday trades. If momentum is positive, I'm inclined to buy dips and upward pivot crosses. If momentum is negative, I'm more inclined to fade the rallies and sell pivot breakdowns.
It's also possible to build a short term options trading model based on RSI2 Qs behavior, buying near term calls (or selling puts) on positive momentum and buying puts (or selling calls) on negative momentum. I pefer to be a seller of options, rather than a buyer, but the risk potential is considerably greater and not every trader is comfortable with this type of exposure.
A couple RSI focused articles in this month's Stocks and Commodities. One tests the RSI to trade a basket of 100 ETFs (capital requirements and commissions make this impractical for most traders), and another expounding the virtues of RSI as an indicator that works. This is a statistical study and helps explain the why of the Qs support and resistance numbers used in today's study (28 and 80).

1 comment:

LP said...

yeah I've been looking at the RSI 2/4. Very nice. However, since I cancelled my breakstation account, I've needed to do everything in SQL. Makes it lot harder but it's nice to do multi-time frame backtest/analysis in a matter of seconds.

Also, another nice gap fill on all indicies on the futs.