1. Successful operations planning requires good forecasts.

2. Forecasting is imprecise, but the errors in prior forecasts are measurable.

3. There are qualitative and quantitative forecast systems.

1. The qualitative systems include expert or executive opinions, sales force composites, opinion surveys and the Delphi technique. The Delphi method includes a sequence of questionnaires administered to a select group of qualified experts; the design of each questionnaire is based upon the results of the previous questionnaire.
2. Quantitative forecasting methods include naïve forecasts, exponential smoothing, moving averages, and associative (regression-based) systems. A forecast system may be a combination of several of these.
3. The "naive" model is a simple special case: the value for the next period is predicted to be the same as it was in the previous period. For data that has cyclical or seasonal variations, the "last period" would be the previous corresponding period, such as "the forecast for this Friday is actual demand for last Friday." A trend version of a naive model is that the difference between the value for this period and the value for the next period will be the same as the difference between the last period and this one.

4. The accuracy of a forecast system depends upon:

1. accuracy of the historical time series data
2. similarity of patterns between the past and the future
3. grouping or aggregation of the data series
4. time lapse between the historical periods and the period for which the prediction is being made choice of a model.

5. Exponential smoothing is an adaptive forecasting technique with some advantages over other types of moving averages and other statistically based measures. These advantages include:

1. the calculations are simple.
2. the weighting pattern can be changed simply by changing the smoothing constant.
3. Both exponential smoothing and simple moving averages smooth the data and lag changes in a time series.

6. If there is trend in the historical data, single exponentially smoothed forecasts tend to lag behind the actual values. Therefore, it is necessary to incorporate trend adjustments, with double smoothing.

7. Associative techniques involve the use of predictor (independent) variables in equation form to estimate values of the variable of interest (dependent variable). Least squares analysis is used to obtain the coefficients of the regression equation.

8. Moving averages and trend lines can be used to compute monthly, weekly or daily indexes that show how one part of a "season" compares to the average value of a time series. These seasonal indexes are used in conjunction with trend calculations to generate predictions that take account of fluctuations in demand or economic activity within a period of a year.

9. This chapter shows how to monitor and control the accuracy of forecasts. The mean absolute deviation (MAD) is a measure of how far the actual values were from the predictions for previous periods, on the average. The tracking signal (TS) is a measure of the bias of the differences between the actual values and the predictions.

10. A forecast is deemed to be in control when forecast errors are judged to be random.