OPEN-SOURCE SCRIPT
Zaktualizowano Linear Regression Forecast (ADX Adaptive)
Linear Regression Forecast (ADX Adaptive)
This indicator is a dynamic price projection tool that combines multiple linear regression forecasts into a single, adaptive forecast curve. By integrating trend strength via the ADX and directional bias, it aims to visualize how price might evolve in different market environments—from strong trends to mean-reverting conditions.
Core Concept:
This tool builds forward price projections based on a blend of linear regression models with varying lookback lengths (from 2 up to a user-defined max). It then adjusts those projections using two key mechanisms:
ADX-Weighted Forecast Blending
In trending conditions (high ADX), the model follows the raw forecast direction. In ranging markets (low ADX), the forecast flips or reverts, biasing toward mean-reversion. A logistic transformation of directional bias, controlled by a steepness parameter, determines how aggressively this blending reacts to price behavior.
Volatility Scaling
The forecast’s magnitude is scaled based on ADX and directional conviction. When trends are unclear (low ADX or neutral bias), the projection range expands to reflect greater uncertainty and volatility.
How It Works:
Regression Curve Generation
For each regression length from 2 to maxLength, a forward projection is calculated using least-squares linear regression on the selected price source. These forecasts are extrapolated into the future.
Directional Bias Calculation
The forecasted points are analyzed to determine a normalized bias value in the range -1 to +1, where +1 means strongly bullish, -1 means strongly bearish, and 0 means neutral.
Logistic Bias Transformation
The raw bias is passed through a logistic sigmoid function, with a user-defined steepness. This creates a probability-like weight that favors either following or reversing the forecast depending on market context.
ADX-Based Weighting
ADX determines the weighting between trend-following and mean-reversion modes. Below ADX 20, the model favors mean-reversion. Above 25, it favors trend-following. Between 20 and 25, it linearly blends the two.
Blended Forecast Curve
Each forecast point is blended between trend-following and mean-reverting values, scaled for volatility.
What You See:
Forecast Lines: Projected future price paths drawn in green or red depending on direction.
Bias Plot: A separate plot showing post-blend directional bias as a percentage, where +100 is strongly bullish and -100 is strongly bearish.
Neutral Line: A dashed horizontal line at 0 percent bias to indicate neutrality.
User Inputs:
-Max Regression Length
-Price Source
-Line Width
-Bias Steepness
-ADX Length and Smoothing
Use Cases:
Visualize expected price direction under different trend conditions
Adjust trading behavior depending on trending vs ranging markets
Combine with other tools for deeper analysis
Important Notes:
This indicator is for visualization and analysis only. It does not provide buy or sell signals and should not be used in isolation. It makes assumptions based on historical price action and should be interpreted with market context.
This indicator is a dynamic price projection tool that combines multiple linear regression forecasts into a single, adaptive forecast curve. By integrating trend strength via the ADX and directional bias, it aims to visualize how price might evolve in different market environments—from strong trends to mean-reverting conditions.
Core Concept:
This tool builds forward price projections based on a blend of linear regression models with varying lookback lengths (from 2 up to a user-defined max). It then adjusts those projections using two key mechanisms:
ADX-Weighted Forecast Blending
In trending conditions (high ADX), the model follows the raw forecast direction. In ranging markets (low ADX), the forecast flips or reverts, biasing toward mean-reversion. A logistic transformation of directional bias, controlled by a steepness parameter, determines how aggressively this blending reacts to price behavior.
Volatility Scaling
The forecast’s magnitude is scaled based on ADX and directional conviction. When trends are unclear (low ADX or neutral bias), the projection range expands to reflect greater uncertainty and volatility.
How It Works:
Regression Curve Generation
For each regression length from 2 to maxLength, a forward projection is calculated using least-squares linear regression on the selected price source. These forecasts are extrapolated into the future.
Directional Bias Calculation
The forecasted points are analyzed to determine a normalized bias value in the range -1 to +1, where +1 means strongly bullish, -1 means strongly bearish, and 0 means neutral.
Logistic Bias Transformation
The raw bias is passed through a logistic sigmoid function, with a user-defined steepness. This creates a probability-like weight that favors either following or reversing the forecast depending on market context.
ADX-Based Weighting
ADX determines the weighting between trend-following and mean-reversion modes. Below ADX 20, the model favors mean-reversion. Above 25, it favors trend-following. Between 20 and 25, it linearly blends the two.
Blended Forecast Curve
Each forecast point is blended between trend-following and mean-reverting values, scaled for volatility.
What You See:
Forecast Lines: Projected future price paths drawn in green or red depending on direction.
Bias Plot: A separate plot showing post-blend directional bias as a percentage, where +100 is strongly bullish and -100 is strongly bearish.
Neutral Line: A dashed horizontal line at 0 percent bias to indicate neutrality.
User Inputs:
-Max Regression Length
-Price Source
-Line Width
-Bias Steepness
-ADX Length and Smoothing
Use Cases:
Visualize expected price direction under different trend conditions
Adjust trading behavior depending on trending vs ranging markets
Combine with other tools for deeper analysis
Important Notes:
This indicator is for visualization and analysis only. It does not provide buy or sell signals and should not be used in isolation. It makes assumptions based on historical price action and should be interpreted with market context.
Informacje o Wersji
bias calculationInformacje o Wersji
cleanerInformacje o Wersji
bias calculationInformacje o Wersji
1:1 Gann angle benchmark ratio calculation for Biasbetter bias display
Informacje o Wersji
forecast actually blendsInformacje o Wersji
anti-trend weightingInformacje o Wersji
made logistic range an inputInformacje o Wersji
logistic formula non-flipSkrypt open-source
W duchu TradingView twórca tego skryptu udostępnił go jako open-source, aby traderzy mogli analizować i weryfikować jego funkcjonalność. Brawo dla autora! Możesz korzystać z niego za darmo, ale pamiętaj, że ponowna publikacja kodu podlega naszym Zasadom Społeczności.
Wyłączenie odpowiedzialności
Informacje i publikacje przygotowane przez TradingView lub jego użytkowników, prezentowane na tej stronie, nie stanowią rekomendacji ani porad handlowych, inwestycyjnych i finansowych i nie powinny być w ten sposób traktowane ani wykorzystywane. Więcej informacji na ten temat znajdziesz w naszym Regulaminie.
Skrypt open-source
W duchu TradingView twórca tego skryptu udostępnił go jako open-source, aby traderzy mogli analizować i weryfikować jego funkcjonalność. Brawo dla autora! Możesz korzystać z niego za darmo, ale pamiętaj, że ponowna publikacja kodu podlega naszym Zasadom Społeczności.
Wyłączenie odpowiedzialności
Informacje i publikacje przygotowane przez TradingView lub jego użytkowników, prezentowane na tej stronie, nie stanowią rekomendacji ani porad handlowych, inwestycyjnych i finansowych i nie powinny być w ten sposób traktowane ani wykorzystywane. Więcej informacji na ten temat znajdziesz w naszym Regulaminie.