Top 3 NHL Playoff Surprise Candidates (CURSD Score)
| Rank | Team | Points | xPoints | CURSD Score | Why |
|---|---|---|---|---|---|
| 1 | Vegas Golden Knights | 95 | 97.7 | -48 | Most cursed near the playoff line. Shootout variance and finishing rate both against them. |
| 2 | Edmonton Oilers | 93 | 96.7 | -46 | Running cold in every situation that flips outcomes. ~4 points stolen by variance. |
| 3 | Utah Mammoth | 92 | 102.5 | -44 | 10-point gap between actual and expected. Not just a bubble team. |
What the CURSD Score Measures
The CURSD Score is a composite metric built on xG differentials, Pythagorean expectation, close-game record, shootout outcomes, and one-goal game variance. It asks a simple question: which teams performed above their underlying numbers, and which performed below them?
NHL points standings are a useful fiction. They tell you who made the playoffs, but not why, and not whether it will hold up over a best-of-seven. The CURSD Score strips away the noise.
The Cursed: Built for More
The Vegas Golden Knights finished with 95 points and a CURSD Score of -48, the most cursed team anywhere near the playoff line. Their expected points total sits at 97.7, meaning the underlying performance was there. Shootout variance, close-game misfortune, and a finishing rate that underperformed expected goals all contributed. Vegas played like a stronger team than the standings reflect.
Edmonton posted 93 points with an xPts of 96.7 and a CURSD Score of -46. The underlying numbers suggest they should have roughly four more points. A team that generates that level of expected production and still lands at -46 has been running cold in exactly the situations that flip outcomes.
Utah Mammoth is the most extreme case. Their 92 points came with an xPts of 102.5, a gap of more than ten points. A CURSD Score of -44 reflects a team that played better hockey than their point total shows. The data suggests they are not just a bubble team.
Ottawa (99 pts, xPts 103.6, -31) and Tampa Bay (106 pts, xPts 113.0, -23) round out the cursed side. Tampa's gap is worth flagging: a team that should have 113 expected points and finished with 106 is an underrated contender. The market may be overvaluing the raw point totals.
The Blessed: Regression Risk
| Rank | Team | Points | xPoints | CURSD Score | Risk |
|---|---|---|---|---|---|
| 1 | Buffalo Sabres | 109 | 108.6 | +68 | Most blessed in the NHL. Close games + shootouts consistently their way. |
| 2 | Montreal Canadiens | 106 | 101.6 | +61 | 5-point overperformance, massive luck signal. |
| 3 | Dallas Stars | 112 | 111.5 | +57 | Strong team, but margins have leaned their way more than expected. |
Buffalo leads the NHL in luck. The +68 CURSD Score against 109 actual points and 108.6 expected looks like a tight gap on the surface, but the score reflects how consistently they won situations that could have gone either way. Close games, shootouts, finishing above xG rates. The Sabres are a legitimately good team. They are also a team that variance like this typically regresses against in a best-of-seven.
Montreal is the more striking case. The Canadiens posted 106 points with an xPts of 101.6 and a CURSD Score of +61. Nearly five points of overperformance, packaged inside a very large luck signal. The underlying numbers say this is a good team. The score says they have been running hot at the right moments. Both things can be true, and the second one tends to normalize under playoff pressure.
Dallas (112 pts, xPts 111.5, +57) shows a similar pattern. Boston (100 pts, xPts 100.1, +50) sits at nearly dead-even on expected vs. actual points, but the +50 CURSD Score suggests their underlying numbers are not as dominant as a 100-point team might imply.
What This Means for the 2026 Stanley Cup Bracket
The CURSD Score is not a prediction model. It is a signal about where a team's actual performance stands relative to their process. Historically, the teams with the largest positive luck signals in the regular season tend to underperform expectations in the playoffs, and the most cursed teams tend to outperform theirs.
The underlying numbers say Vegas, Edmonton, and Utah enter the bracket with more room to run than their seeds suggest. Buffalo and Montreal are carrying weight that does not show up on the standings page, but will show up in a series.
Methodology
The CURSD Score for NHL teams combines six signals: Pythagorean gap (goals for/against vs actual points), xG finishing rate, xGA defensive rate, close-game record, shootout net, and OTL rate. Each signal is z-scored independently, weighted, and scaled to a -100 to +100 range. Full methodology at cursd.com/methodology.