Linear regression is a statistical tool used to predict future values from past values.
In the case of security prices, it is commonly used as a quantitative way to determine the underlying trend and when prices are overextended.
A Linear Regression trend line uses the least squares method to plot a straight line through prices so as to minimize the distances between the prices and the resulting trend line
Interpretation:
If you had to guess what a particular security's price would be tomorrow, a logical guess would be “fairly close to today’s price.” If prices are trending up, a better guess might be “fairly close to today’s price with an upward bias.” Linear regression analysis is the statistical confirmation of these logical assumptions.
A Linear Regression trend line is simply a trend line drawn between two points using the least squares fit method. The trend line is displayed in the exact middle
of the prices. If you think of this trend line as the “equilibrium" price, any move above or below the trend line indicates overzealous buyers or sellers. A Linear Regression trend line shows where equilibrium exists.
Raff Regression Channels show the range prices can be expected to deviate from a Linear Regression trend line. Developed by Gilbert Raff, the regression channel is a line study the plots directly on the price chart. The Regression Channel provides a precise quantitative way to define a price trend and its boundaries. The Regression Channel is constructed by plotting two parallel, equidistant lines above and below a Linear Regression trend line.
The distance between the channel lines to the regression line is the greatest distance that any one high or low price is from the regression line.
Raff Regression Channels contain price movement, with the bottom channel line providing support and the top channel line providing resistance. Prices may extend outside of the channel for a short period of time. However, if prices remain outside the channel for a long period of time, a reversal in trend may be imminent.
Example:
