Basketball is the most predictable major sport. Point differentials predict playoff seeding better than win-loss records. But even in the NBA, sequencing luck matters: a team that loses every close game but blows out weaker opponents will underperform their Pythagorean expectation. The NBA CURSD Luck Score (CLS) uses 7 base signals plus a scaled injury burden factor: garbage-time-adjusted Pythagorean (margins capped at 15 to strip blowout noise, 25%), close-game luck as a continuous W/L + net margin signal (17%), offensive 3PT luck (11%), offensive 2PT luck (7%), defensive 3PT luck (11%), defensive 2PT luck (7%), and opponent free throw luck (7%, the purest luck signal in basketball since defense has zero influence on FT%). Shooting is decomposed into 3PT/2PT/FT components because three-point variance is the highest (purest luck), while 2PT shooting is more skill-heavy.
53 wins from a +145 point differential. One of the luckiest teams in the NBA - winning tight games and getting favorable bounces at an unsustainable rate.
60 wins from a +669 point differential. One of the luckiest teams in the NBA - winning tight games and getting favorable bounces at an unsustainable rate.
62 wins from a +681 point differential. One of the luckiest teams in the NBA - winning tight games and getting favorable bounces at an unsustainable rate.
Frequently Asked Questions
CLS Evolution
How each team's CLS has changed throughout the season
Above zero = lucky. Below zero = unlucky. Scale: -100 (cursed) to +100 (blessed).
NBA Luck by Signal
The CLS combines several signals. Pick one to see which NBA teams are getting robbed (or blessed) on that signal individually.
Pythagorean Delta
Expected wins from goal/point differential vs actual