Stochastic Oscillator

This is an analysis of an indicator called the Stochastic Oscillator.  The idea is similar to the Relative Strength Index that I described in another analysis.

A value over 80 indicates the stock is overbought, and expected to be weaker than normal going forward. A value under 20 indicates the stock is oversold, and expected to be stronger than normal going forward.

I’ve applied the Stochastic Oscillator to four stocks below (AAPL, BA, CAT, and XOM), for the period 2007 – 2016. That’s a 10-year period that included a major sell-off, a major rally, and some time spent moving sideways.

Before looking at the specific results, let’s discuss what we would expect to see if the indicator has predictive value. We expect that whenever the indicator is less than 20, the average gain over the next N days will be greater than when the indicator is greater than 20. We also expect that whenever the indicator is greater than 80, the average gain over the next N days will be less than when the indicator is less than 80.

In the table I’ve included values of N for 5 trading days, 30 trading days, and 110 trading days (about 6 months). Let’s see how the indicator worked out:

 AAPLBACATXOM
Days < 20362354490441
Days > 80915739711702
Days 20 - 801376156014521510
Avg 5 day Return, <20+0.06%+0.48%+0.52%+0.58%
Avg 30 day Return, <20+2.95%+1.93%+2.29%+0.93%
Avg 110 day Return, <20+17.12%+3.82%+5.69%+2.93%
Avg 5 day Return, >80+0.82%+0.27%+0.32%-0.09%
Avg 30 day Return, >80+4.36%+2.51%+2.02%+0.37%
Avg 110 day Return, >80+13.99%+9.58%+9.92%+2.87%
Avg 5 day Return, 20-80+0.60%+0.31%+0.19%+0.06%
Avg 30 day Return, 20-80+3.58%+1.58%+1.40%+0.53%
Avg 110 day Return, 20-80+13.38%+7.55%+5.85%+1.15%

Five day returns when the indicator is less than 20 are better than when they’re greater than 20 for three of the four.  Thirty day returns seem to be about the same, as do 110 day returns.

Five day returns when the indicator is greater than 80 don’t seem to be much different than when they’re less than 80.  Thirty day returns also seem to be about the same.  110 day returns, opposite of what we would expect, seem to be a bit better when the indicator is greater than 80 than when it’s less.

So in summary, more investigation would be warranted to see if an indicator value of less than twenty has meaningful predictive value over the next five days, but otherwise the Stochastic Oscillator doesn’t seem to have much predictive value.

 

This is an analysis of past performance, but past performance is not a guarantee of future performance.

Comments / Questions: joseph AT StockMarketMovement.com