NHL Best Bets, Predictions & Expert Picks Tuesday, October 20, 2026

Analyse

The New York Islanders enter this home matchup at 4 and 3, while the Anaheim Ducks sit at 4 and 4. That record gap is small, and the model sees a competitive game rather than a clear mismatch. New York is projected to win 58.3% of the time, which gives the Islanders a modest advantage, but not enough to justify the listed moneyline as a model backed play. The spread tells the same story. The Islanders are priced at minus 1.5, while the model projects them at minus 0.46. That is a major difference in margin expectations. New York may be the more likely winner, but the model doesn't see a strong chance of the Islanders winning by multiple goals. Laying the puck line asks for more separation than the projection supports. The total is also too close to attack. The model projects 5.78 goals against a line of 5.5, but that gap produces no listed edge. There's a slight lean toward more scoring, yet not enough to turn it into a bet. New York is also without Calum Ritchie and Matias Maccelli, while the Ducks have no listed injuries. With both teams carrying similar records and no market offering value, this is a clean pass.

À contre-courant du public

The public may see New York at home and assume the Islanders should handle a Ducks team with a .500 record. But the model only gives New York a 58.3% win probability and projects less than a full goal of separation. That isn't enough support for the Islanders at minus 1.5.

Analyse

This is the most interesting matchup of the three because the records point in opposite directions from the home designation. The Boston Bruins are 7 and 2, the best record listed in this group, while the Dallas Stars are 4 and 3 at home. Still, the model gives Dallas a 60.1% win probability, suggesting the home ice advantage matters more here than the raw standings snapshot. That doesn't mean the Stars are a playable side. Dallas is listed at minus 1.5, but the model projects only a minus 0.56 margin. The difference is substantial. A Dallas win is more likely than a Boston win according to the model, but a two goal victory is not the expected outcome. The puck line requires too much from the favorite. The total is similarly tight. The model lands at 5.62 goals, just above the posted 5.5, and the listed edge remains zero. There isn't enough separation to support either side of the total. Boston's 7 and 2 start could attract attention, but the model hasn't converted that record into a betting edge on the Bruins. At the same time, Dallas carries the stronger projected win probability without offering enough value at the spread. This is a pass in every listed market.

À contre-courant du public

Casual bettors will likely gravitate toward Boston because a 7 and 2 record is the loudest number in this matchup. But the model still gives Dallas a 60.1% win probability. The problem is that Dallas is being asked to win by more than the projection supports, so neither the favorite nor the underdog offers value.

Analyse

The records create an obvious story. The Carolina Hurricanes are 7 and 2, while the Vancouver Canucks are 2 and 7. Carolina is clearly the stronger team by the standings provided, yet the model still doesn't find a bet at the available numbers. That matters because strong teams can be correctly favored without being profitable favorites. Carolina's main problem is the spread. The Hurricanes are listed at minus 1.5, but the model projects Vancouver at plus 1.28. That projection sits close to the full goal and a half being offered, so there isn't enough room to support either side. Carolina can win this game and still fail to cover, while Vancouver can lose and remain inside the number. The total is nearly perfectly aligned with the market. The model projects 6.49 goals against a line of 6.5, producing no edge. That makes an over or under decision especially thin. Vancouver's poor 2 and 7 record may push bettors toward Carolina or toward a higher scoring game, but the model has already accounted for the matchup and still finds no value. Mark Jankowski is out for Carolina, though the data does not provide a quantified impact from that absence. The confidence score is 72%, but that reflects the model's assessment, not a betting edge. Pass.

À contre-courant du public

The public will probably back Carolina because the Hurricanes are 7 and 2 and Vancouver is 2 and 7. That record gap is real, but the market already reflects it with Carolina laying 1.5 goals. The model sees Vancouver at plus 1.28, so the favorite isn't giving us enough margin to bet.