Before I continue, it is well to remember that past performance is no guarantee of the future. I have tweaked the settings on these indicators to fit certain past market behaviour, which is otherwise known as curve-fitting. The sample size (or at least the one I am working with) for market crashes is low, so it is unwise to make definitive statements about the best indicator or best setting. Even with the data for SPX , starting in 1980, there is no one perfect indicator or setting.
Moving Average Crossover
One of the most popular approaches for market timing - with good reason - is the moving average crossover. I've found that the exponential moving averages (EMA's) tend to work better than simple averages for this application, and my favoured combination is the 6 and 12-month exponential.
At present the blue 6-month of closing price is well above the red 12-month , indicating an uptrend. The crossings seen in the above chart are in early 2008 ( ) and mid to late 2009 ( ). During the bumpy market of 2010 & '11, the lines came very close to crossing over but did not cross. So the signal ever since 2009 has been to 'buy' or 'hold' stocks (at least those in the S&P 500 index ).
The surge in October (see the long tail highlighted in yellow) brought out quite a few "Chicken Littles"; predictions of 50% markdowns and the next market abounded. In truth the swoon was a little harrowing but just by focusing on these moving averages - and ignoring the noise - you would have sailed through relatively unscathed.
Taking the moving average concept one step further, we know that the oscillator is the difference between two EMA's, so we should be able to replicate the crossover system with this indicator. Indeed a crossover would correspond to the (light blue line) going negative, and as seen above its zero crossings coincide with the moving average crossovers on the main chart.
Now the advantage of the is that a signal line is provided, which is just a moving average of . As the moving average crossover can be a bit slow to happen after a market bottom, we can instead look for a signal line crossover. The signal crossover in mid 2009 provides an excellent example; it occurs a full 4 months before the crossover.
In this application I ignore the signal line when the is above zero; I don't take signal crossovers as sell signals and only consider the crossovers. I use the 6 and 12 month EMA's with a 14 month signal line; you could use a shorter period signal line for faster response but this generates some false signals.
I use (14) with a cross downward through 50 as the sell signal and a cross upward through 30 for the buy. The background colour on the indicator above covers only the 30 to 50 region. The reacts quickly to market movement which is both its strength and its weakness; it provides timely signals but they can be spurious (see orange circles).
After the bottom in 2009 the performed the best out of all these indicators for getting back into the market; it crossed above 30 after the second month of the uptrend. However during 2010 and '11 it oscillated around the 50 level, giving some unnecessary sell signals. In these cases I would take the recovery above 50 as a new buy signal. The question might also be raised, what if the doesn't reach as low as 30 during a market sell-off ? The safest answer would be to wait until it crosses 50 again, but judgement is called for; a low of 35 or less would probably suffice.
The S & P 500 has strong resistance at 2,094 points and the Dow Jones in 18,300 points, the indices can not reach these levels.
The reversal of the awards may begin soon in principo in January 2015.
'Blau Divergence RSI' (BDRSI)
This a somewhat experimental indicator (not yet published), inspired by the work of William Blau. I'm not really sure what to call it, nor if it has been used by Mr Blau or others already. It's an attempt to smooth out RSI by using double EMA of price change - the numerator of the True Strength Index - as substitute for price. Using 30 for both the buy and sell level, this indicator moves quite decisively at opportune times and tends not to give false signals. Like MACD, this indicator can also be improved with the addition of a signal line for buy decisions.
For this study, the long EMA is set to 36 periods while the short EMA is 10. The 7-period RSI is employed to produce the main (black) indicator line. The signal line (shown in red) is a 5-period EMA of the BDRSI; interpretation of signal crossovers is as described for the MACD. This indicator performs rather well over the 2007 to 2014 period, being similar in timeliness to the other two without coming close to giving a false signal.
This is a bit of a long ramble and I'd like to discuss other time periods as well as my findings. If you'd like to continue, please see my blog post over at http://blindfreddy.postagon.com/4symvwzun