The Brooklyn Nets finished the 2025-26 NBA season with a Luck Index of -61, the worst mark across both the NBA and NHL. Their actual point differential sat nearly six points per game below what the underlying shot quality, turnover rates, and opponent shooting variance said it should have been. The standings recorded one version of the Nets' season. The data recorded a meaningfully different one.
Due a Correction
Brooklyn Nets (NBA) — Luck Index: -61, xPoints gap: -5.80. Brooklyn's gap was the largest negative deviation in the league this season. Their opponent three-point shooting percentage ran well above league average for months at a time, a classic marker of bad variance rather than bad defense. Close-game finishing was equally brutal. The Nets lost 19 games decided by four points or fewer, a clip that historically reverts hard. With their season now over, the actionable question shifts to how the market prices Brooklyn's core heading into the offseason and next October's futures.
Washington Wizards (NBA) — Luck Index: -56, xPoints gap: -6.10. Washington actually posted the largest raw xPoints gap of any team on this list at -6.10, meaning their results understated their process by the widest margin in the league. Their defensive efficiency metrics graded closer to a play-in team than to their actual record. Variance hit them hardest in the second and third quarters, where opponent shooting ran nearly two percentage points above expected. The Wizards were not secretly good. They were, however, secretly less bad than their record.
Utah Jazz (NBA) — Luck Index: -49, xPoints gap: -5.60. Utah's gap was smaller but still substantial. Their fourth-quarter net rating was positive in games they lost by single digits, a pattern that typically screams noise. Bookmakers pricing Jazz futures for next season may want to consider the 5.60-point cushion between what Utah showed and what Utah was.
Sometimes a team's record is just its record. Three teams finishing this far below expectation is not one of those times.
Living on Borrowed Luck
Montréal Canadiens (NHL) — Luck Index: 71, xPoints gap: +4.40. Montréal's Luck Index of 71 was the highest of any team in either league. Their save percentage ran well above the goaltending talent on the roster, and their power play converted at an unsustainable rate for three consecutive months. The 4.40 xPoints gap means roughly two extra standings points per month that the underlying play did not earn. Montréal is in the playoffs now, and series outcomes can flip on a single game's shooting variance, but the regular season evidence is clear: the Canadiens' point total outpaced their process.
Buffalo Sabres (NHL) — Luck Index: 66, xPoints gap: +0.40. Buffalo's case is unusual. The Luck Index of 66 is driven almost entirely by sequencing and score effects rather than a massive xPoints gap. Their 0.40 gap is tiny, but they won a disproportionate share of one-goal games and benefited from opponent empty-net mismanagement at a rate that strains credulity. The wins were real. Whether the pattern that produced them is repeatable is a different conversation.
Oklahoma City Thunder (NBA) — Luck Index: 60, xPoints gap: +5.90. OKC's 5.90 xPoints gap was the largest positive deviation in the NBA. Their clutch shooting was historically good, and opponent free-throw rates in close games ran below average. The Thunder were excellent this season by any measure, but the margin between excellent and historically dominant was filled in by variance. As they move deeper into the playoffs, the market has priced them as a title favorite. The data agrees they are very good. It disagrees about the margin.
The Regression Window
In the NBA and NHL playoffs, regression does not operate on the comfortable 10-to-15 game timelines of a regular season. A seven-game series can see an entire season's worth of luck correction compressed into a single overtime period. Teams carrying large positive Luck Index scores into the postseason have historically seen their win probability dip by the second round. Teams carrying large negative scores, of course, are already on the couch. The information is the information. What anyone does with it is their own business.