The Bruins held 57 percent of shot attempts — 31 to 23 — yet expected goals said the game was a toss-up. They lost 4-3.

The scoreline overstates Florida's edge. Expected goals split 3.57–3.73, nearly even. Boston's shot volume didn't translate to dangerous chances; the Bruins averaged 0.057 xG per attempt while the Panthers managed 0.085. High-danger looks went 7–6 Florida's way.

S. Bobrovsky stopped 28 of 31 (0.903 SV%), outperforming his expected level by 0.57 goals. J. Swayman stopped 19 of 23 (0.826 SV%), a save rate that made every shot costly.

Boston's power play was dominant — 2 goals on 13 shots (0.82 xG). But five-on-five, the Panthers prevailed. Both teams converted rushes at nearly the same rate (Boston 2-of-26, Florida 3-of-24). High-danger finishing separated them: Florida 3-of-6 (50%), Boston 2-of-7.

In the third, Carter Verhaeghe's wrist shot from the high slot at 19:34 put the Panthers ahead.

Three Stars:

• D. Pastrnak (BOS) — 2A in 23:05
• B. Marchand (FLA) — 2A in 19:38
• P. Zacha (BOS) — 1G, 1A in 19:10

This is a genuine coin flip: the model projects 2.6–2.6 with the Bruins at exactly 50% win probability. One team dominates the underlying play. The other dominates the scoreboard. Neither advantage is large enough to guarantee victory.

Florida's edge is territorial and measurable. At 52.0% Corsi and 49.0% expected goals percentage, the Panthers generate volume from high-leverage positions. This means extended offensive pressure—stretches where Boston spends more time defending than attacking, wearing down in retreat. Territorial control of this magnitude usually converts to goals if the shooting percentage holds.

Boston's edge is precarious. At 47.0% xG%, they're being outshot and out-chanced relative to play quality. What keeps them competitive is a 1.023 PDO—combined shooting and save percentage running at the edge of plausibility. This will eventually regress. The question is whether it happens tonight.

For Florida, Sam Bennett (0.97 xG/60, 18G, six high-danger goals, 13.1% shooting) generates elite volume at a sustainable rate. He's not riding variance; he's executing. Brad Marchand (17G, 19.5% shooting, three HD goals) operates at a riskier level—his shooting percentage is elite, which means either he's genuinely exceptional or he's borrowed against next month.

Viktor Arvidsson leads Boston at 1.04 xG/60 and 18 goals, but only two come from high-danger areas. He's a volume scorer working at peak efficiency, vulnerable if that efficiency retreats. Tanner Jeannot (0.75 xG/60, 6G) scores at a rate his underlying metrics don't support. If Boston's PDO normalizes, Jeannot normalizes with it.

Both teams register 0% of goals from rebounds and 0% from high-danger areas—an anomaly suggesting either classification inconsistency or a genuine team construction that avoids those scoring sources.

Defensively, vulnerability runs both directions. Brad Marchand is Florida's weak link; Elias Lindholm is Boston's. If the Panthers establish territorial dominance—as the metrics suggest—Lindholm becomes exposed. If Boston forces pace and chaos, Marchand becomes a liability.

The deciding factor is identity. Florida has the underlying advantage; Boston has the scoreboard advantage. Whoever enforces their style for 60 minutes wins. The model, correctly, cannot choose.