0809 HOLD RUNAWAY
08:00–09:00 ET zone holds into RTH, then price runs away from it. Rare (~13/yr) but close-direction matches hold direction 96.6% bull / 75.7% bear.
updated 2026-04-21The setup
During 08:00–09:00 ET, a tight zone forms. If RTH opens and holds above that zone (no retrace back into it), the day closes bull 96.6% of the time. Holds below close bear 75.7%. It doesn’t fire often — 66 events over 5 years — but when it does, it is the most reliable directional signal we have.
Shown on our chart HUD as 0809 HOLD UP / 0809 HOLD DN.
Entry rules
- Zone: high & low of the 08:00–09:00 ET one-hour bar
- Trigger: post-open, price closes outside the zone and does not retrace back into it
- HOLD UP: holds above → close BULL expected
- HOLD DN: holds below → close BEAR expected
- Target: ny_high + 185 (UP) or ny_low − 210 (DN) — aggressive excursion targets
- Probability: 95 UP / 82 DN
- TTL: through RTH close
Backtest
audit_0809_hold_hrr.py, n = 1,365 sessions, 5-year MNQ:
| Cohort | n | Close past zone | Close-direction | Avg close pts |
|---|---|---|---|---|
| HOLD UP | 29 | 100% | 96.6% BULL | +140 |
| HOLD DN | 37 | 81.1% | 75.7% BEAR | −148 |
Edge above baseline: +43 (bull), +29 (bear).
Why the target is aggressive
The aggressive excursion targets (±185 / −210) only hit by close ~30–41% of the time — but the close-direction is 96.6% / 75.7%. The signal is a directional bet, not a precise-target bet. When it works, it works big (avg +140 / −148pt), so we size the target for the tail, knowing most resolve in direction without hitting the line.
Why Tier 1
- Single highest close-direction rate in the model suite (96.6% bull).
- Zero rescue needed — the 08:00–09:00 zone is directly observable; the hold is a clean real-time trigger.
- Small but not tiny sample (n=66) with clean edge both sides.
- Matches backtest sample exactly (same dataset, independent audit confirmed numbers).
History
- 2026-04-21 — Validated via Audit Round 6. Probabilities 95/82 confirmed accurate. No changes needed; audit-confirmation comment added to both .cs files.