Recursive Least Squares Forecast

Recursive Least Squares Forecast fits an adaptive linear regression whose slope and intercept are re-estimated on every bar, plotting the result as an equilibrium RLS Mean with volatility-adjusted bands. A forward "Ghost Line" projects the current estimate ahead, wrapped in standard-deviation forecast bands that visualize the model's uncertainty, a low-lag alternative to static fits.

How to Trade the Recursive Least Squares Forecast?

• RLS Mean slope: a rising mean with an upward Ghost Line frames bullish conditions; a declining structure signals bearish pressure.
• Outer band excursion: price beyond the Upper or Lower Band is statistically overextended for the current regime, the start of a mean-reversion setup.
• BUY / SELL labels: printed when price crosses back inside the band; the RLS Mean serves as the natural profit objective.
• Forgetting Factor tuning: values near 0.95 privilege recent action and pivot fast; values near 0.99 favor stability and smoothness.

The engine is an adaptive filter updated through a gain vector in a state-space framework, a form of online, incremental learning that discounts older observations instead of refitting a fixed window. That is what lets it pivot quickly through regime shifts without excessive lag or noise.

Recursive Least Squares Forecast Settings

• Forgetting Factor (λ): how quickly older data is discounted. Higher is smoother and more stable, lower reacts faster to new prices.
• Band Multiplier: scales the forecast standard deviation that sets the band width.
• Forecast Horizon: how many bars ahead the Ghost Line and forecast bands project.
• Show Ghost Line & Bands: toggles the forward projection and uncertainty bands.
• Show Mean Reversion Signals: toggles the BUY and SELL labels.

Frequently Asked Questions

Does the Recursive Least Squares Forecast repaint?

How does it compare to other regression tools in the library?

Where can I get this indicator?