The sharpest question in this week's breakout-or-fluke debates is the one about Buffalo, because the team and its best scorer point in opposite directions. The Sabres' drought-breaking season was, by the underlying numbers, substantially borrowed: a 49 percent team that scored 283 goals against roughly 25 goals of finishing luck and got another 25 from its netminders. Regression is coming for some of that, and I have written the warning already. But inside the luck, one thing was structural, and it is the thing that decides how hard the regression bites. Tage Thompson's 40 goals were real, and proving that distinction matters more to Buffalo's season than anything else on the roster.
| Metric | Number | The read |
|---|---|---|
| Tage Thompson | 40 goals and 81 points | On a 57% on-ice share |
| Rasmus Dahlin | 74 points | A plus-13 swing from the blue line |
| Team expected-goals share (all situations) | 49% | Below break-even |
| Goals scored | 283 | Roughly 25 above expectation |
The distinction between a lucky team and a lucky player is where this gets interesting. Buffalo as a group out-finished its chances, the classic signature of a season that does not repeat. Thompson individually did what he has now done across multiple seasons: converted at an elite rate on heavy chance volume while driving a 57 percent on-ice share, eight points above his team's baseline. One of those is variance. The other is a skill set, a 6-foot-6 frame with a wrist shot that beats goalies from distances the league's models barely credit, and skills do not regress on schedule.
The repeat-season evidence
The fluke hunt usually asks whether a scorer has done it once or twice, and Thompson keeps answering. This is not his first 40-goal season, his shooting talent has survived every percentage correction thrown at it, and his chance creation has grown alongside the conversion. At 28, in his physical prime, with the largest release window in hockey, he profiles as exactly the kind of shooter whose percentages run permanently hot because the shots themselves are better than average, not luckier. The model that flags him as regression fodder is the same model his career keeps embarrassing.
Why he matters double this season
Here is the structural point for Buffalo: when a team is due to give back 30-odd goals of collective luck, the players whose production is real are the levees. Rasmus Dahlin's plus-13 swing holds the blue line side of that equation, and Thompson holds the scoring side. If both sustain, the Sabres' fall is cushioned to something survivable, a step back that still makes the playoffs in a soft division. If either wobbles, the regression compounds instead of averaging out, and the drought-breaker year starts looking like a one-off. Buffalo's season is not really a bet on its luck repeating. It is a bet on its two real stars being real enough to absorb the correction.
The honest hedge
What could genuinely drag Thompson down is not his shot but his context. A 49 percent team that loses its percentage cushion generates fewer leads, fewer power plays and worse score states, and even structural scorers produce less inside a losing environment. His health record has wobbled before, and 40 goals always requires the 75-plus games nobody can promise. A 33-goal Thompson season on a fading team is entirely plausible and would prove nothing about him. The claim is narrower and stronger: whatever Buffalo gives back, his production is the part built on skill rather than sequencing, and the difference will be visible by December.
What Buffalo should do with the truth
The practical implication for the front office is about honesty in roster-building. A team carrying known regression should spend its summer improving the five-on-five structure underneath its stars, the shot suppression and depth scoring that a 49 percent share exposes, rather than assuming 283 goals is the new normal. Thompson and Dahlin buy the Sabres time to do that work; they do not do it for them. The teams that survive a luck correction are the ones that treated the lucky year as a head start, and Buffalo's whole trajectory rests on which interpretation its management chooses.
The shot science, briefly
It is worth pausing on why Thompson specifically breaks the regression models, because the mechanism is concrete rather than mystical. Expected-goals frameworks grade shots by location, type and context, but they cannot see release point, and release point is Thompson's entire edge. A 6-foot-6 frame shooting off a long reach changes the geometry a goaltender faces: the puck leaves from coordinates the positional textbook does not cover, arrives earlier than the depth cues suggest, and screens itself behind a wingspan that hides the blade until release. The models call his mid-range wrist shots average chances. Goaltenders who face them call them something else, and the gap between those two assessments is where his percentage lives, year after year.
This is the same reason the fluke hunt keeps misfiling him. Regression analysis works by comparing a shooter's conversion against the league average for similar shots, and it flags anyone running hot as due for correction. That logic holds for players whose shots genuinely are average, and it fails systematically for the handful whose physical tools make their shots categorically different from the comparison set. Thompson has now sustained an elevated conversion across enough seasons and enough volume that the honest statistical conclusion is not that he is lucky, but that the model's shot categories are too coarse to price him. Analysts who have watched this movie with other outlier shooters know how it ends: the model keeps predicting the fall, and the fall keeps not arriving.
None of which makes him immortal; it makes him legible. His goals will track his health and his chance volume, not a percentage correction, and that is exactly the kind of production a team can plan around. Buffalo's front office should treat the distinction as operational: the team's luck needs replacing, and its sniper does not.
Where this lands
Buffalo's breakthrough was half real and half borrowed, and the fluke debates keep failing to separate the halves. The borrowed part, the team-wide finishing and the goaltending surge, follows the same gravity it always does. The real part is a 6-foot-6 sniper with a repeat 40-goal pedigree and a 57 percent share, and betting against that specific skill set has been a losing trade for years. When the regression arrives, and it will, watch what stands after it recedes. The prediction here is that Tage Thompson still will be, and that he is the reason the Sabres' step back stops short of a collapse.