TOPIC TITLE

03 QUANTITATIVE METHOD IN FORECASTING

OBJECTIVES:
At the end of the topic session, the students should be able to:
▪ Learn the importance of forecasting in decision making
▪ Understand different quantitative techniques in forecasting
▪ Know trend analysis, exponential smoothing, decomposition methods and
casual method of forecasting
▪ Find out the suitability of forecasting models and calculate the errors in
forecasting

TOPIC PRESENTATION:
Forecasting is the art and science for predicting future events. Forecasting was
largely an art, but now it has now become science as well.

‘You can never plan the future by the past’ -Edmund Burke

‘I know of no way of judging the future but by the past’ -Patrick Henry

‘The quality of the forecast strongly related to the information that can be extracted from
past data’

0.1 FORECASTING

Forecasting is viewed as the key elements in the operations structure. It is an

essential tool in decision-making process. Forecasting may be short-term or long-
term by nature.

A Forecast is an estimate of a future event achieved by systematically combining

and casting forward in a predetermined way data about the past. Prediction on the
other hand, is an estimate of a future event achieved through subjective
considerations of managers.

Long-run planning decisions require consideration of many factors: general

economic conditions, industry trends, probable competitor’s actions, overall political
climate, and so on.

Forecasts are often classified according to time period and use. In general, short-
term (up to one year) forecasts guide current operations. Medium term (one-to-
three years) and long-term (over five years) forecasts support decisions on plant
location and capacity. The accuracy of the forecast and its costs are interrelated.
0.2 APPLICATION TO DIFFERENT FUNCTIONAL AREAS

Forecasting is one input to all types of business planning and control, both inside
and outside the operations function. The main focus is on forecasting on operations
function.

MARKETING To plan products, promotion and pricing.

FINANCE For managing cash flows and as input to financial planning.

HUMAN RESOURCE Need for recruiting

ACCOUNTANTS rely on forecast of costs and revenues for tax planning

It serves as an input for decision in process design purposes, forecasting is needed

to decide on the type of process and the degree of automation to be used.

0.2.1 FORECASTING IN OPERATIONS MANAGEMENT

Forecasting is an important component of operations planning. It is absolutely
necessary for planning, scheduling, and controlling the system to facilitate
effective and efficient output of goods and services.
The resource forecasts are used to plan and control operation subsystem as
shown below.

Figure 1: Using demand forecasting and production/operations subsystems

These operations sub-functions are planning the system, scheduling the system,
and controlling the system.

0.2.1.1 PLANNING THE SYSTEM

Managers need to forecast demands so that they can design or redesign
processes necessary to meet demand. Automated, continuous flows facilitate
high production volumes; manual or semi-automated, intermittent flows are
generally more economical for smaller production volumes.

0.2.1.2 SCHEDULING THE SYSTEM

Job scheduling in intermittent and continuous operations is more stable if
demand forecasts are accurate. Accurate demand forecasting is needed for
best utilization of the existing conversion system.

0.2.1.3 CONTROLLING THE SYSTEM

In regards to controlling inventory, production, labor, and overall costs,
managers need accurate demand forecast. Accurate forecasts are needed for
the immediate future-hours, days and weeks ahead. Thus, a computerized
forecasting system may be needed for these decisions.

There are different types of decisions in operations and different associated

forecasting requirements as shown table below:

Table 1: Forecasting uses and methods

Quantitative forecasting is used when:

▪ Past data is available, and

▪ Past data can be fitted into a pattern that can be expected to continue into
the future
0.3 SELECTING A SUITABLE FORECASTING METHOD

The most important factors for selecting a suitable qualitative, time-series and
casual methods of forecasting are indicated as follows:

▪ USER AND SYSTEM SOPHISTICATION

It has been found that the forecasting method must be matched to the
knowledge and sophistication of the user. The method chosen must not be too
advanced or sophisticated for its user or too far advanced beyond the current
forecasting system.

▪ TIME AND RESOURCE AVAILABLE

The selection of a forecasting method will depend on the time available in
which to collect the data and prepare the forecast. This may involve the time
of users, forecasters and data collectors.

▪ USE OR DECISION CHARACTERISTICS

The forecasting method must be related to the use or decision required. The
use is closely related to such characteristics as accuracy required, time
horizon of the forecast, and number of items to be forecast.

▪ DATA AVAILABILITY
The choice of forecasting method is often constrained by available data.

▪ DATA PATTERN
The pattern in the data will affect the type of forecasting method selected. If the
time-series is flat, a first method can be used. However, if the data show trends
or seasonal patterns, more advanced method will be needed.

0.4 SPECIFIC FORECASTING METHODS

The long-range strategic planning and facilities decisions use less analytical
qualitative methods of forecasting, and operational planning in production and
inventory control uses more analytical, time series analysis. Causal forecasting
techniques are used for a variety of planning situations but are specifically in
intermediate-term planning.

There are three types of forecasting techniques: qualitative forecasting, naïve (time-
series) forecasting and causal forecasting.

The most frequently used forecasting techniques in operation management are the
qualitative and naïve models. The casual models are very costly models to
implement and are not suitable for short-term forecasting.
 MODEL TYPE DESCRIPTION
1. QUALITATIVE MODELS

▪ DELPHI METHOD Questions panel of experts for opinions

▪ HISTORICAL DATA Makes analogies to the past is a judgmental manner
▪ NOMINAL GROUP TECHNIQUE Group process allowing participation with forced voting

2. NAÏVE (TIME-SERIES)
QUANTITATIVE METHODS

▪ SIMPLE AVERAGE Average past data to predict the future based on that
average
▪ EXPONENTIAL SMOOTHING Weights old forecasts and most recent demand

3. CAUSAL QUANTITATIVE
MODELS

▪ REGRESSION ANALYSIS Depicts a functional relationship among variables

▪ ECONOMIC MODELING Provides an overall forecast for a variable such as
gross national product

Table 1: Summary of representative forecasting techniques

0.5 MAIN CLASSES OF QUANTITATIVE MODELS

Some important forecasting techniques are explained below:

05.1 TIME SERIES MODELS

Time series models attempt to relate some quantity strictly to time. A time series
is a time ordered sequence of observations which have been taken at regular
intervals over a period-of time (hourly, daily, weekly, monthly, annually etc.)

05.1.1 TREND PROJECTION (LONG-TERM)

Reasons for studying trends are:

▪ Allows us to describe a historical pattern.
▪ Permits us to project past patterns, or trends, into the future
▪ Allows us to eliminate the trend component from the series

Trend can be a straight line or a curvilinear.

Example: A straight line trend is increase of

pollutants in the environment.
Common example of curvilinear relationship is the life cycle of a new product

i. When a new product is introduced, its

sale volume is low
ii. As the product gains recognition and
success, unit sales grow at an
increasing rate
iii. After the product is firmly established,
the unit sales grow at a stable rate
iv. Finally, as the product reaches the end
of its life cycle, unit sales begin to
decrease.

Trend reflects the effect of global movements in the time series. Consider the
linear trend.
̂ = a + bx
𝒚

̂ = estimated value of the dependent variable

Where: 𝒚
a = y-axis intercept
b = slope of the trend line
x = independent variable (time in trend analysis)

Objective is to find out the value of slope (b) and intercept (a) from the given
set of independent values of x.

*The general trend of many time-series using a straight line but when faced
with the problem of finding the best-fitting line, we can use the least-squares
methods to calculate the best-fitting line, or equation.

LEAST-SQUARES METHOD

Statisticians have developed equations that we can use to find the values of a and b
for any regression line.

( ∑ 𝒙𝒚) − 𝒏(𝒙
̅)(𝒚
̅)
b= a=𝒚
̅ – (b) 𝒙
̅
(∑ 𝒙𝟐 )−𝒏(𝒙
̅𝟐 )

where: b = slope of regression

x = known values of the independent variable
y = known values of the dependent variable
̅ = (“x-bar”) average of the x-values
𝒙
̅ = (“y-bar”) average of the y-values
𝒚
n = number of data points or observations
a = y-axis intercept

Express the trend line with the equation: ̂ = a + bx

𝒚
EXAMPLE PROBLEM: Time Period Sales Value
1 32
Sales data of a company over the past 14 time 2 28
periods (x) is shown in the following table with 3 30
respective sales value (y). Forecast the sales 4 34
value for the period 15 demand by fitting a 5 30
straight-line trend to these data. 6 43
7 36
8 42
9 42
10 55
11 47
12 56
13 54
14 57

SOLUTION:

STEP 1: Calculate the value of xy and x2 from the given sales value and its
summation.

Time Period Sales Value

Year xy x2
(x) (y)
1 1 32 32 1
2 2 28 56 4
3 3 30 90 9
4 4 34 136 16
5 5 30 150 25
6 6 43 258 36
7 7 36 252 49
8 8 42 336 64
9 9 42 378 81
10 10 55 550 100
11 11 47 517 121
12 12 56 672 144
13 13 54 702 169
14 14 57 798 196
n = 14 ∑ 𝒙 = 105 ∑ 𝒚 = 586 ∑ 𝒙𝒚 = 4927 𝟐
∑ 𝒙 = 1015

STEP 2: Calculate the 𝑥̅ , 𝑦̅ , b and a.

∑𝒙 𝟏𝟎𝟓 ∑𝒚 𝟓𝟖𝟔
̅=
𝒙 = = 7.5 ̅=
𝒚 = = 41.86
𝒏 𝟏𝟒 𝒏 𝟏𝟒

(∑ 𝑥𝑦) − 𝑛(𝑥̅ )(𝑦̅) 4927 − (14)(7.5)(41.86)

b= = = 2.34
(∑ 𝑥 2 )−𝑛(𝑥̅ 2 ) 1015 −(14)(7.52 )

a = 𝑦̅ – (b) 𝑥̅ = 41.86 – (2.34) (7.5) = 24.31

STEP 3: Calculate the trend line for period 15.

x = 15
̂ = a + bx = 24.31 + (2.34) (15) = 59.41
𝒚

∴ Forecast for period 15 is 59.41

STEP 4: Draw the graph.

05.1.2 DECOMPOSITION METHODS (INTERMEDIATE-TERM)

This method of time-series analysis is based on the belief that the demand can
be separately and distinctly broken down into its components, viz, trend, cycle;
seasonality and randomness. Decomposition method is the forecasting
required for intermediate range forecasting which helpful in production
planning. It may also use smoothing in their analysis.

Thus, demand usually is considered as a product of these components:

𝒀𝒕 = 𝑻𝒕 𝒙𝑺𝒕 𝒙𝑪𝒕 𝒙𝑹𝒕

where: 𝒀𝒕 = Demand observed

𝑻𝒕 = Trend component at period t
𝑺𝒕 = Seasonal index at period t
𝑪𝒕 = Cyclical factor at period t
𝑹𝒕 = Randomness as an index at period t

* 𝑻𝒕 is measure in units of 𝑌𝒕 but 𝑺𝒕 , 𝑪𝒕 and 𝑹𝒕 are all measured in relative

terms. Values of 𝑺𝒕 , 𝑪𝒕 and 𝑹𝒕 > 1 indicate effects above the trend (above the
average level)
Moving Average is an average that is repeatedly updated. The message given
by the moving averages technique is that while history helps to plan the future,
only a small fragment of the past is relevant.

EXAMPLE PROBLEM: The following table Year Quarter Patients

gives the number of patients in a family 1 920
practice for last four years. Each year is 2 916
1
divided into four quarters and the respective 3 895
patients being treated are included for 4 905
forecasting. 1 947
2 930
2
How do you make forecasts for the period 17 3 902
i.e. for the first quarter of 5 year, considering 4 940
the trend component, seasonal component, 1 996
cyclical component and random component? 2 952
3
3 923
4 960
1 1201
2 1142
4
3 1106
4 1163
SOLUTION:

*Number of digit/s after the decimal point must be based on the computation however,
maximum of 3 digits only and round-off.

STEP 1: Identify seasonal factors: For the actual series D, compute a moving
average (MA), whose length is equal to the seasonality (e.g. 12 in monthly data,
4 in quarterly data, 7 in daily data).

*Based on the problem, 4 in quarterly data. Time period t = 1, 2, 3, …. 16

4-period MA
At period 4 = (920 + 916 + 895 + 905) / 4 = 909
At period 5 = (916 + 895 + 905 + 947) / 4 = 915.75
(continue the computation up to …)
At period 16 = (1201 + 1142 + 1106 + 1163) / 4 = 1153

Note that the first moving average (= 909) is for the whole first year i.e. it is
associated with all the four quarters of the whole year. Therefore, it is
reasonable to center it. This is the (𝑻𝒕 𝑪𝒕 ) component. It can be between 2nd and
3rd quarter.

4th Quarter MA and 1st Quarter MA

= (909 + 915.75) / 2 = 912.375 (Centered at 3rd Quarter Year 1)

2nd Quarter MA and 3rd Quarter MA

= (915.75 + 919.25) / 2 = 917.50 (Centered at 3rd Quarter Year 1)
Here’s the calculated and tabulated values for reference.

Patients Centered Mas

Year Quarter (𝒀 𝒕 ) 4-MAs (𝑻𝒕 𝑪𝒕 )
1 920
2 916
1
3 895 912.375
4 905 909 917.50
1 947 915.75 920.125
2 930 919.25 925.375
2
3 902 921 935.875
4 940 929.75 944.75
1 996 942 950.125
2 952 947.5 952.25
3
3 923 952.75 983.375
4 960 957.75 1032.75
1 1201 1009 1079.375
2 1142 1056.5 1127.625
4
3 1106 1102.25
4 1163 1153

STEP 2: From 𝑻𝒕 𝑪𝒕 component (centered MA), identify the seasonal-random

component.

𝒀 Year Quarter 𝑺𝒕 𝑹𝒕
𝑺𝒕 𝑹𝒕 = 𝑻 𝑪𝒕
𝒕 𝒕
3 0.981
1
4 0.986
For Year 1:
1 1.029
2 1.005
Q3 => 895 / 912.375 = 0.981 2
3 0.964
Q4 => 905 / 917.50 = 0.981 4 0.995
…. and so on. 1 1.048
(see table for computed and tabulated values)
2 0.997
3
3 0.939
4 0.930
Note: Lose 4 points, 2 at the beginning
1 1.113
(Year 1: Q1 and Q2) and 2 at the end (Year 4: 4
2 1.013
Q3 and Q4)

STEP 3: Seasonal factor (𝑺𝒕 ) of each quarter can be obtained by taking

averages from each year.

1st Quarter Effect = (1.029 + 1.048 + 1.113) / 3 = 1.063

2nd Quarter Effect = (1.005 + 0.997 + 1.013) / 3 = 1.005
3rd Quarter Effect = (0.981 + 0.964 + 0.939) / 3 = 0.961
4th Quarter Effect = (0.986 + 0.995 + 0.930) / 3 = 0.970
STEP 4: Deseasonalize the series.

𝑻𝒕 𝑪𝒕 𝑹𝒕 = 𝑺𝒕
𝒀 Patients
𝒕 Year Quarter (𝒀𝒕 ) 𝑻𝒕 𝑪𝒕 𝑹𝒕

For Year 1: 1 920 865.157

2 916 911.443
1
Q1 => 920 / 1.063 = 865.157 3 895 931.322
Q2 => 916 / 1.005 = 909.641 4 905 932.990
Q3 => 895 / 0.961 = 931.204 1 947 890.875
Q4 => 905 / 0.970 = 932.700 2 930 925.373
2
3 902 938.606
…. and so on.
4 940 969.072
(see table for computed and tabulated 1 996 936.971
values) 2 952 947.264
3
3 923 960.458
4 960 989.691
1 1201 1129.821
2 1142 1136.318
4
3 1106 1150.884
4 1163 1198.969

STEP 5: Identify the trend in the deseasonalized series.

Time Period Patients

Year Quarter (ty) t2
(t) (y)
1 1 920 920 1
2 2 916 1832 4
1
3 3 895 2685 9
4 4 905 3620 16
1 5 947 4735 25
2 6 930 5580 36
2
3 7 902 6314 49
4 8 940 7520 64
1 9 996 8964 81
2 10 952 9520 100
3
3 11 923 10153 121
4 12 960 11520 144
1 13 1201 15613 169
2 14 1142 15988 196
4
3 15 1106 16590 225
4 16 1163 18608 256
n = 16 ∑ 𝒕 = 136 ∑ 𝒚 = 15798 ∑ 𝒕𝒚 = 140162 ∑ 𝒕𝟐 = 1496

∑𝒕 𝟏𝟑𝟔 ∑𝒚 𝟏𝟓𝟕𝟗𝟖
𝒕̅ = 𝒏 = 𝟏𝟔 = 8.5 ̅=
𝒚 = = 987.375
𝒏 𝟏𝟔
 (∑ 𝑡𝑦) − 𝑛(𝑡̅)(𝑦̅) 140162 − (16)(8.5)(987.375)
b1 = = = 17.291
(∑ 𝑡 2 )−𝑛(𝑡̅ 2 ) 1496 −(16)(8.52 )

b0 = 𝑦̅ – (b1) 𝑡̅ = 987.375 – (17.291) (8.5) = 840.402

Period 17: 𝑻𝒕 = b0 + b1 t = 840.402 + 17.291 (17) = 1134.350

𝒀
STEP 6: Obtain cyclical component 𝑪𝒕 𝑹𝒕 where 𝑺𝒕 is deseasonalized series.
𝒕

Before calculating 𝑪𝒕 𝑹𝒕 , you must calculate the values of 𝑻𝒕 i.e. the trend line
in the deseasonalized series for all the 16 periods.

𝑻𝒕 = b0 + b1 t 𝒀𝒕
𝑻𝒕
𝑺𝒕 𝑪𝒕 𝑹𝒕
T1 = 840.402 + 17.29 (1) = 857.691 865.157 857.691 1.009
T2 = 840.402 + 17.29 (2) = 874.982 911.443 874.982 1.042
…. and 931.322 892.274 1.044
T16 = 840.402 + 17.29 (16) = 909.565 932.990 909.565 1.026
890.875 857.691 1.039
925.373 874.982 1.058
𝒀 938.606 892.274 1.052
𝑪𝒕 𝑹𝒕 = 𝑺 𝑻𝒕 969.072 909.565 1.065
𝒕 𝒕
936.971 857.691 1.092
865.157
C1R1 = 857.691 = 1.009 947.264 874.982 1.083
960.458 892.274 1.076
989.691 909.565 1.088
911.443
C2R2 = 874.982 = 1.042 … and so on 1129.821 857.691 1.317
1136.318 874.982 1.299
1150.884 892.274 1.290
see table for computed and tabulated values 1198.969 909.565 1.318

Considering cyclical factor which is usual forecasted by using smoothing

techniques.

For Period 17 (Year 5: 1st Quarter)

It is required to forecast the values Tt for the period 17 (see step 5) and
St seasonal factor (see Step 3 ).

𝑻 𝟏𝟏𝟑𝟒.𝟑𝟓𝟎
C17 = 𝑺𝒕 = = 0.941 T17C17 = 1134.350 (0.941) = 1067.121
𝒕 𝟏.𝟎𝟔𝟑

*Seasonal factor for 1st Quarter = 1.063

FORECAST = 𝑻𝒕 𝑪𝒕 𝑺𝒕 = 1067.121 (1.063) = 1134.350

∴ Forecast for 1st Quarter of the 5th Year is 1134.350

05.1.3 SMOOTHING METHODS (SHORT-TERM)

It is suitable for short-term forecasting. It is generally averaged by weighting

that past data. The extreme case is ‘naïve’ forecasting. The weight decreases
in magnitude the further back in time the data are weighted, the decrease is
nonlinear.

05.1.3.1 EXPONENTIAL SMOOTHING

Exponential smoothing is an averaging method that exponentially

decreases the weighting of old demands.

They are popular because of the fact that

▪ Readily available in standard computer software packages
▪ Require relatively little data storage and computation

The limitations of exponential smoothing

▪ It is not suitable for data that include long-term upward or
downward movement (trend) in data
▪ The use of exponential smoothing would produce forecast that
were too low for upward trend and too high for downward trend
(i.e. either overestimated or underestimated)

EXAMPLE PROBLEM:
t 𝐘𝐭
Sales data for 10 periods are given.
1 700
Forecast the data for the next period i.e. for
2 724
the 11 period. 3 720
Assume a = 0.4 and Ft = 700 (initial value). 4 728
5 740
6 742
7 758
8 750
9 770
10 775

SOLUTION:

*Number of digit/s after the decimal point must be based on the computation
however, maximum of 2 digits only and round-off.

STEP 1: Calculate Ft (forecasting for the period t)

𝐅𝐭 = 𝐅𝐭−𝟏 + 𝐚 (𝐘𝐭−𝟏 − 𝐅𝐭−𝟏 )

 t 𝐘𝐭 𝐅𝐭
Assume first value of Ft is equal to 700
which same with first value of Yt 1 700 700
2 724 700
For next forecasting: 3 720 709.6
4 728 713.76
Ft-1 = 700 (just previous value)
Yt-1 = 700 (just previous value) 5 740 719.46
Ft = 700 6 742 727.67
a = 0.4 7 758 733.4
8 750 743.24
F2 = 700 + 0.4 (700-700)
9 770 745.95
F3 = 700 + 0.4 (724 – 700)
F4 = 709.6 + 0.4 (720 – 709.6) 10 775 755.57
F5 = 713.76 + 0.4 (728 – 713.76)
…. and so on (see the calculated and tabulated values)

F11 = 755.57 + 0.4 (775 – 755.57) = 763.34

∴ Forecast for period 11 is 763.34

05.2 CAUSAL MODELS

05.2.1 REGRESSION ANALYSIS

It is a causal forecasting model in which, from historical data, a functional

relationship is established between variables and the use to forecast dependent
variable values.

𝐅𝐭 = 𝐚 + 𝐛𝐗 𝐭

where: 𝐗 𝐭 = given the value of the variable X in period t

a = coefficient intercept
b = slope of the line

The coefficient a and b are computed by the following two equations:

𝒏 ( ∑ 𝑿𝒕 𝑫𝒕 ) −( ∑ 𝑿𝒕 )( ∑ 𝑫𝒕 ) ∑ 𝑫𝒕 −𝒃 ∑ 𝑿𝒕
b= 𝟐 a=
𝒏 (∑ 𝑿𝒕 )−(∑ 𝑿𝒕 )𝟐 𝒏

𝑫 = 𝐚 + 𝐛𝐗

where: D = dependent variable values

 EXAMPLE PROBLEM: A paper box company makes carryout pizza boxes. The
operations planning department knows that the pizza sales of a major client
are a function of the advertising pesos the client spends, an account of which
they can receive in advance of the expenditure. Operation planning is
interested in determining this relationship between the client’s advertising and
sales. The amount of pizza boxes the client will order, in peso volume, is known
to be a fixed percent of sales. Forecast for the period 11.

Quarter Advertising Sales

(100 000 php) (1 000 000 php)
QUARTERLY ADVERTISING
1 4 1
AND SALES 2 10 4
3 15 5
4 12 4
5 8 3
6 16 4
7 5 2
8 7 1
9 9 4
10 10 2

SOLUTION:

STEP 1: Calculate the value of a and b where advertising is Xt.

Quarter Advertising Sales

(t) (𝑿𝒕 ) (𝑫𝒕 ) 𝑿𝒕 ∗ 𝑿𝒕 𝑿𝒕 ∗ 𝑫𝒕

1 1 32 32 1
2 2 28 56 4
3 3 30 90 9
4 4 34 136 16
5 5 30 150 25
6 6 43 258 36
7 7 36 252 49
8 8 42 336 64
9 9 42 378 81
10 10 55 550 100
t = 55 ∑ 𝑿𝒕 = 96 ∑ 𝑫𝒕 = 30 ∑ 𝑿𝒕 𝟐 = 1060 ∑ 𝑿𝒕 𝑫𝒕 = 328

𝒏 ( ∑ 𝑿𝒕 𝑫𝒕 ) −( ∑ 𝑿𝒕 )( ∑ 𝑫𝒕 ) 𝟏𝟎 (𝟑𝟐𝟖)−𝟗𝟔(𝟑𝟎) ∑ 𝑫𝒕 −𝒃 ∑ 𝑿𝒕 𝟑𝟎−𝟎.𝟐𝟗 (𝟗𝟔)

b= = = 0.29 a= = = 0.22
𝒏 (∑ 𝑿𝒕 𝟐 )−(∑ 𝑿𝒕 )𝟐 𝟏𝟎 (𝟏𝟎𝟔𝟎)− 𝟗𝟔𝟐 𝒏 𝟏𝟎

Thus, the estimated regression line, the relationship between future sales Ft and
advertising Xt is:

Forecast Period 11: Xt = 11 𝐅𝐭 = 𝐚 + 𝐛𝐗 𝐭 = 0.22 + 0.29 (11) = 3.41

So, sales are forecast as 3, 410, 000 php

0.6 FORECAST ERROR

Whether it is a simple smoothing or more advanced smoothing, an estimate of

forecast error should be computed along with the smoothed average. This error
estimate might be used for several purposes.

▪ To set safety stocks or safety capacity and thereby ensure a desired level of
protection against stockout.
▪ To monitor erratic demand observations.
▪ To determine when the forecasting method is no longer tracking actual
demand and needs to be reset.

In computerized forecasting systems it is extremely important to incorporate error

controls. This will ensure that the system does not run out of control.