Mammoth hold off the Rangers

Utah has the cleaner numbers in this matchup. The model prices the Mammoth at 56% win probability with a projected 2.3–2.6 goal outcome—a narrow margin that reflects Utah's underlying edge across shot quality and territorial control. At 5v5, the Mammoth generate 52.0% expected goals to the Rangers' 48.0%, and their 53.0% Corsi rate suggests they'll control most of the hockey. When a team dominates possession and shot quality, it eventually shows up on the scoreboard.

But Utah's recent results don't match their underlying dominance. The Mammoth are -12 in goal differential over their last 10 games, a collapse that questions whether these metrics can be trusted this late in a season shaped by variance and exhaustion. That's the tension embedded in this pick: Utah has earned the model's lean through consistent process, yet their form is conspicuously broken.

Goaltending adds a layer of uncertainty. K. Vejmelka is on a back-to-back, which typically signals fatigue or a backup start if Utah chooses to rest him—either way, a disadvantage at net. I. Shesterkin faces the same burden on the Rangers' side, meaning neither team can reliably count on their starter's best performance.

The player matchups tip toward Utah's advantage. Dylan Guenther has been the Mammoth's engine, generating 1.02 xG/60 with 28 goals and 15 assists; he's delivered 3 high-danger chances and is shooting at 16.8%, well above league baseline. The Rangers' best offensive engine, Adam Sykora, generates 0.91 xG/60 with 3 goals and 1 assist—a stark gap in volume and finishing. Gabe Perreault (8 goals, 11 assists, 13.1% shooting) provides secondary depth, but the Rangers still trail Utah's forward advantage.

The Mammoth's edge in xG and Corsi should be decisive against a Rangers team that hasn't generated the underlying play to keep pace. The model is right to favor Utah at 56%. The risk is execution: Utah's recent form suggests they've misfired on these processes, and tired goaltending compounds the uncertainty. But when the process is this clean and the matchup this lopsided, the model's lean holds up.