I have used StdErr in some of my indicators. I've looked up Std Error in the reference book "Standard Math Tables ", and here is what I think it does:
1. It takes the computed best fit y = mx + b equation and computes the difference between the fitted value at a given point along the line and the value of the data at the same point. Now you have an array of differences for each value of x. 2. Take the absolute value of each of the array. 3. Take the standard deviation of the array. Reef-Break --- In [email protected], "Ara Kaloustian" <[EMAIL PROTECTED]> wrote: > > Is the StdErr() function normalized? > > Example: > > I want to find the number of bars, (in the range of 3 to 10 bars) that have best fit to a Linear Regression Trend Line. > > If I loop though the data, can I simply pick the iteration with the smallest stderr without any further calculations or do I need to normalize it for the number of bars associated with each reading? > > > Tx > > Ara >
