The Simple Moving Average
Time series data such a stock prices, weather observations or experimental results often contain random fluctuations which are often referred to as "noise". Noise can make it difficult to infer usefull things form the data stream such a trend and cycles. The simplest way of reducing the effect of noise is to use an historical average of a set number of observations.
Often moving averages are described by the number of observations used in the calculations. for example, MA20 is an average based on the 20 most recent observations, MA50 and MA100 use fifty and one hundred observations respectively.
Simple Example Calculation
The last ten observations from some experiment are shown in the series below:
420,390,423,414,410,383,382,419,399,390
On a graph, they look like this
The arithmetic for the most recent value of MA5 is shown below:
(383,382,419,399,390)/5 = 1973/5 = 395 (rounded up to nearest integer)
The calculations for MA20, MA50 and MA100 are similar, they just use more historic data.
The Simple Moving Average is a Lagging Indicator
Because it uses historical data, the simple moving average is described as a lagging indicator.
Related Topics