Kalman Hull Trend Score [BackQuant]

Kalman Hull Trend Score [BackQuant]

Overview
Kalman Hull Trend Score is a trend-strength and regime-evaluation indicator that combines two ideas, Kalman filtering and Hull-style smoothing, then measures persistence of that filtered trend using a rolling score. The goal is to produce a cleaner, more stable trend read than typical moving average tools, while still reacting fast enough to be practical in live markets.

Instead of treating a moving average as a simple line you cross, this indicator turns the filtered trend into an oscillator-like score that answers: “Is the smoothed trend consistently progressing, or is it stalling and degrading?”

Core idea
The indicator is built from two components:

• A Kalman-based smoothing engine that estimates price state and reduces noise adaptively.
  
• A Hull-style construction that uses multiple Kalman passes to create a responsive, low-lag trend filter.
  

Once the Kalman Hull filter is built, a persistence score is calculated by comparing the current Kalman Hull value to many past values. The result is a trend score that rises in sustained trends and compresses or flips during deterioration.

Why Kalman instead of standard smoothing
Traditional moving averages apply fixed smoothing rules regardless of market conditions. A Kalman filter behaves differently, it is designed to estimate an underlying state in noisy data, adjusting how much it “trusts” new price information versus prior estimates.

This script exposes that behavior through two key controls:

• Measurement Noise: how noisy the observed price is assumed to be.
  
• Process Noise: how much the underlying state is allowed to evolve from bar to bar.
  

Together, these settings let you tune the balance between smoothness and responsiveness without relying on blunt averaging alone.

Kalman filter mechanics (conceptual)
Each update cycle follows the classic structure:

• Prediction: assume the state continues, and expand uncertainty by process noise.
  
• Update: compute Kalman Gain, then blend the new price observation into the estimate.
  
• Correction: reduce uncertainty based on how much the filter accepted the new information.
  

When measurement noise is higher, the filter becomes more conservative, smoothing harder. When process noise is higher, the filter adapts faster to regime changes, but can become more reactive.

Check out the original script:

Kalman Hull construction
The “Hull” component is not a standard HMA built from WMAs. Instead, it recreates the Hull idea using Kalman filtering as the smoothing primitive. The structure follows the same intent as HMA, reduce lag while keeping the line smooth, but does it with Kalman passes:

• Apply Kalman smoothing over multiple effective lengths.
  
• Combine them using the Hull-style weighting logic.
  
• Run the combined output through another Kalman pass to finalize smoothing.
  

The result is a Kalman Hull filter that aims to track trend with less jitter than raw price, and less lag than slow averages.

Another Kalman Hull with Supertrend

Trend scoring logic
The trend score is computed by comparing the current Kalman Hull value to past Kalman Hull values over a fixed lookback range (1 to 45 bars in this script):

• If current kalmanHMA > kalmanHMA, add +1
  *If current kalmanHMA < kalmanHMA, add -1
  

This produces a persistence score rather than a simple direction signal. Strong trends where the filter keeps advancing will accumulate positive comparisons. Weak trends, chop, or reversals will cause the score to flatten, decay, or flip negative.

Interpreting the score
Read the score as trend conviction and persistence:

• High positive values: bullish persistence, the filtered trend is progressing consistently.
  
• Low positive values: trend exists but is fragile, progress is slowing.
  
• Near zero: indecision, range behavior, frequent challenges to structure.
  
• Negative values: bearish persistence or sustained deterioration in the filtered trend.
  

The rate of change matters:

• Score expansion suggests trend is gaining traction.
  
• Score compression often signals consolidation or exhaustion.
  
• Fast flips usually accompany regime transitions.
  

Signal thresholds and regime transitions
User-defined thresholds convert the score into regimes:

• Long threshold: score must exceed this level to confirm bullish persistence.
  
• Short threshold: a crossunder of the score triggers bearish regime transition.
  

This is intentionally conservative. Long bias is maintained while the score holds above the long threshold. Short transitions are event-triggered on breakdown via crossunder, helping avoid constant flipping during minor noise.

Signals are only plotted on regime changes (first bar of the flip), keeping them clean for alerts and backtests.

Visual presentation
The indicator provides multiple layers depending on how you want to use it:

• Kalman Hull Trend Score oscillator, color-coded by active regime.
  
• Optional Kalman Hull filter plotted on the price chart for structure context.
  
• Optional threshold reference lines for quick regime mapping.
  
• Optional candle coloring and background shading for instant readability.
  

You can run it as a pure score panel or as a combined panel + on-chart trend overlay.

How to use in practice

Trend filtering

• Favor long setups when the score remains above the long threshold.
  
• Reduce directional aggression when score compresses toward zero.
  
• Treat a short-threshold breakdown as a regime risk event, not just a signal.
  

Trend quality assessment

• Rising score supports continuation trades and adds confidence to breakouts.
  
• Flat or falling score warns that trend persistence is fading.
  
• If price trends but score fails to expand, trend may be weak or liquidity-driven.
  

Trade management

• Use the Kalman Hull line as dynamic structure reference on chart.
  
• Use score deterioration to scale out before a full regime flip.
  
• Use regime flips as confirmation for bias shifts rather than prediction.
  

Tuning guidelines

Measurement Noise

• Higher: smoother filter, fewer false shifts, slower to adapt.
  
• Lower: more responsive, more sensitive to microstructure noise.
  

Process Noise

• Higher: adapts quicker to sudden changes, but can become twitchy.
  
• Lower: steadier state estimate, but slower during sharp regime transitions.
  

A practical approach is to first tune measurement noise until the Kalman Hull line matches the “clean trend structure” you want, then adjust process noise to control how quickly it reacts when the regime genuinely changes.

Summary
Kalman Hull Trend Score transforms a Kalman-based Hull-style trend filter into a quantified persistence oscillator. By combining adaptive Kalman smoothing with low-lag Hull logic and a rolling comparison score, it provides a cleaner read on trend quality than basic moving averages or single-condition trend tools. It is best used as a regime filter, trend strength gauge, and structure-aware trade management layer.