Question

Table below shows monthly beer sales at Gordon’s Liquor Store in 2017.

The sales manager had predicted that January 2017 beer sales would be 950.

Month	Sales
Jan 2017	900
Feb	725
Mar	1000
Apr	800
May	750
Jun	1200
Jul	1000
Aug	1100
Sep	1250
Oct	1050
Nov	1400
Dec	1600
Jan 2018	??

• Forecast beer sales needed for January 2018 using the moving average with 3 periods.

• Forecast beer sales needed for January 2018 using an exponential smoothing method with α = 0.2 and α = 0.9

• What is the best method to forecast the beer sales according to MSE. (moving average with 3 periods, exponential smoothing method with α = 0.2, or exponential smoothing method with α = 0.9)

Answer 1

1)

forecasted beer sales with moving average with 3 period =(1050+1400+1600)/3=1350

2)

for exponential smoothing: next period forecast =previous period actual*alpha+(1-alpha)*previous period forecast

for alpha =0.2

month	value	forecast		error^2
1	900
2	725	900.00		30625.00
3	1000	865.00		18225.00
4	800	892.00		8464.00
5	750	873.60		15276.96
6	1200	848.88		123285.25
7	1000	919.10		6544.16
8	1100	935.28		27131.62
9	1250	968.23		79396.27
10	1050	1024.58		646.11
11	1400	1029.66		137148.01
12	1600	1103.73		246281.93
13		1202.99	total	693024.33
			average=MSE=	63002.21

forecast for Jan 2018 with alpha=0.2 is 1202.99

with alpha =0.9

month	value	forecast		error^2
1	900
2	725	900.00		30625.00
3	1000	742.50		66306.25
4	800	974.25		30363.06
5	750	817.43		4546.13
6	1200	756.74		196477.21
7	1000	1155.67		24234.47
8	1100	1015.57		7128.86
9	1250	1091.56		25104.27
10	1050	1234.16		33913.31
11	1400	1068.42		109948.24
12	1600	1366.84		54362.86
13		1576.68	total	583009.66
			average=MSE=	53000.88

forecast for Jan 2018 with alpha=0.9 is 1576.68

3)

as MSE is less with alpha= 0.9 , exponential smoothing method with α = 0.9 seems better.