Trend Extrapolation

Trend extrapolation is the simplest form of foresighting (Millet, 1991). This method is based on an assumption that patterns in the past will continue into the future. To perform this method, information is collected about a variable over time, and then extrapolated to some point in the future. This analysis can be either qualitative or quantitative. In the most simple form, trend extrapolation can be based on linear or other straightforward projections. Other foresighting methods that are more elegant variations of simple trend extrapolation (and that can be considered a subset of modeling) include system dynamics, s-curves, regression analysis, and substitution analysis. All of these methods hold a common assumption that the future will follow some pattern based on the past.

This method has two major weaknesses. First, it is often a fallacy to assume that the future will follow the pattern of the past. While people often make such assumptions due to a lack of better information, any picture of the future that is developed on this basis can be inaccurate. The second weakness of this method is that it typically provides information on only a single variable. Especially in current world conditions, it is rare for any variable to act independently. More often, the influence of outside forces can dramatically alter the future of any one event or condition.

Reference:
http://www.blogger.com/www.seattle.battelle.org/Services/ES/foresite/ch02.htm