Wyckoff (WY)

What It Measures

Wyckoff classifies phase (accumulation/markup/distribution/markdown/range), events, and directional probabilities.

Inputs

• Trend, momentum, effort-vs-result, ADX, and volume-quality features.
• Detected Wyckoff event stream with confidence tags.
• Optional higher-timeframe phase context.

• wy_min_confidence_to_display: 0.35
• wy_event_lookback_bars: {"1d": 90, "1w": 64, "4h": 120}
• wy_phase_smoothing_bars: 5
• wy_volume_zscore_window: 34
• wy_spring_threshold: 0.35
• wy_upthrust_threshold: 0.35

Calculation Pipeline

1. Compute feature scores per phase and apply cross-timeframe adjustment.
2. Pick phase with highest score and compute score spread.
3. Convert bull/bear component delta to probabilities.
4. Build confidence from spread, events, and volume quality.
5. Generate zones, forecast scenarios, and invalidation.

Formulas & Rules

raw_delta = (accumulation + markup) - (distribution + markdown)
delta_scale = 8.0 if phase != 'range' else 5.0
bull_prob = clip(round(50 + raw_delta * delta_scale), 5, 95)
bear_prob = 100 - bull_prob

confidence = 18
confidence += (score_spread / top_score) * 45
confidence += min(20, events * 3.5)
confidence += min(8, abs(raw_delta) * 1.1)
range and poor volume quality apply penalties

Live Worked Example

Failure / Fallback Behavior

• If history/volume quality is insufficient, returns neutral fallback payload with reduced confidence.
• If confidence is below minimum threshold, events are filtered and output can neutralize.
• Forecast still returns deterministic schema with default range-persistence mode.

Interpretation Notes

Wyckoff phase is a structure lens. Use phase + event alignment with invalidation for risk control.