Fixed before evaluation
1H signals use a 4-hour window; 4H signals use 24 hours; 1D signals use seven calendar days. Expired windows remain in the denominator.
Demo network ready Model 0.1.4
AI DEMO
BTC$118,420+1.84%ETH$3,812+0.72%SOL$194.22-0.36%DXY97.64-0.18%US10Y4.21%+3 bpGOLD$3,344+0.44%
SIMULATED06SIMULATED DATA
Verification
Methods, failures and unknowns are first-class evidence.
METHOD LEDGER · V0.1Evidence before
Evidence before
performance claims.
Confirmed means a predefined outcome condition was observed within its fixed evaluation window. It does not mean profitable, executable or financially advised.
62SIGNAL OBSERVATIONS
18INVALIDATED RETAINED
02FAILED RISK EVENTS
Swipe horizontally to inspect every ledger field →
VERIFICATION AREASTATUSRESULTMETHODDATA STATE
Signal PerformanceVERIFIED62 observationsFixed 4H / 1D outcome windowsSIMULATED
Strategy PerformancePROPOSED3 research setsCost-adjusted historical replayBACKTEST
Agent AccuracyUNKNOWNCalibration onlyBrier score design pending reviewSIMULATED
Invalidated SignalsPASS18 retainedNo outcome selectionSIMULATED
Risk EventsFAILED2 threshold breachesShown without exclusionBACKTEST
Model Update 0.1.4DEFERREDReview requiredNo production promotionNOT RUN
External DataBLOCKEDNot connectedDemo dataset onlyNOT RUN
Reference to outcome price
Returns use the stated reference price to the fixed horizon close. Strategy replays include modeled fees of 10 bps and slippage of 8 bps unless otherwise noted.
Never merge performance types
Live uses verified external observations. Paper uses real-time data without funds. Backtest uses historical replay. Simulated is generated for this interface demo.