Support / Resistance (S/R)

What It Measures

S/R identifies objective support and resistance clusters from pivot interactions, reaction strength, and recency.

Inputs

• Candidate level prices from pivots and local structure.
• Touch counts and reaction magnitude over lookahead bars.
• Recency weighting and side classification (support/resistance).

• sr_zone_history_bars: 20
• sr_zone_future_bars: -1
• sr_zone_alpha_base: 0.06
• sr_zone_alpha_strength_mult: 0.1
• sr_zone_alpha_cap: 0.22

Calculation Pipeline

1. Generate level candidates and merge nearby prices.
2. Compute touch stats and post-touch reaction scores.
3. Compute recency and strength (with decay).
4. Filter by recency/strength thresholds with nearby level fallback.
5. Return sorted levels plus nearest support/resistance.

Formulas & Rules

raw = (touches / 6.0) + (reaction_score * 6.0)
base_strength = clip(raw / 1.7, 0.0, 1.0)
recency_decay = exp(-(1 - recency) / tau)
strength = clip((base_strength * recency_decay) + (0.20 * recency), 0.0, 1.0)

recent_ok = recency >= min_recency_ratio OR strength >= stale_relax_strength
strength_ok = strength >= strength_threshold
level kept if recent_ok AND strength_ok

Live Worked Example

Failure / Fallback Behavior

• If timeframe data is empty, returns no levels and null nearest levels.
• If filtering removes all levels on one side, nearby fallback can inject nearest side levels.
• Role-reversal markers (S->R / R->S) are derived from prior side-vote touches.

Interpretation Notes

S/R levels are probabilistic zones. Stronger levels carry more reaction history and recency support.