Calgary won 5-3. Florida should have won 5-3.

The Panthers dominated expected goals at 58% (5.42–3.94), owned shot attempts at 55% (41–34), and tied high-danger chances 7-7. On paper, this was a Florida victory. On the scoreboard, it was Calgary's heist.

D. Cooley did the stealing. His 0.927 SV% meant he stopped 2.42 goals above expected — the swing that flipped the game. D. Tarasov's 0.879 SV% for Florida couldn't compete.

Where Calgary punched above its class: three goals came from low-danger spots, locations an NHL goaltender should routinely deny. On the rush, Calgary converted 2 of 22 chances; Florida converted 3 of 37. At the net front, neither team found rhythm — Florida crashed for 12 rebound attempts and scored on none; Calgary created 7 chances and converted one.

The Panthers' problem wasn't process. They underperformed their xG by 2.4 goals, which doesn't happen often to teams this efficient. Joel Farabee's snap from the wing at 18:57 of the third period gave Calgary the breathing room it needed.

N. Kadri led Calgary's finish: 1 goal and 2 assists in 18:59. R. Andersson added 2 assists in 22:59. For Florida, S. Reinhart recorded 2 assists in 21:20 — solid work that still wasn't enough.

Three Stars

• #91 N. Kadri (CGY) — 1G, 2A
• #4 R. Andersson (CGY) — 2A
• #13 S. Reinhart (FLA) — 2A

The model sees this as nearly a pick'em. It projects Calgary at 2.5 expected goals and Florida at 2.7, giving the Panthers a 54% win probability.

Calgary is the worry. The Flames are -5 in goal differential over their last ten games, suggesting recent play has been sloppy or unlucky or both. But there's a safety valve: their PDO of 0.985 is running cold. When a team's shot quality and goaltending both underperform luck, regression is nearly certain. Calgary should expect better results in similar games from here.

The season-long picture shows Florida with a structural advantage. The Panthers drive play at 52.0% Corsi while Calgary sits at 49.0%. That territorial edge translates to xG%: Florida 49.0%, Calgary 47.0%. None of this is overwhelming, but it's consistent. Florida wins close-range battles.

Goaltending could muddy things. S. Bobrovsky plays on a back-to-back, as does D. Cooley. Either starter may rest, or the real one takes the ice tired. That unpredictability is worth noting but hard to bet on.

The players tell the story:

Blake Coleman (Calgary) carries 0.89 xG per 60 minutes with 16 goals, 12 assists, and an 11.4% shooting rate—two of those goals came from high-danger chances. Matt Coronato (Calgary) is at 0.80 xG/60 with 10 goals and 16 assists, but his 6.6% shooting rate flags as a regression risk; he's likely to score more on similar volume.

Sam Bennett (Florida) leads at 0.97 xG/60 with 18 goals, 19 assists, and six high-danger tallies. His 13.1% shooting rate is sustainable. Brad Marchand (Florida) generates 0.93 xG/60 with 17 goals, nine assists, and three high-danger goals—but his 19.5% shooting rate is an outlier. Marchand won't stay this hot.

Both teams clock 0% on high-danger and rebound goals, which signals either exceptional goaltending or randomness so far. One or both rates will move toward the league mean.

The matchup hinges on whether Calgary's regression window opens before puck drop. If it does—if the Flames' luck follows their underlying play—they'll keep this tight despite Florida's structural edge. If not, Florida's territorial dominance will separate them. The model, correctly, won't pick a side.