업데이트됨
Williams %R Jaggedness fix
[I. THE PROBLEM OF RETAIL JAGGEDNESS]
The standard Williams %R is extremely jagged. Relying on a naive (Highest - Close) / (Highest - Lowest) look-back formula makes it a victim of spatial illusion. It is highly vulnerable to "salt-and-pepper" noise, intraday microstructure stop-hunts, and high-frequency volatility spikes. It generates an erratic, jagged signal that triggers false regime shifts, trapping retail liquidity.
[II. THE MULTI-STAGE SOLUTION] Solving the jaggedness problem is not about applying a single, lagging moving average. It requires a continuous-time signal processing pipeline. I developed 4, sovereign hypotheses to isolate the true kinematic intent of the market.
► HYPOTHESIS I: The Robust Nonlinear Butterworth %R (RNB-R)
The Mechanism: Amplitude and Noise Filtering.
The Execution: We cleanse the raw data using a 3-period Median Filter to instantly reject 1-bar liquidity sweeps without introducing lag. The resulting %R is then dynamically smoothed via a Volatility-Adaptive SuperSmoother. The lookback adapts kinematically—tightening during explosive volatility to capture breakouts, and expanding during consolidation to ignore chop.
► HYPOTHESIS II: The Decycled Zero-Lag Cycle Oscillator (DZL-R)
The Mechanism: Frequency Cancellation.
The Execution: We abandon the concept of "averaging" entirely. Using a 48-bar High-Pass filter, we isolate and subtract the high-frequency spectral noise directly from the price (Decycling). By mapping the Williams %R onto this Decycled Price, we produce a zero-lag, wave-like signal that perfectly reflects pure trend movement while ignoring micro-structural noise.
► HYPOTHESIS III: The IMF-Extracted Persistent Shape Wave (IMF-W)
The Mechanism: Spectral Shape Fitting & Empirical Mode Decomposition (EMD).
The Execution: We deconstruct the continuous price manifold into its intrinsic cyclical modes. We discard IMF 1 (high-frequency noise) and the residual (ultra-slow drift). By reconstructing the asset using only IMF 2 and IMF 3, we force the Williams %R to measure purely persistent, rhythmic structural cycles.
► HYPOTHESIS IV: The Bipolar Inverse Fisher Williams %R (BIF-R)
The Mechanism: Probability Distribution Transformation.
The Execution: To maximize state persistence, we smooth the standard oscillator with a 2-pole Gaussian filter, normalize it, and compress it through an Inverse Fisher Transform. This mathematical thermodynamic shift forces the oscillator into a bimodal, polarized state. It snaps quickly between extremes and stays there, eliminating nervous jaggedness in favor of pure regime tokenization.
The standard Williams %R is extremely jagged. Relying on a naive (Highest - Close) / (Highest - Lowest) look-back formula makes it a victim of spatial illusion. It is highly vulnerable to "salt-and-pepper" noise, intraday microstructure stop-hunts, and high-frequency volatility spikes. It generates an erratic, jagged signal that triggers false regime shifts, trapping retail liquidity.
[II. THE MULTI-STAGE SOLUTION] Solving the jaggedness problem is not about applying a single, lagging moving average. It requires a continuous-time signal processing pipeline. I developed 4, sovereign hypotheses to isolate the true kinematic intent of the market.
► HYPOTHESIS I: The Robust Nonlinear Butterworth %R (RNB-R)
The Mechanism: Amplitude and Noise Filtering.
The Execution: We cleanse the raw data using a 3-period Median Filter to instantly reject 1-bar liquidity sweeps without introducing lag. The resulting %R is then dynamically smoothed via a Volatility-Adaptive SuperSmoother. The lookback adapts kinematically—tightening during explosive volatility to capture breakouts, and expanding during consolidation to ignore chop.
► HYPOTHESIS II: The Decycled Zero-Lag Cycle Oscillator (DZL-R)
The Mechanism: Frequency Cancellation.
The Execution: We abandon the concept of "averaging" entirely. Using a 48-bar High-Pass filter, we isolate and subtract the high-frequency spectral noise directly from the price (Decycling). By mapping the Williams %R onto this Decycled Price, we produce a zero-lag, wave-like signal that perfectly reflects pure trend movement while ignoring micro-structural noise.
► HYPOTHESIS III: The IMF-Extracted Persistent Shape Wave (IMF-W)
The Mechanism: Spectral Shape Fitting & Empirical Mode Decomposition (EMD).
The Execution: We deconstruct the continuous price manifold into its intrinsic cyclical modes. We discard IMF 1 (high-frequency noise) and the residual (ultra-slow drift). By reconstructing the asset using only IMF 2 and IMF 3, we force the Williams %R to measure purely persistent, rhythmic structural cycles.
► HYPOTHESIS IV: The Bipolar Inverse Fisher Williams %R (BIF-R)
The Mechanism: Probability Distribution Transformation.
The Execution: To maximize state persistence, we smooth the standard oscillator with a 2-pole Gaussian filter, normalize it, and compress it through an Inverse Fisher Transform. This mathematical thermodynamic shift forces the oscillator into a bimodal, polarized state. It snaps quickly between extremes and stays there, eliminating nervous jaggedness in favor of pure regime tokenization.
릴리즈 노트
new features릴리즈 노트
2 new
면책사항
해당 정보와 게시물은 금융, 투자, 트레이딩 또는 기타 유형의 조언이나 권장 사항으로 간주되지 않으며, 트레이딩뷰에서 제공하거나 보증하는 것이 아닙니다. 자세한 내용은 이용 약관을 참조하세요.
면책사항
해당 정보와 게시물은 금융, 투자, 트레이딩 또는 기타 유형의 조언이나 권장 사항으로 간주되지 않으며, 트레이딩뷰에서 제공하거나 보증하는 것이 아닙니다. 자세한 내용은 이용 약관을 참조하세요.