Smoothing: How You Profit from Your Mistakes
IN THIS CHAPTER
Familiarizing yourself with smoothing
Working with the Exponential Smoothing tool
Deciding on a smoothing constant
Dealing with smoothing problems
Smoothing, in this case exponential smoothing, is a kind of modified moving average. Don’t let the name put you off. You won’t have to deal with any exponents (or, for that matter, proponents, deponents, opponents, or components). It’s a simple idea, tricked out — as so many simple ideas are — with a fancy name. The idea is to correct a prior forecast and use that correction in making the next forecast. That’s pretty straightforward, no?
The Data Analysis add-in does exponential smoothing on your behalf. The Exponential Smoothing tool is one of the tools in the Data Analysis add-in that returns a formula, so if your input data changes, the forecasts will update automatically. But you want a bit more control over the formulas than the Exponential Smoothing tool gives you, and this chapter shows you how to get that control.
Exponential smoothing uses something called — here’s another bit of jargon — a smoothing constant. It helps determine the amount of the error in an earlier forecast to use in making the next one.
After you’ve chosen a smoothing constant, even just for the moment, you’ve automatically chosen a damping factor — yet another piece of jargon. I wouldn’t even bring it up here except that the Data Analysis add-in’s Exponential Smoothing tool uses that ...