NEW TECHNIQUES
But Not The Met
Visiting Moma
by Stephan Bisse
Give your moving averages a head start with move-adjusted moving
averages.
The most popular of all technical indicators
may be the moving average. The backbone of many trading systems, moving
averages are great for filtering out the noise from time series and showing
the underlying trend of where the time series has been.
By definition, however, moving averages on their own can never say anything
about the future direction of a time series, regardless of the lookback
period used or any standard adjustments or weightings applied to the datapoints.
The only way that moving averages can give any indication of the coming
direction of a time series is if additional information with some predictive
power -- in other words, a leading indicator for the time series in question
-- is incorporated into the calculation.
TYPES OF MOVING AVERAGES
A simple moving average (SMA) is calculated by taking the sum of a series
of datapoints and dividing it by the number of points in the series. Each
subsequent value of the SMA is calculated by including the new datapoint
in the calculation and dropping the oldest to keep the number of datapoints
in the calculation constant. In an SMA, all the datapoints are given equal
weight.
In addition to the SMA, numerous variations of moving averages exist
based on giving unequal weights to the datapoints, usually with the idea
that the newer data has more relevance to future price action; for
these variations, the most recent points are weighted more heavily than
older ones.
The weighted moving average (WMA) and the exponential moving average
(EMA) are the most common examples of unequally weighted moving averages.
A WMA weights the datapoints in a linear manner according to the order
in which they occur. For example, in a 10-period WMA the most recent datapoint
is given 10 times the weight of the oldest datapoint.
The EMA, on the other hand, weights the datapoints using an exponent
that determines the rate at which the weight of previous datapoints diminishes.
Neither of these methods introduces any additional information into the
calculation, however, and as a consequence, they are useful only for following
where the time series has already been, albeit with the advantage of cutting
down the lag inherent in an SMA.
FIGURE 1: 10-PERIOD SMA AND 10-PERIOD MOMA. Since they are
applied to a zigzag, both SMA and MOMA exhibit similar behavior. This is
because the change between datapoints during reversals in the direction
of the zigzag remains constant, giving equal weighting to all datapoints.
VARIATIONS OF MOVING AVERAGES
The two most famous variations on moving averages that introduce additional
information into the calculations are the volume-adjusted moving average
(VAMA), developed by Richard Arms of Arms index fame, and the volatility-index
dynamic average (VIDYA), developed by Tushar Chande.
As its name suggests, VAMA weights the datapoints in the lookback period
based on their corresponding relative volume or tick volume. VIDYA, on
the other hand, uses a volatility index instead of volume for weighting
the datapoints. If this additional information is a leading indicator for
future changes in the time series, then a moving average adjusted in this
manner can, in theory, have predictive power and go beyond simply showing
where a time series has been.
MOVE-ADJUSTED MOVING AVERAGE
Starting with the basics, one of the easiest pieces of additional information
that can be introduced into a moving average calculation is the magnitude
of the change leading up to each datapoint. Applied to financial market
time series, the assumption implicit in such a constructed moving average
is that large moves in a given direction are leading indicators of future
moves in the same direction.
...Continued in the February issue of Technical Analysis
of STOCKS & COMMODITIES
Excerpted from an article originally published in the February 2005
issue of Technical Analysis of STOCKS & COMMODITIES magazine. All rights
reserved. © Copyright 2005, Technical Analysis, Inc.
Return to February 2005 Contents
