The Penguins are the model's lean over the Jets — with PDO regression in play

The model projects Pittsburgh at 56% win probability, expecting the Penguins to score 3.5 goals against Winnipeg's 3.1. That lean sits on one fact: Pittsburgh controls the puck quality.

At 5v5, the Penguins generate expected goals at 62.2% to Winnipeg's 50.3%. This gap translates to better scoring chances and more efficient shooting. That advantage should carry through the match, even if the goal totals stay close.

Pittsburgh's territorial grip is real. The Penguins hold 52.4% of shot attempts versus Winnipeg's 42.7%, the kind of play-possession split that normally locks teams down and starves the opponent of high-danger looks.

A second edge lives in Pittsburgh's finishing. The Penguins score at 14% on chances from the slot (5 goals on 35 attempts this season). Winnipeg's role becomes damage prevention at the hash marks.

The goaltending matchup cuts both ways. S. Skinner starts for Winnipeg on no rest—the second night of a back-to-back—carrying a +1.2 advantage in expected goals saved (2 games, 90.8% save percentage) into what may be a tired performance. Pittsburgh counters with A. Silovs, running at -0.7 expected goals saved (2 games, 88.6% save percentage), opening the door for a hot start if Pittsburgh sustains early pressure.

Variance looms large. Pittsburgh's luck is running hot (1.071), suggesting regression is built in. Winnipeg sits at 0.984, a cold patch that typically normalizes upward. If both teams drift toward average, Winnipeg closes the gap the model projects.

The model's lean favors Pittsburgh on quality and volume. Skinner's back-to-back status is the one thread that could unravel it.