Live evidence command centre
Research dashboard
Tracks captured paper signals versus settled results across back tips, false-favourite lays and place alternatives. This is intentionally evidence-first: no strategy is promoted from paper data without sample size, ROI and drawdown guards.
Automated research decision
Railside Ratings uses this without you manually inspecting every signal.
- Lay mode remains paper-only: 16 settled signals at -25.35% ROI.
- Inverse back-favourite mode is being paper-tracked: 13 settled at -11.92% ROI.
Promotion guard status
109 settled · -26.08% ROI · £-48.35 paper P&L
- Need at least 250 settled paper signals before considering promotion.
- Live paper ROI must be positive before considering promotion.
Executable-price evidence
Separates manual Betfair Exchange, automated Betfair delayed API, and proxy/SP paper settlement.
Exchange capture status
Checks whether automated Betfair delayed snapshots are reaching open paper signals.
Cumulative P&L trend
Settled paper evidence only, grouped by race date.
Drawdown trend
Shows peak-to-trough damage in the live paper record.
Mode x price-source breakdown
Profitability must survive per-mode and per-source checks before any promotion.
Exchange matching diagnostics
Open proxy-priced signals and persisted Betfair capture failures. These are the first candidates for market/runner matching fixes.
Captured/open signal health
Today
Short-term live paper evidence
Last 7 days
Near-term validation window
Last 30 days
Complete persisted paper record
All time
Mode performance
lay drifting morning favourite
Mode performance
inverse lay back favourites
Mode performance
lay false favourites cycle filtered
Mode performance
lay false favourites
Mode performance
lay gb high draw top weight favourites
Mode performance
back tips
Mode performance
lay exchange slice candidates
Mode performance
lay exchange slice b candidates
Mode performance
false favourite each way alternatives
Recent captured signals
Latest 200 persisted paper signals. Use Paper for capture controls and detailed filtering.