1) What the (public) models say — collected & averaged
Many reputable prediction outlets ran algorithmic / AI previews for NY Islanders @ Dallas Stars (Nov 18, 2025, American Airlines Center). A lot of the heavy-hitters (SportsLine, BetQL) keep their full simulation outputs behind paywalls, but several public / syndicated computer-model previews do publish projected final scores and totals:
-
AM1300 / Ticket760 (computer model): Stars 4 — Islanders 3; model also projects ~6.3 total goals (so “Over 6” lean).
-
Leans.ai (their “Remi” AI): publishes an AI preview (over/under ~6.0–6.5 and favors Dallas in pricing; best-prop angle Heiskanen SOG). (no explicit final-score in free preview, but the implied scoring + edge is similar).
-
SportsLine: shows their simulation framework and market/projected lines (spread/moneyline/OU around DAL -1.5 / O/U 6–6.5); full score sims are subscribers-only but the public outputs imply a one-goal Dallas win and an OU around 6–6.5.
-
Pickswise / Docsports / SportsInteraction: expert + model writeups that converge on Dallas (ML) and an O/U ~6 and similar one-goal margin narratives (public previews lean Stars).
Averaging the explicit public score outputs (the only widely-published explicit final-score projections I could find were the syndicated computer-models that published 4–3 Stars) gives an average model prediction ≈ Dallas 4 — New York 3 (total ≈ 7 by summing, but the models themselves stated ~6.0–6.5 totals; the explicit score predictions most often published were 4-3 with a modeled total ≈ 6.0–6.3).
2) Key recent news / injury checks (important situational inputs)
-
ESPN / SportsLine injury reports show Dallas has some lineup questions: Ilya Lyubushkin — questionable, Thomas Harley — out; other longer-term/IR statuses (Jamie Benn listed IR-LT elsewhere). SportsLine and ESPN both list these items in their pregame notes. Those absences affect Dallas’ defensive depth.
-
Form / momentum: Dallas has been rolling recently (multi-game winning streak; recent 5-1 win and Jason Robertson hat trick referenced in previews). Isles are coming off a 4-1 loss to Colorado and have been rotating lines heavily. That favors a Dallas edge at home.
3) Team fundamentals / Pythagorean check & SOS
I used publicly reported season totals (Hockey-Reference / team stat pages) to run a quick Pythagorean expected-win calculation (simple, transparent approach using GF & GA totals and exponent = 2):
Sources: Hockey-Reference / team pages show (season-to-date, at time of the previews)
-
Dallas Stars (through this sample date) — Goals For 62, Goals Against 53. (GF/GA → strong offense + good defense).
-
New York Islanders — Goals For 61, Goals Against 60. (more marginal GD).
Pythagorean expected win% (GF²/(GF²+GA²)):
-
Dallas ≈ 57.8% expected (pythag).
-
Islanders ≈ 50.8% expected (pythag).
(Interpretation: on goals-alone, Dallas projects as the stronger side — roughly mid-50s win probability before adjusting for matchup specifics). (calculation based on the season totals above).
SOS: hockey-reference season pages show Islanders’ SOS is tougher (Islanders SOS ~ +0.21, top-5), while Dallas SOS ~ -0.07 (weaker schedule). That means part of the Islanders’ numbers are earned against stronger opposition (so the Islanders’ raw GA looks worse but vs tougher opponents), and Dallas’ numbers may be slightly inflated by an easier schedule. I use this to temper the Pythagorean gap — it narrows the practical edge but doesn’t erase it.
4) Other external factors I included
-
Starting goalie situation: Sports previews flagged Jake Oettinger slated for Dallas starts in the writeups (and he’s been key). Goaltending matchups matter — SportsLine/Pickswise notes Oettinger in net for Dallas. That nudges the edge toward Dallas.
-
Injuries/defense depth: Harley out and Lyubushkin questionable (Dallas D depth impacted) — small push back toward the Isles being able to generate chances.
-
Rest / schedule: Both teams are mid-week; Dallas is home and riding momentum — home advantage and momentum favor Dallas.
