Linear Regression

Linear Regression
February 28, 2009
Definition:
Linear regression is a statistical tool used for forecasting future price. The concept behind linear regression is to find the best estimate of the trend given a noisy sample of data points. It is calculated by using the “Least Squares” method over a given period, which is drawn as a trendline extending through the defined period that attempts to filter out market noise.
You can add a Linear Regression Channel, which forms lines above and below the linear regression to help identify support and resistance.LR Channel can be added by clicking “Utilities” >> “Parameters” and selecting “LR Channel” in the resulting pop-up window. The default LR Channel is one standard deviation above and below the linear regression.
Interpretation:
There are two conventional interpretations for the linear regression line.
The first interpretation is to use the linear regression as the overall trendline for that given period. If the line is positive, it may suggest a buying opportunity, whereas a turn downwards suggests one may consider selling the stock. Price divergences below the line indicate a possible buying opportunity, for the market is oversold, while divergences above the line indicate the market is potentially overbought. Linear regression will work best when the period being studied is similar to the cycle length or typical trend length of the security in question.
A second interpretation is to construct a linear regression channel, consisting of two parallel lines at fixed distances above and below the linear regression line. These lines can be used as support and resistance lines, which are used to watch the battle between buyers and sellers.
Support and resistance lines are drawn as the upper and lower limits of a trading range, whereby the support line is the bottom line, and is the point at which “bulls” will not let the price fall below, and the resistance line is the top line, the point above which the “bears” will not let the price rise above.
Conventionally, a breakout above resistance or below support indicates that there is either a) some news about the company which justifies recreating the upper and lower trading limits or b) there is about to be a correction towards the range as trader\’s are hesitant about the stock\’s new value.
Using the Linear Regression Channel can assist in finding support and resistance levels from the Linear Regression.

Written by Larry Swing · Filed Under Indicators, Oscillators and Overlays