The following is a brief review of the Regression studies looked at during an intensive session over the course of 4 days.
Linear Regression
A new channeling method was added, called Raff Channels. This method bases the channel on a maximum distance between any closing price and the regression line. We also determined that the method for standard deviation calculation needed to be modified (thanks to Bill S.) to consider price distance from regression trendline as opposed to price distance from average price. Several helpful strategies for using regression lines were shared by users. Information on Raff Channels, Standard Deviations, User Strategies and more can all be found on the web page mentioned above.
Linear Regression Forecast
We discovered that the LRF indicator is really several great indicators built into one: a forecasting tool, a moving regression channel, and a standard deviation oscillator. We applied our new standard deviation calculation method to this indicator as well in computing the bands and the oscillator. The Linear Regression Oscillator was added as a new option to this indicator and provides a useful value representing the number of standard deviations of price from the linear regression line. Several useful strategies and a custom indicator were shared. All can be found on the web page above.
Linear Regression Slope
A normalizing option was added to make this indicator comparable among instruments that don't necessarily trade in the same range. This indicator tells us how steep the regression line (using period specified) is at any time.
Linear Regression Acceleration
A normalizing option was also added to Linear Regression Acceleration, making it comparable among instruments that don't necessarily trade in the same range. This indicator tells us the change in slope. A positive value of Acceleration tells us the slope is increasing while a negative value of Acceleration tells us the slope is decreasing. I would still like to encourage everyone to share their strategies using the indicators above and I will post these strategies to the respective pages on the web site. After looking at a chart involving the linear regression oscillator, slope, and acceleration yesterday, I want to further investigate creating (and optimizing) a signal that involves all 3 tokens.