ALMA
Arnaud Legoux Moving Average
November 24, 2009
Copyright: Arnaud Legoux / Dimitrios Kouzis Loukas
ALMA; In search for the perfect Moving Average - Created by Arnaud
Legoux & Dimitris Kouzis-Loukas
Every trading system based on technical analysis uses Moving Averages. Usually Moving Averages/
crosses are considered to be entry or exit signals. Furthermore Moving Averages are often responsible
for distinguishing random price movements (jitter) from the real trend. This means that a significant
amount of your profits depends on the quality of the Moving Averages you use for your trading system.
One of the benefits of Moving Averages is that they have quite standard form which allows us to easily
switch from one to another, compare them and choose the best for our trading system.
There are two key features we are looking for on an excellent Moving Average named smoothness and
responsiveness. Smoothness is important because it allows us to take decisions according to true
trends instead of random noise. Responsiveness on the other hand is important in order to take
decisions timely. Deciding that a trend is true with big latency wastes precious profit PIPs. Now every
Moving Average is a Discrete Time Filter and as such it is ruled by the Uncertainty Principle which
means that smoothness and responsiveness are conflicting requirements. Everybody who has basic
experience with any moving average e.g. SMA has knows firsthand that a 9-day SMA is much more
responsive and much less smooth than a 21-day SMA. It looks like we can/t have both at the same time
and that holds true if we restrict ourselves in a single type of Moving Average. On the other hand
different types of Moving Averages have different performance when it comes to smoothness and
responsiveness.
Table 1. Formulas and Kernels of different Moving Averages