JuiceBox CRSI Enhanced

5월 1일

JuiceBox “CRSI Enhanced” is a single-pane, zero-lag Connors RSI indicator supercharged with multi-theory lenses, Jurik smoothing, and multi-timeframe consensus.

1. Base Oscillator (JL-CRSI):
- Computes Connors RSI (3‐period price RSI, 2‐period streak RSI, 100-period percentile rank)
- Smooths it with a true Jurik Moving Average (configurable length & phase)

2. Sliding‐Window Divergence Filter:
- Detects classic price–indicator divergences over a recent look-back window
- Only lets signals fire when CRSI and price lows or highs diverge in the same direction

3. MTF Consensus (Ultra-product):
- For each lens, checks that at least 2 of {1m, 3m, 5m, 15m} agree on the same condition
- Ensures you see only the tightest, zero-lag multi-timeframe confirmation

4. Four “Lenses” (overlaid on the CRSI line):

• Jerk (1ˢᵗ derivative) as a histogram, volume-weighted and ATR-scaled for adaptive sensitivity
  
• Infinitesimal Divergence (2ⁿᵈ derivative) as a thin histogram, using a dynamic ε based on recent volatility
  
• Zero-Cross markers (up/down labels) on the detrended CRSI midline, filtered by MTF consensus
  
• Recurrence crosses, spotting 3-bar “W”/“M” micro-patterns that exceed a minimum amplitude and extend when volume surges
  

5. Classic RSI Reference Lines:
- 30, 50, 70 thresholds drawn with customizable solid, dashed or dotted styles

5월 4일

릴리즈 노트

True KNN (K-Nearest Neighbors) Classifier:

Machine learning classification using historical JL-CRSI, jerk, and Inf-Div values.
Enhances signal prediction accuracy based on learned historical patterns.
Visualized as dynamic colour changing CRSI lines.

Fractal Dimension Drops:

Identifies simultaneous fractal dimension declines across multiple timeframes.
Flags moments of reduced market randomness, indicating possible directional moves.
Visualized as purple labels.

Laguerre Filter Refinement:

Additional smoothing (Laguerre) on Inf-Div signals for noise reduction.
Strengthens the reliability of momentum divergence signals.

Adaptive Overbought/Oversold (OB/OS):

Percentile-based dynamically adjusted OB/OS thresholds.
Adapts automatically to recent volatility conditions.

5월 5일

릴리즈 노트

📌 1. Adaptive Thresholds & Midline

• Thereshold colors adapt based on recurrence (finalRecS, finalRecB) clearly distinguishing bullish/bearish.
  
• Midline clearly changes color (fdrop) based on fractal dimension stability, providing visual cues.
  

📌 2. Histogram Plotting

• Bullish histograms start from 0 upwards clearly.
  
• Bearish histograms are inverted and start clearly from 100 downwards.
  
• Scaling (1)*100 ensures values strictly remain within 0–100.