Linear Regression Forecast

26 May

Description:
This indicator computes a series of simple linear regressions anchored at the current bar, using look-back windows from 2 bars up to the user-defined maximum. Each regression line is projected forward by the same number of bars as its look-back, producing a family of forecast endpoints. These endpoints are then connected into a continuous polyline: ascending segments are drawn in green, and descending segments in red.

Inputs:
maxLength – Maximum number of bars to include in the longest regression (minimum 2)
priceSource – Price series used for regression (for example, close, open, high, low)
lineWidth – Width of each line segment

Calculation:
For each window size N (from 2 to maxLength):
• Compute least-squares slope and intercept over the N most recent bars (with bar 0 = current bar, bar 1 = one bar ago, etc.).
• Project the regression line to bar_index + N to obtain the forecast price.
Collected forecast points are sorted by projection horizon and then joined:
• First segment: current bar’s price → first forecast point
• Subsequent segments: each forecast point → next forecast point
Segment colors reflect slope direction: green for non-negative, red for negative.

Usage:
Apply this overlay to any price chart. Adjust maxLength to control the depth and reach of the forecast fan. Observe how shorter windows produce nearer-term, more reactive projections, while longer windows yield smoother, more conservative forecasts. Use the colored segments to gauge the overall bias of the fan at each step.

Limitations:
This tool is for informational and educational purposes only. It relies on linear regression assumptions and past price behavior; it does not guarantee future performance. Users should combine it with other technical or fundamental analyses and risk management practices.

26 May

Sürüm Notları

flip