The Ducks are the model's lean over the Bruins — with PDO regression in play

The model projects Boston 3.1 and Anaheim 3.5, and it backs the Ducks to win. The lean holds, but it rests on a narrower edge than the score suggests, and Boston's results may be running ahead of its play.

Anaheim's advantage shows up in expected goals, the model's read on scoring chances weighted by how dangerous they are. At five-on-five this season, the Ducks take 49.3% of that share. Boston does better by the simpler count of shot attempts, holding 49.5% of the total. Both teams sit just under even, so neither has taken control of the territory fight.

The bigger concern for Boston is luck. PDO adds a team's shooting percentage to its save percentage, and it runs well above average when a club has been getting bounces. The Bruins sit at 1.076, a level that usually drifts back toward normal. That is why the draft flags them as the regression watch.

Goaltending is where the model finds its margin. J. Swayman has posted a 93.0% save percentage over six games. His work in net has also run 6.8 goals above what the shots should have allowed this season. L. Dostal has an 89.2% save percentage over five games, and on goals saved above expected sits at -0.0. That means the shots he has faced have been neither better nor worse than the model predicted.

The lean holds: the model gives Anaheim a 56% chance. Boston's best arguments are its shot-attempt share and the goalie edge, but the luck behind that 1.076 is the thing to watch.

Data: NHL API, Backcheck xG model