GMA
Author: karla2010
Creation Date: 2/4/2012 7:57 AM
profile picture

karla2010

#1
The geometric moving average seems to have trouble with longer windows.

It produces infinity for windows sizes around the 200+ size.

My guess is that you are not using the log formulation. I wrote my own version using the log version and it works fine.

GMA has the flavor of a factorial function and best handled with logs.

Please take a look. Thanks.
profile picture

Eugene

#2
GMA is a part of Community Indicators that is a community driven project. If you don't like something about it or some other class, then I'd suggest the following. Download the solution, find the GMA class, review (it's open source), then please feel free to submit your own version. Preferably with a documentation/prooflink on the web. Thank you in advance.
profile picture

thodder

#3
Could we have some documentation on the GMA indicator? I don't see it on the wiki site or "More info" within WLP in the Technical Indicators dialog.

According to the book "Technical Analysis" by Kirkpatrick and Dahlquist (copyright Oct 2015; page 316):

QUOTE:
Geometric Moving Average (GMA)
The geometric moving average (GMA) is used mostly in indexes. It is a simple moving
average of the percent changes between the previous bar and the current bar over some past
predetermined period. Using percentages rather than points does not change its range or
dimensions like a price-based moving average. However, it still has all the other problems of
equal weight and lag.


The description above does not seem to match this indicator. I had to test it first before using it as there was no description in WLP.

Fyi, I'm still waiting for access to the code so I can review it myself.
profile picture

Eugene

#4
The GMA here is a component of GAMDO by David Varadi. The formula used is non-standard i.e. Math.Pow(CumulativeReturn, period). Check out this source thread: Geometric and Arithmetic Mean Divergence Oscillator (GAMDO)
This website uses cookies to improve your experience. We'll assume you're ok with that, but you can opt-out if you wish (Read more).