Parlay Correlation Lab
Parlays don't raise your win probability — each added leg lowers it and compounds the book's hold. The one place value can hide is correlation: books price legs as independent, so when outcomes actually move together the parlay price is wrong. This lab measures that correlation on NBA markets and tests it honestly.
Research and informational use only. Model outputs are probabilistic, not guarantees; historical performance does not guarantee future results.
Outcome correlation
Loading correlation matrix…
Independence mispricing
Scanning the board…
Parlay backtest
Bets one fixed 2-leg same-game combo on every game in range — leak-free, it just reports what happened. Leg B is the game total.
Results
Pick a combo and hit run backtest. Most naive parlays lose — that's the point.