Moves & reversals

Each row has its own entry price. All hits are inclusive and observed before that session ends. Later entries have less time; short sessions have no 14:00 entry. First-hit timing is conditional on a hit.

Up/down columns use the threshold of each row. Barriers are nested, so their percentages must not be added. Both sides means both +threshold and −threshold relative to entry, not a range of that size.

Denominator: only days that reached the first barrier. Return to entry and opposite barrier must occur afterwards. Wait times below run from the first impulse and include successful cases only; failures are not assigned a wait time.

Winter = Jan/Feb/Dec of the same year; spring = Mar–May; summer = Jun–Aug; autumn = Sep–Nov. Rates are weighted by actual eligible days, not averaged across years.

Only complete 390-minute sessions: identical days in every half-hour. Path and sampled range are percentages of the session open. Share is the average half-hour share of each non-flat day’s path; shares sum to 100%. Entry and barrier filters do not affect this block or RV.

RV = 100 × √Σ ln(Pᵢ / Pᵢ₋₁)². Not the absolute path and not annualized. RV5 samples every five minutes; RV1 samples every minute. Both include short days without scaling. These numbers are known at the end of the day, not forecasts.

Daily rows compare RV5 with the previous 20 and 60 scheduled sessions. Today is excluded from both baselines. Missing prior sessions leave the reference blank; no future data or shorter substitute window.

Largest absolute net return across all sliding windows of exactly 5, 15 or 30 minutes. Minute endpoints only, relative to the start of each window. Not a range, price path or executable profit.