Intermittent Demand Forecasting

9Accuracy and Its Implications

9.1 Introduction

So far in this book, we have seen that there are a number of choices that can be made regarding inventory policies, demand distributions, and forecasting methods. All of these choices will have an effect on the overall performance of the inventory system.

When choosing forecasting methods, we want to know which is ‘best’. To answer this question, we must first pose another, namely, ‘What do we mean by best?’ One answer seems obvious: the best method is the most accurate method. However, forecasting is not an end in itself; it is a means towards the goal of better organisational performance. So, there is a need to assess not only the accuracy of a forecasting method but also its performance implications.

For inventory management, we need to assess the effects of forecasting on inventory holdings and service, for a given inventory policy and demand distribution. The measures of these effects are examples of accuracy‐implication metrics (Boylan and Syntetos 2006).

In short, forecasting methods may be evaluated using:

1. Accuracy measures, focusing on forecasting performance. Accuracy may be measured in terms of the ability to predict mean demand, or to predict the whole demand distribution.
2. Accuracy‐implication measures, focusing on organisational performance.

The choice of forecasting method is not the only factor affecting these measures. The accuracy of demand prediction also depends on the choice of distribution. Organisational ...