The War Room
Every screen — what it does, how to use it
Seven screens, one job each. Here's what to open and when. A normal day flows Scanner → RaceDay → Desk → (settle) → Scorecard, with the Harness as the gate that decides whether any of it can go live.
Your home base — bankroll, open paper bets, and the day’s strongest edges at a glance.
How: Glance here first to see where you stand and what’s live.
Sweeps every ingested card and surfaces overlays automatically — the mispriced horses, ranked by edge.
How: Your daily starting point. Work the top of the list down.
Builds exacta / trifecta / superfecta tickets for chaos races where favorites are weak. Separate scorecard.
How: Use sparingly, only when a race screams chaos. Treat as lottery tickets with an edge.
Today’s cards, race by race, with the model’s fair odds and overlays in context.
How: Drill from a card into a race to see the full field before you stage a bet.
Your working ticket — stage overlays, set stakes, and place (paper) bets into the ledger.
How: Where the decision becomes a bet. Confirm the overlay and the stake, then fire.
The honest ledger — CLV and ROI by generator, win vs exotics, with full provenance. The truth.
How: Review after races settle. CLV trending up is the early sign you’re winning.
The validation gate — walk-forward verdict (EDGE / NO_EDGE) and the live-betting switch.
How: Check before trusting the model at scale. It controls whether live mode can even arm.
Behind the screens: the data pipeline
The screens only show what the pipeline feeds them. It runs from the command line (or on a schedule), and each step has a self-test so nothing silently breaks: