14d CORE evidence window

CORE evidence. Route simulation and resolved prediction outcomes, side by side.

This page separates the simulated route report from resolved public CORE prediction outcomes. Neither stream is an exchange fill, a customer trade, or customer PnL.

Current endpoints load on each visit. The route report labels live simulation versus static fallback; prediction outcomes identify their predictions-table source. Weak windows stay visible.

Evidence order

How to read this page

The hierarchy is deliberately plain: resolved prediction outcomes first, route simulation second, open-tape counts as unrealized context only.

Resolved prediction outcomes

Rows from the predictions-backed public endpoint. Runtime-forward CORE prediction rows are recorded at signal fire; a path-based outcome is appended later from stored bars. This is research evidence, not customer execution.

Route simulation

PF, HR, and n from /api/scanner/tier-performance. Useful for routing review, but still a simulated route report rather than customer PnL.

Current status

loading evidence

No weak-window alert from the loaded evidence streams.

Route simulation

CORE route-simulation context

CORE tier route simulation (sweet: live (under watch); liqsqueeze: shadow-only — not currently delivered). Source-labeled, separate from the resolved prediction-outcome archive.

Evidence terminal

Open context and resolved CORE outcomes

The public terminal pairs trailing-7-day open CORE counts with outcomes now resolved for CORE signals fired in the trailing 14 days. Outcomes may resolve later. Open-row details stay member-only; aggregate counts and resolved prediction rows stay public.

Open tape

Resolved outcomes for CORE signals fired in 14d

Loading resolved CORE prediction outcomes.

Validation health

Live model posture and evidence quality

Model posture

The telemetry behind the receipts page.

Route-simulation PF/HR telemetry and unresolved-queue posture are pulled from the same public route that feeds receipt framing on /performance.

Core route source
...
loading
Top deployed bucket
loading
Backtest confidence
awaiting backtest
Backtest source
unavailable
1h IC / n unavailable

Telemetry is still loading.

SENTINEL mascot (waiting)
After the evidence

Verify, then decide.

Read the ledger first, including the worst row and loss context. If it still earns your attention, connect your wallet.

Want to verify the ledger first? Open the loss ledger.

Model operations

Model and alert diagnostics

Model tuning diagnostics

Public view of ML edge evidence and threshold-risk posture from live bucket telemetry.

Loading optimization telemetry…

Loading model telemetry...

Reading notes

Metric definitions

PF — profit factor

Profit factor from the route-simulation endpoint for the selected report window.

HR — hit rate

Hit rate for the same route-simulation sample. Thin samples are context, not claims.

n — simulated route rows

Simulated route count. A small n should make the reader more cautious, not more excited.