Research Lab
PL_TRUSTED_RESEARCH_EPOCH_V2 · RUST TRUTH KERNEL · live_trading=false
One control room. One KPI contract. Legacy research is forensic-only.
RESEARCH LAB — ALPHA CONTROL ROOM
Maximize new behavior → OOS → Independent forward evidence → proven alpha · live_trading=false
Forward Proven is research-independent evidence. Paper Verified is closed Paper evidence. They are different.
INDEPENDENT FORWARD OUTCOMES / DAY · —
UNIQUE-PRINT · FEE-TRUE $ primary · R may diverge via sizing/stop normalization (not an adder bug). Strong standardized $ is not rejected solely for negative R when reconciliation is A_MATHEMATICALLY_CORRECT_DIVERGENCE_SIZING_VS_R.
| Window | Unique outcomes | W/L | Win $ | Loss $ | Net $ |
|---|---|---|---|---|---|
| — | —/— | — | — | — | |
| — | —/— | — | — | — | |
| — | —/— | — | — | — | |
| — | —/— | — | — | — | |
| — | —/— | — | — | — |
Forward Activity — LIFECYCLE ACTIVITY (not unique-print economics)
Not a conversion funnel. Stage timestamps · LIFECYCLE ACTIVITY unit (not unique-print).
| Win | Sig | Entry-Ready | Fills | Opens | Lifecycle closes |
|---|---|---|---|---|---|
| 1H | — | — | — | — | — |
| 6H | — | — | — | — | — |
| 24H | — | — | — | — | — |
| 7D | — | — | — | — | — |
Signal Cohort Conversion — same signal_id
signal_id = spec|signal_ts · cohort_closes ≤ lifecycle closes · economics Unique outcomes above remain UNIQUE-PRINT
| Win | Sig | Entry-Ready | Filled | Opened | Closed |
|---|---|---|---|---|---|
| 1H | — | — | — | — | — |
| 6H | — | — | — | — | — |
| 24H | — | — | — | — | — |
| 7D | — | — | — | — | — |
LIVE=0 · Observability only · VM SSOT → snapshot → API → UI · GitHub deploy=false · INDEPENDENT_FORWARD_OUTCOMES_PER_DAY
Auto-generated hypotheses (not hardcoded). BBO seats 0 → 0 under LEGACY_FAMILY_SATURATED.