Dialogue with Van about Market Type
Van K. Tharp
the last few weeks, I�ve been talking to you about market type.
Those articles (see
back issues) have stimulated a series of questions from readers
along with some confusion due to a mistake that I made.
My goal in this article is to revisit this topic by answering
some of the questions generated, and to correct that mistake which was in how I
explained the data calculation. The data
itself was correct.
Van, are there not a lot of ways that you could calculate market
Yes, there are as many ways as there are systems. Your way of calculating market type should depend upon how
you trade. My original
goal was to visually look at markets by quarter and then define each
quarter with respect to how a long- term position trader might
experience them. For
example, if you visually inspected each quarter it should be obvious
whether it was up or down. And
if you could not tell, then the market was sideways.
I wanted to be a little more flexible when I reported it in the
monthly update and basically did rolling 13 week windows (quarters).
In addition, I wanted things to be automatically calculated.
a result, I elected to do the following.
I looked at the average gain (from 30 years of data) for the
absolute value of the 13 week change.
That average amounted to 5.646%.
Then on a weekly basis, I asked the question,
�Is the absolute value of the 13 week change greater than 5.646% (or
whatever the number is since it is updated weekly but doesn�t
change much since we have 30 years of data).
If the answer is no, then I consider the last 13 weeks to be
sideways. If the answer
is yes, then I determine whether the real 13 week change is greater
than +5.646%. If it is,
then the 13 weeks are called bullish.
Otherwise, the real 13 week change is going to be a drop of
greater than -5.646% and I�d call the 13 weeks bearish.
That�s how I classify the market on a rolling 13 week
also have a daily market classification, where I look at the 5- day
change (i.e., weekly) in the same manner to classify the short term
market as up, down, or sideways.
The absolute weekly change over 30 years is about 2.3%.
But I�m not reporting that in Tharp�s Thoughts because my
update is monthly, not weekly.
So how do you classify whether the market is quiet or
This is where there has been a little controversy.
I basically look at absolute value of the weekly change and
compare the absolute value of the weekly change with the historical
absolute value which is 2.3%. If
it is above the historical norm then it is volatile and if it is
below, then it is quiet.
When I calculated the weekly change it was not the same as
yours. What gives?
There was the mistake in the way I reported it.
I actually don�t calculate the weekly change, I calculate
the average absolute value weekly change over the last 7 weeks.
It�s sort of a moving average.
I do that for several reasons.
I�m not interested in market classifications that change
violently from week to week, so using a moving average makes the
changes a little slower. In
addition, there could be some weeks where the market is volatile but
the weekly change is not. That�s
not likely to happen often, and by using a rolling 7 week window,
the market would have to do that many times to mask the data.
That just doesn�t happen, so I feel pretty accurate about
what we are getting.
Still the market could be all over the place on a daily basis
with huge changes of 3
to 4% each day, but the weekly change could be
zero. If that happened,
we�d have a volatile market but your system would say it was
quiet. Shouldn�t you
use the ATR instead?
I said, I think the rolling average of the last seven weeks covers
us there. But what you
say is still possible. However,
to do ATRs I�d have to gather a lot more information than just the
close and that would make a huge database that would take a lot
longer to process. As a
result, I prefer this method.
do plan to look at ATRs and see how that changes our results.
I plan to publish a special report that will include 1) long
term market classifications with both the rolling 7 week windows and
also the ATR. I�m also
going to publish the daily changes in that special report so that
short term traders can see what�s going on in terms of market
example, the market right now is much quieter on a daily basis
classification than it is on a weekly basis classification.
And day traders would need to know that.
Are there not a lot of ways to do market classification?
Yes, and what I do might not be right for you.
For example, Ken Long publishes a market classification and
he has 9 categories � up, sideways, and down vs quiet, normal, and
volatility, he looks at the ATR and considers the mean +/- one
standard deviation to be normal.
More than one standard deviation is volatile and less than
one standard deviation is considered quiet.
And with his classification, by definition, about 2/3rds of
his days will have a normal volatility.
addition, Ken only looks at a look- back period of 100 days.
Thus, his statistics are only relative to the last 100 days.
Thus, in a very volatile environment, he might suddenly
classify the market as quiet, simply because it is quiet compared to
the norm of the last 100 days.
actually uses his market classification as a screening tool to know
what to expect from the market and what systems he should be using.
I�ve developed my classification so that longer term
traders can know how their systems work under various market
I could not match up your dates with the weekly close in the market
Answer: I think
we will be able to fix that in the future, but it is an XLQ problem.
XLQ gives the Monday date and then the weekly close on Friday
when the week is over. And
if the week is still going, it gives the Monday date and the last
closing data as the weekly close.
In the future, we�ll make adjustments to fix that.
About Van Tharp: Trading
coach, and author, Dr. Van K. Tharp is widely recognized for his
best-selling book Trade Your Way to Financial Freedom and his
outstanding Peak Performance Home Study program - a highly regarded
classic that is suitable for all levels of traders and investors.
You can learn more about Van Tharp at www.iitm.com.