Market Sentiment Oscillator at Lows

Earlier this fall, Dr. Brett Steenbarger of TraderFeed published an excellent series of articles outlining three market sentiment indicators based on inter-market ratios. The well developed articles presenting the rational follow (with permission):

1) Gold: A Sentiment Measure For Technology

2) Using Sector Relationships to Discover Trader Sentiment

3) Using Global Performance As A Gauge of Trader Psychology

Intrigued and inspired by the good doctor, I developed a simple unified sentiment oscillator and hypothetical trading system incorporating the core concepts from these three posts plus an additional ratio relating the Dow Transports to the S&P500. The indicator is calculated by first adding the log ratios of EEM:SPY, XLK:^XAU, XLY:XLP, and IYT:SPY. A ten-day ratio of the result then converts this first calculation to an oscillator.

You will find that the oscillator ranges fairly consistently between 0.99 and 1.01, which indicate short-term market despair and euphoria, respectively. Last week’s ugly market pushed this indicator down below .990, registering in the bottom 1% of all daily readings since early 2003. A hypothetical trading system based on this oscillator follows:

Buy – Buy the SPY at the close when the oscillator falls below 0.996 and the two-day RSI of the SPY is below 40. These criteria were most recently satisfied on November 5, suggesting a long position held against the sentiment extreme. Filtering signals when short-term RSI is high reduces total gains somewhat, but reduces system volatility to an even greater extent by avoiding high-level entries.

Sell – Sell after the earlier 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, last week being no exception!

During the 1,221 trading days back through January 1, 2003, excluding the open trade 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. A five-year equity curve for the system is shown below:

During this same period the S&P 500 (SPY) gained about 76%. As nice as that equity curve looks from a 20,000 foot level (RSQ = 97%), as a system it undoubtedly requires further stress testing and development before trading. 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 during this volatile market period.

Disclosure: The author holds long positions in SPY