Smoothing is a way to get your forecast to respond quickly to events that occur during the baseline period. Regression approaches such as TREND and GROWTH apply the same formula to all the forecast points, and getting a quick response to a shift in the level of the baseline becomes quite complex. Smoothing is a useful way around this problem.
The fundamental idea behind the smoothing approach is that each new forecast is obtained in part by moving the prior forecast in a direction that would have improved the old forecast. Here's the basic equation:
t is the time period, ...