OPEN-SOURCE SCRIPT

WaveTrend RBF

147
What it does

WT-RBF extracts a “wave” of momentum by subtracting a fast Gaussian-weighted smoother from a slow one, then robust-normalizes that wave with a median/MAD proxy to produce a z-score (z). A short EMA of z forms the signal line. Optional dynamic thresholds use the MAD of z itself so overbought/oversold levels adapt to volatility regimes.

How it’s built:
  • Radial (Gaussian) smoothers
  • Causal, exponentially-decaying weights over the last radius bars using σ (sigma) to control spread.
  • fast = rbf_smooth(src, fastR, fastSig)
  • slow = rbf_smooth(src, slowR, slowSig)
  • wave = fast − slow (band-pass)
  • Robust normalization
  • A two-stage EMA approximates the median; MAD is estimated from EMA of absolute deviations and scaled by 1.4826 to be stdev-comparable.
  • z = (wave − center) / MAD


Thresholds

  • Dynamic OB/OS: ±2.5 × MAD(z) (or fixed levels when disabled)


Reading the indicator
  • Bull Cross: z crosses above sig → momentum turning up.
  • Bear Cross: z crosses below sig → momentum turning down.
  • Exits / Bias flips: zero-line crosses (below 0 → exit long bias; above 0 → exit short bias).
  • Overbought/Oversold: z > +thrOB or z < thrOS. With dynamics on, the bands widen/narrow with recent noise; with dynamics off, static guides at ±2 / ±2.5 are shown.


Core Inputs
  • Source: Price series to analyze.
  • Fast Radius / Fast Sigma (defaults 6 / 2.5): Shorter radius/smaller σ = snappier, higher-freq.
  • Slow Radius / Slow Sigma (defaults 14 / 5.0): Larger radius/σ = smoother, lower-freq baseline.

Normalization
Robust Z-Score Window (default 200): Lookback for median/MAD proxy (stability vs responsiveness).
Small ε for MAD: Floor to avoid division by zero.

Signal & Thresholds
  • Dynamic Thresholds (MAD-based) (on by default): Adaptive OB/OS; toggle off to use fixed guides.
  • Visuals
  • Shade OB/OS Regions: Background highlights when z is beyond thresholds.
  • Show Zero Line: Midline reference.

(“Plot Cross Markers” input is present for future use.)

Отказ от ответственности

Все виды контента, которые вы можете увидеть на TradingView, не являются финансовыми, инвестиционными, торговыми или любыми другими рекомендациями. Мы не предоставляем советы по покупке и продаже активов. Подробнее — в Условиях использования TradingView.