Question

QUANTITATIVE – You work for Lynn, Inc. Your boss has asked you to complete a forecast for February and March using two forecasting methods and then to evaluate the two methods to identify the “best” method. You decide to try a weighted moving average of the last two months using weights of 60% (previous month) and 40% (2 months prior). You also try exponential smoothing with an alpha constant of .3Y. Complete the table by providing numbers for A, B, C, D, E, F, G, and H and calculate the MAD for each method (I and J). Which is a better forecasting method for this data? Why?

Month	Actual	Forecast – Weighted Average	Absolute Deviation – Weighted Average	Forecast – Exponential Smoothing (alpha = .3Y)	Absolute Deviation – Exponential Smoothing
November	98	Do not enter	Do not enter	Do not enter	Do not enter
December	101	Do not enter	Do not enter	Do not enter	Do not enter
January	10X	Do not enter	Do not enter	102	Do not enter
February	9X	A	B	C	D
March	105	E	F	G	H

MAD for Weighted Average = I

MAD for Exponential Smoothing = J

Answer 1

For weighted average method, the forecast is calculated as the sum of product of weights and the previous 2 months actuals. Hence we have the below

Forecast for February = 0.6*106+0.4*101 = 104.

Forecast for March = 0.6*96+0.4*106 = 100

Absolute deviation is given using the following formula - |Actuals - Forecast|. Hence we have the below

Absolute deviation for Febraury = |96-104| = 8

Absolute deviation for March = |105-100| = 5

Exponential smoothing forecast is given as below

Forecast = Previous Forecast + Alpha*(Previous Actuals - Previous Forecast).

Hence we have the below

Forecast for February = 102 + 0.3*(106-102) = 102+1.2 = 103.2

Forecast for March = 103.2+0.3*(96-103.2) = 103.2-2.16 = 101.04

Absolute deviation is given using the following formula - |Actuals - Forecast|. Hence we have the below

Absolute deviation for Febraury = |96-103.2| = 7.2

Absolute deviation for Macrh = |105-101.04| = 3.96

The complete table is as below

MAD is calculated as the average of absolute deviation. Hence we have the below

MAD for weighted average = (8+5)/2 = 6.5

MAD for exponential smoothing = (7.2+3.96)/2 = 5.58

We know that the method with lower MAD is preferred. Hence Exponential Smoothing is better as MAD is lower for this method.