The Sabres stole one. They lost the expected-goals battle 3.20–3.54 and won anyway.
The Sabres won a game the underlying metrics almost handed to Detroit. Buffalo overperformed their expected goals by 1.80 — an elite margin of luck or finishing — while Detroit's goaltending gave back what should have been an advantage.
Expected goals went 3.54–3.20 in Detroit's favor. Yet the Sabres scored five goals to four. C. Ellis stopped 24 of 28 shots (0.857 SV%), reaching near-expected performance. J. Gibson gave up 22 of 27 (0.815 SV%), finishing 1.80 goals worse than predicted — a decisive underperformance for a goalie in a one-goal game.
Detroit's case should have been stronger. They held 51% of shot attempts and won the high-danger battle nowhere — chances split 4–4. Yet they scored three goals on twenty-five rush chances, a potent conversion rate. Buffalo answered with four goals on twenty-nine chances, higher volume and higher efficiency. The Red Wings' shooting (50% on close looks) showed poise; the Sabres' rush scoring showed desperation that worked.
Mattias Samuelsson's snap shot from the high slot at 01:05 of overtime broke the tie, finishing a game the Red Wings should have found easier to control. The scoreline masked how much goaltending and rush execution, rather than five-on-five play, decided it.