Shadow Mission Control
Read-only experiment telemetry. Paper promotion stays locked until the evidence gate is met.
Readiness runway
Promotion and calibration use one canonical outcome per settled ticker. Simulated fills do not advance them.
Model quality
Settled evidence, calibration, confidence, and readiness.
Data freshness
Latest durable evidence and scanner heartbeat.
Today’s activity
UTC activity from the current experiment.
Hypothetical Shadow P&L
Simulated only · future Paper P&L is separate.
Top 10 opportunities
Highest observed net edge, not an execution queue.
| City | Contract | Est. prob. | Market prob. | Raw edge | Net edge | Confidence | Decision | Timestamp |
|---|
Recent settlements
Last 10 settled shadow candidates.
| City / Contract | Side | Est. | Market | Result | Immediate | Passive | Net edge | Conf. |
|---|
Rolling performance
Hypothetical shadow results by settled-candidate window.
| Window | Record | Win rate | Net P&L |
|---|
Canonical market audit
One forecast outcome per ticker; execution lanes remain separate.
| Ticker | City/date | Rows | Snapshots | Immediate | Passive | Canonical |
|---|
Canonical ROI analysis
Unique-market groups; immediate executable and passive simulation are shown separately.
Calibration
Unique canonical outcomes · YES probability versus realized YES rate.
Learning
Only populated after settled observations.
Readiness checklist
The actual promotion gates.
Mission progress
Daily experiment trajectory.