HISTORICAL PROOF
Archive, history, and proof of edge.
Feed is the proof layer: archived signals, win history, realized edge, and the next step for users who want the live daily-calls product.
Live success rate
—
— completed signals
Average return
—
Weighted realized return across tracked signal types
Average drawdown
—
Weighted downside profile across tracked signal types
Signal types
—
Active performance categories being tracked
Signal accuracy trend
The realized success rate across the latest evaluation windows.
Accuracy is based on the deterministic decision engine, not OpenAI.
7D accuracy
+0.0%
0 signals, +0.00% average return
14D accuracy
+0.0%
0 signals, +0.00% average return
30D accuracy
+0.0%
0 signals, +0.00% average return
Regime comparison: win rate
This shows which market regimes the engine handles best and where thresholds tighten.
Win rate shows realized accuracy.
Signal archive
Old signals only. This page should never read like an active call surface.
Signals updated: waiting for first run
Loading proof records...
Performance by type
Performance data will populate once signal history accumulates.
Regime performance
Backtest results are grouped by market regime so the decision engine can use different thresholds in trend, range, and chop conditions.
Regime-specific backtest data will appear once enough history has accumulated.
Why this matters
- Shows the edge before asking for live workflow access.
- Keeps live signals separate from historical performance.
- Gives free users proof and a clear next step.
Next step
If the archive proves the edge, Signals is the next step: 2 to 3 daily system-generated calls with confidence, target, stop, timeframe, and reason.