Market Board
Actionable market discrepancies
Bet Tracker
| Selection | Odds | Stake | Close | CLV | Result |
|---|
Analytics
Use the close as a quality check.
A winning bet can still be a bad decision. A losing bet can still have been well priced. Over a meaningful sample, consistently beating the closing market is a cleaner process signal than short-run win/loss results.
Old prices get penalized.
Prices older than the freshness threshold are flagged and can be hidden. This helps avoid “edges” caused only by one sportsbook lagging behind the rest of the market.
Public stats confirm; they do not guarantee.
Parlay Lab checks probable starters, ERA, WHIP, strikeout and walk control, games started, innings pitched, recent starts, team scoring, venue and weather. Small pitching samples are regressed toward neutral, while the strict builder requires reliable samples. Adjustments remain capped because much of this information is already reflected in the market price.
Settings
Bankroll
Position sizing
Cards show quarter-Kelly as a research reference, capped at 2 units by default. Kelly can overbet badly when the estimated probability is wrong, so it should never be treated as certainty.
Live feed
The Cloudflare Function reads the API key from a server-side secret named ODDS_API_KEY. The browser never receives the key.