|
PureBytes Links
Trading Reference Links
|
Below is a copy of the original web-page!
D:\Equis\Z-Divers\faq-MetaStock\TASC\errata.htm
Regards,
Ton Maas
Ms-IRB@xxxxxxxxx
----------------------------------- From TASC mag - its Errata page ---------------------------------------------
APRIL 1997
The following correction was sent in by Mark Vakkur, author of the April 1997 article "The moving average
convergence/divergence histogram."
While producing the April issue, we noticed too late the discrepancies in Figures 1 and 4 between the author's original
hard-copy charts and the files he created for the article, because the two indicators (the MACD and the MACD histogram)
are so similar. We apologize for any inconvenience caused by the error. -- Editor
Mark Vakkur writes:
Contrary to how it was named in the article, the indicator I presented in my April 1997 article, "The moving average
convergence/divergence histogram," was actually a variant of the MACD indicator itself, not of the MACD histogram, which
is an indicator described by Alexander Elder in Trading for a Living.
In my article, I used SuperCharts to create a user function, MACDHis, defined as follows:
MACDHis = MACD(close, shortma, longma)/
MaxList(close,1)*100
In plain language, this is simply the difference between two moving averages (that is, the MACD indicator) of lengths
shortma and longma, divided by the close (I used the MaxList function to ensure against divide-by-zero errors). I
multiplied this result by 100 to create a more readable number, then plotted the result as a histogram.
However, to clarify matters, this is not the MACD histogram as described by Elder in Trading for a Living, which would
be defined as the MACD minus its signal line, plotted as a histogram. The SuperCharts formula for that indicator would
be:
MACDHis2 = MACD(close, shortma, longma)-xaverage
(MACD(close, shortma, longma), thirdma)
where thirdma is the length of the signal line (for example, 6 in a 12/26/6 bar system). This could then be divided by
the close and multiplied by a large constant to normalize the result. At any rate, as discussed in the article, the
value of the indicator is unimportant; what matters is its trend.
The rest of the article, including the systems presented and their results, is essentially correct once you substitute
"MACD indicator" for "MACD histogram" throughout. Figures 1 and 4 shown here have also been modified to reflect this
correction.
The system I used is as follows:
Long entry:
Buy stop high[1] if
MACDHis(SHORTMA,LONGMA)MACDHis(SHORTMA,
LONGMA)[1]
Long exit:
Sell stop low[0] if
MACDHis(SHORTMA,LONGMA)<MACDHis(SHORTMA,
LONGMA)[1]
For testing purposes, I disabled any short selling and used no stops.
I apologize for this error and hope it did not cause too much frustration for readers. I appreciate the feedback from
readers and from the editorial staff of STOCKS & COMMODITIES, which led to my discovery of the error (mainly through
their failure to replicate my trading system results). Moreover, the comments I received from readers demonstrate that
readers are taking to heart the magazine's frequent advice to accept nothing at face value but to test everything before
risking a dime. Thanks for the feedback.
Mark Vakkur, M.D.
vakkur.mark@xxxxxxxxxxxxxxxxxx
-----------------------------------------end of page--------------------------------------------
|