The Sharks won a game they had no business winning. San Jose edged Utah 3–2 in overtime despite the Mammoth generating 3.25 expected goals to their 2.09—a 61–39 split at 5v5 that should have favored the home team. Goaltending theft and surgical finishing made the difference.

Utah controlled play by every measure: 26 shot attempts to San Jose's 22, seven high-danger chances to two. The Sharks converted two goals from low-danger areas. They were lethal on the rush, scoring three goals on 13 chances. Utah scored twice on 25 rush attempts—a stark efficiency gap.

Y. Askarov stopped 1.25 goals above expected. Utah couldn't convert on any of its seven high-danger opportunities, a drought unlikely to repeat. The Sharks' goaltending and their ability to exploit mistakes compensated for being thoroughly outplayed in both volume and quality of chances.

Macklin Celebrini finished the game at 02:52 of overtime with a wrist shot from the circles. He recorded three goals in 19:41, a burst that exemplified San Jose's night: every shot counted.

Models project a UTA 2.6-2.5 SJS game with the Mammoth at 52% win probability — essentially a pick'em. San Jose owns recent form, but Utah controls the puck; both edges feel real.

The Sharks are 6-3-1 over their last 10, the better recent record. Utah's underwater at -5 goal differential in the same span. But San Jose's hot streak has masked underlying play. Utah leads 5v5 xG% 52.0% to 48.0% and drives Corsi at 53.0%—the territorial tilt favors possession play, not results.

Utah's issue isn't play quality, it's execution. Dylan Guenther (28G 15A, 1.02 xG/60) generates chances at an elite rate but converts at 16.8%, a sustainable mark for this volume. Daniil But's 5.6% shooting (2G 2A from 1.01 xG/60) sits well below expectation and should improve eventually, but San Jose shouldn't bank on correction arriving tonight.

For San Jose, Pavol Regenda (5G 1A, 1.10 xG/60) is marked inactive—a significant hit to depth. Igor Chernyshov (6G 8A, 0.91 xG/60, 15.0% shooting) absorbs that load, though his stats suggest he's outperforming chances. That 15.0% won't persist; regression is a when, not an if.

Utah's Maveric Lamoureux (the team's defensive weak link) faces pressure from a suddenly lethal San Jose. Conversely, Chernyshov's own defensive liability—despite offensive production—gives Utah an opening. PDO favors neither team (UTA 1.003, SJS 0.998), so luck isn't explaining the result gap.

The models see what the records miss: Utah's underlying play—52.0% xG—suggests the Mammoth should be winning more. San Jose's 6-3-1 run is real, but Utah's possession edge, if converted at historical rates, flips this into an away win. That's why the projection sits so tight: Utah's depth of play justifies favorites' odds, but San Jose's momentum is recent and measurable.

This is a true flip. Take the underlying advantage or the hot streak. Either case is defensible.