OmmAlpha · VCP Stage Engine V3 Shadow

Read-only live comparison. Nothing on this page writes to production tables or changes ML predictions.

2026-09-30Live scan date

1Rows shadow-tested

0Stage classifications changed

vcp-stage-v3-shadow-1.0Shadow engine

Core V3 rule: pivot comes from confirmed swing resistance preceding the final contraction trough. A fresh breakout spike needs future bars before it can become a confirmed swing, so the pivot cannot chase the breakout upward the next day.

Why shadow mode first: Phase 2A ML was trained on the old detector population. Switching stage/pivot definitions without rebuilding historical signals and retraining would create model-population drift.

Stage distribution

Current production

FAILED: 1

V3 shadow

FAILED: 1

Live stage comparison

Symbol	Old stage	V3 stage	Detail	Old pivot	V3 structural pivot	V3 distance	Pivot evidence	Contractions	Breakout evidence	Confidence
ATGL close 592.00	FAILED	FAILED	FAILED_BREAKOUT	624.40 5.19% below	624.40 2026-09-09	5.19% below	CONFIRMED_SWING_RESISTANCE touches 9	14.3% → 3.8%	confirmed breakout later lost pivot on multiple closes	80%

ATGL · V3 reasons

V3 structural pivot 624.40 (CONFIRMED_SWING_RESISTANCE, touches 9)
2 contraction(s); tightening 100%
final contraction 3.8%
5.19% below structural pivot
5D/20D volume 0.22x
day/20D breakout volume 0.23x
confirmed breakout later lost pivot on multiple closes
stage detail FAILED_BREAKOUT

Thresholds: confirm 0.36% · failure buffer 2.16% · pre-breakout zone 4.86% · extended 8.00%