BitLogic - Kalman Composite

BitLogic Kalman Composite (BL-KC)

What it is
A momentum/condition oscillator that filters price with a multi-stage Kalman and blends two normalized branches into one composite line with a compact score histogram. Built for cleaner flips and fewer whipsaws.

How it works

• Kalman filter (5-stage) on your chosen price source; selectable output (Stage1/Stage5/Average).
  
• Branch A: RSI on Kalman price → normalized to ~[−0.5, +0.5].
  
• Branch B (selectable):
  - Residual Z: z-score of the Kalman residual (observation − predicted state), squashed for
  stability (distinct vs classic KSO)
  - Williams %R on Kalman price (normalized).
  
• Gain-weighted blend: the composite weights Branch B by the average Kalman gain (when the filter trusts new info more, residual matters more).
  
• Zero-line hysteresis: small band around 0 to reduce flip noise.
  
• Score (columns): quadrant logic → 1, 0.5, −0.5, −1 for quick read of bias + slope.
  
• No repainting: updates/alerts on bar close.
  

Inputs you’ll care about

• Q/R (process/measurement noise): responsiveness vs smoothness.
  
• Blend: base weight + gain weighting.
  
• Residual Z: lookback & squash scale (controls sensitivity).
  
• Hysteresis band and optional EMA smoothing of the composite.
  

Reading it

• Line (ci): above 0 → bullish zone; below 0 → bearish zone.
  
• Columns (KC_score): show strength/weakness inside each zone (green ≥ 0, orange < 0).
  
• Alerts: bullish/bearish flip fire on close when the composite crosses the band edges.
  

Tips

• For faster markets: raise Q, lower smoothing, keep a small hysteresis (e.g., 0.03–0.05).
  
• For trend following: use Stage5/Average Kalman output and a slightly wider band (0.06–0.10).
  
• Want “classic” feel? Switch Branch B to Williams %R.
  

Credits
Inspired by the community idea behind the Kalman Synergy Oscillator (Kalman + RSI + %R). This is an independent, from-scratch implementation with a residual z-score branch and gain-weighted blending for distinct behavior.

Disclaimer
For educational purposes only. Not financial advice. Past performance does not guarantee future results.