CHETBETSDATA · DRIVEN · DOMINANCEAI Capper Report
Historical model baseline: Sep 17, 8:55 AM PT
Live rows use the current schedule and market feed. They are research observations, not published official positions.
Market checked Sep 30, 5:54 PM PT · refreshes every 15 minutes
Green = favorable model discrepancy; red = unfavorable. Sort uses probability points for ML and scoring points for other markets. Estimates are experimental, not calibrated betting recommendations.
How the projection is calculated
10,000 seeded Monte Carlo simulations per matchup. Up to 30 completed games per team, recency weights, prior-season results weighted at 45%, and eight league-average games of shrinkage. Scoring offense is matched to opponent defense. Home advantage comes from observed results and is capped. MLB/NHL use a Poisson scoring baseline; football/basketball/soccer use score simulations with observed variance. Ties are retained where the sport permits them.
ML is a model fair price. This baseline does not yet adjust for confirmed lineups, injuries, starters, surface, weight cuts, pace or weather and has no verified predictive track record. Missing sportsbook quotes are never manufactured.
Edition coverage and sources
138 teams · 138 team histories · 288 games with archived DraftKings markets. ATS/ROI/trends use the last 10 completed games per team; modeling uses up to 30. Previous-season games can be included and all dates remain visible in source records.
Results and boxscores: ESPN. Opening and current/archived prices: DraftKings via ESPN. Team-total prices, player-prop closing lines and verified bettor identities are not supplied by this connection.