Question

Problem 6 - Based on the data provided in the table below is this new forecasting technique biased? If so how is it biased? Calculate tracking signal values. Is there any evidence this New Forecast should not be used?

	Actual	New
Period	Demand	Forecast
1	650	610
2	883	932
3	950	917
4	1025	988
5	680	853
6	895	788
7	988	942
8	1178	1083
9	725	952
10	915	820
11	1068	992
12	1256	1162

Answer 1

Period	Actual Demand	New Forecast	Forecast Error = Actual - Forecast	Running Sum of Forecast error (Keep adding forecast errors)	Square of error	Absolute deviation
1	650	610	40	40	1600	40
2	883	932	-49	-9	2401	49
3	950	917	33	24	1089	33
4	1025	988	37	61	1369	37
5	680	853	-173	-112	29929	173
6	895	788	107	-5	11449	107
7	988	942	46	41	2116	46
8	1178	1083	95	136	9025	95
9	725	952	-227	-91	51529	227
10	915	820	95	4	9025	95
11	1068	992	76	80	5776	76
12	1256	1162	94	174	8836	94
		# of Observations	12	Total	134144	1072
				MSE = 134144/12	11179
				MAD = 1072/12	89.33
				Tracking Signal = RSFE / MAD = 174/89.33	1.95	Underestimating the demand

MSE = SUm of Mean square errors / # of Observations

MAD = Sum of absolute deviation / # of Observations

Tracking Signal =RSFE / MAD = 1.95. A positive tracking signal indicates that the forecast has a bias and is underestimating the demand consistently i.e. demand is greater than forecast. Hence, it is not a good forecasting method.

Further, the Mean square error is quite high and the forecasting technique is not good.