Wednesday, November 7, 2007

Sentiment Trading System

In my prior post, I outlined a market sentiment oscillator building on work published by Dr. Brett Steenbarger. In this post I will present a hypothetical trading system based on that oscillator, as follows:

Buy – Buy the SPY at the close when the oscillator falls below 0.995 and the two-day RSI of the SPY is below 40. This criteria was met three days ago on November 5. Filtering signals when the RSI is high reduces total gains somewhat, but reduces returns volatility to an even greater extent by avoiding high level entries.

Sell – Sell after the earliest of the following: the oscillator reaches 1.01, the two-day RSI of the SPY exceeds 90, or twenty days has passed. It’s that last condition that gets me… this simple system can generate long holds during very rocky times, today being no exception!

During the 1,221 trading days back through January 1, 2003, this system would have generated a 52% gain before trading costs with 24 wins out of 28 signals (86%), and was invested 25% of the time for an average hold period of 11 days. The equity curve for the system is shown below:

During this same period, the SPY has gained about 76%. As nice as that equity curve looks (RSQ = 97%), I probably wouldn't trade the model because: a) I can’t promise based on my quick study that it is not "best fit" in some way; and, b) I have alternate systems based purely on price that I prefer. However, I find it to be an interesting concept worthy of further research. I also wonder if it holds any intra-day promise. If nothing else, I hope that you find it instructive and a nice distraction on this otherwise brutal market day.



1 comment:

Jeff Pietsch CFA, Esq said...

As of Friday, November 9, this indicator is below .991, where it has been less than 2% of the time during the last 1,200 trading days. Ugly close to an ugly week.