by Thomas Aspray
In the fall of 1986, I completed work on a new indicator: the MACD Histogram/Momentum. As many
of you are already aware, the MACD, or Convergence/Divergence, is an excellent indicator, but its
signals often lag when run on weekly data. Convergence/Divergence was developed by Gerald Appel,
who has created many excellent technical tools. I first presented my work on the MACD at a CompuTrac
conference in 1984. At that time, very few analysts were aware of it, much less using it on the commodity
markets. Now it is a widely used technical tool.
I have done some work on optimizing the MACD by testing various combinations of inputs for the three
exponential moving averages. This has been discussed in previous educational articles, and, by altering
the moving average variables, the results are dramatically different.
For example, the default for silver ( 12 days, 26 days and 9-period exponential average of the difference)
were not always profitable, while the (10, 20, 9) combination was profitable 54% of the time, with an
average profit of 67.5 points and an average loss of only 34 points.
After studying and using MACD for almost five years, I felt that, besides optimization, a method of
anticipating crossovers would be to my distinct advantage. After looking at (and testing) several
alternatives, I found that by running a 10-day momentum (with 3-day smoothing) of the MACD in
histogram form (MACD-H), the results were quite good. For the MACD-H in these weekly charts I will
use 10 and 20 days for the averages with a 10-period exponential average of the difference between the