Chapter 20

Method III — Reversals

Somewhere in the early 1970s the idea of shifting a moving average up and down by a fixed percentage to form an envelope around the price structure caught on. All you had to do was multiply the average by one plus the desired percent to get the upper band or divide by one plus the desired percent to get the lower band, which was a computationally easy idea at a time when computation was either time consuming or costly. This was the day of columnar pads, adding machines and pencils, and for the lucky, mechanical calculators.

Naturally market timers and stock pickers quickly took up the idea as it gave them access to definitions of high and low they could use in their timing operations. Oscillators were very much in vogue at the time and this lead to a number of systems comparing the action of price within percent bands to oscillator action. Perhaps the best known at the time--and still widely used today--was a system that compared the action of the Dow Jones Industrial Average within bands created by shifting its 21-day moving average up and down four percent to one of two oscillators based on broad market trading statistics. The first was a 21-day sum of advancing minus declining issues on the NYSE. The second, also from the NYSE, was a 21-day sum of up-volume minus down-volume. Tags of the upper band accompanied by negative oscillator readings from either oscillator were taken as sell signals. Buy signals were generated by tags of the lower band accompanied by positive oscillator readings from either oscillator. Coincident readings from both oscillators served to increase confidence. For stocks for which broad market data wasn't available, a volume indicator such as a 21-day version Bostian's Intraday Intensity was used. This approach and a myriad of variants remain in use today as useful timing guides.

Many modifications to this approach are possible and many have been made. My own contribution was to substitute a departure graph for the 21-day summing technique used for the oscillators. A departure graph is a graph of the difference of two averages, a short-term average and a long-term average. In this case the averages are of daily advances minus declines and daily up-volume minus down-volume and the periods to use for the averages are 21 and 100. The plot is of the short-term average minus the long-term average. 

The prime benefit of using the departure technique to create the oscillators is that the use of the long-term moving average has the effect of adjusting (normalizing) for long-term biases in market structure. Without this adjustment a simple Advance-Decline oscillator or Up Volume-Down Volume oscillator will likely fool you from time to time. However, using the difference between averages very nicely adjusts for the bullish or bearish biases that cause the problem.

Choosing the departure technique also means that you can use the widely available MACD calculation to create the oscillators. Set the first MACD parameter to 21, the second to 100 and the third to nine. This sets the period for the short-term average to 21 days, the period for the long-term average to 100 days and leaves the period for the signal line at the default, nine days. The data inputs are advances-declines and up volume-down volume. If the program you are using wants the inputs in percents, the first should be 9%, the second 2% and the third 18%. Now substitute Bollinger Bands for the percentage bands and you have the core of a very useful reversal system for timing markets.

In a similar vein we can use indicators to clarify tops and bottoms and confirm reversals in trend. To wit, if we form a W2 bottom with %b higher on the retest than on the initial low--a relative W4--check your volume oscillator, either MFI or VWMACD, to see if it has a similar pattern. If it does, then buy the first strong up day; if it doesn't, wait and look for another setup.

The logic at tops is similar, but we need to be more patient. As is typical, the top takes longer and usually presents the classic three or more pushes to a high. In a classic formation, %b will be lower on each push as will a volume indicator such as Accumulation Distribution. After such a pattern develops look at selling meaningful down days where volume and range are greater than average.

What we are doing in Method III is clarifying tops and bottoms by involving an independent variable, volume in our analysis via the use of volume indicators to help get a better picture of the shifting nature of supply and demand. Is demand increasing across a W bottom? If so, we ought to be interested in buying. Is supply increasing each time we make a new push to a high? If so, we ought to be marshalling our defenses or thinking about shorting if so inclined.

The bottom line here is clarification of patterns that are otherwise interesting, but on which you might not have the confidence to act without corroboration.

• Buy setup: lower band tag and the oscillator positive
• Sell setup: upper band tag and oscillator negative
• Use MACD to calculate the breath indicators