With Simple Moving Averages, each data value in the "window" in which the calculation is performed is given an equal significance or weight. It is often the case, especially in financial price data analysis, that more chronologically recent data should carry a greater weight. In these cases, Weighted Moving Average (or Exponential Moving Average - see the following topic) functionality is often preferred.
Consider the same table of Sales data values for twelve months:
To calculate a Weighted Moving Average:
Calculate how many intervals of data are participating in the Moving Average calculation (i.e. the size of the calculation "window").
If the calculation window is said to be n, then the most recent data value in the window is multiplied by n, the next most recent multiplied by n-1, the value prior to that multiplied by n-2 and so on for all values in the Window.
Divide the sum of all of the multiplied values by the sum of the weights to give the Weighted Moving Average over that window.
Place the Weighted Moving Average value in a new column according to the trailing averages positioning described above.
To illustrate these steps, consider if a 3-month Weighted Moving Average of Sales in December is required (using the above table of Sales values).
The term "3-month" implies that the calculation "window" is 3, therefore the Weighted Moving Average calculation algorithm for this case should be:
{(Dec Sales value * 3) + (Nov Sales value * 2) + (Oct Sales value * 1)} divided by (3 + 2 + 1) = {(40 * 3) + (37 * 2) + (35 * 1)} / 6 = (120 + 74 + 35) / 6 = 229 / 6 = 38.17Or, if a 3-month Weighted Moving Average were evaluated over the entire original range of data, the results would be: