NHL Market Audit: Selective Model Posts +5.13 Units

The nhl_selective_v1 model settled 12 spread picks between October 4 and October 10, 2026, finishing with 6 wins, 6 losses, and 0 pushes for a total return of +5.13 units. Despite a flat 0.5 win rate across 12 games, the model generated a positive return by targeting plus-money pucklines and high-payout spreads.

A week of results in the NHL represents a very small sample size, and our NHL model is currently under audit. An even record over six days neither proves long-term edge nor exposes systemic failure; it simply demonstrates how payout structure can yield positive returns even without a dominant win percentage.

Oct 9 Recorded Outcomes and On-Ice Execution

Three games on October 9 carried full scorelines in the tracking window, delivering a 1-2 record on the day but illustrating the impact of long-shot pucklines.

The lone cover came in Washington, where the Washington Capitals (-1.5) defeated the New York Rangers 3-1 at a +199 price to net +1.99 units. Forward Boone Jenner scored a goal on three shots, and Tom Wilson added a goal on three shots to secure the two-goal margin.

In Columbus, the Columbus Blue Jackets (-1.5) secured a 3-2 victory over the Pittsburgh Penguins, but failed to cover the 1.5 puckline (-1.00 unit). Defenseman Damon Severson posted a goal and an assist, Cole Sillinger recorded two assists, and Mathieu Olivier scored a goal on four shots for Columbus. Evgeni Malkin scored the solitary goal on three shots for Pittsburgh, keeping the final margin at one goal.

The biggest execution breakdown occurred in Winnipeg, where the Winnipeg Jets (-1.5) fell 5-1 to the Anaheim Ducks (-1.00 unit). Kyle Connor scored Winnipeg's lone goal. Anaheim's offensive output featured multi-point games from Judd Caulfield (1 goal, 1 assist), Mikael Granlund (1 goal, 1 assist), Tim Washe (1 goal, 1 assist), Jeff Malott (2 assists), and Beckett Sennecke (2 assists).

Margin Analysis and Plus-Money Spreads

The remaining nine plays on October 10 yielded a 5-4 record. Final scorelines were unrecorded for these contests in the tracking data, but the actual margins settled the spread lines.

Game Date	Matchup	Pick Line	Price	Actual Margin	Result Units
2026-10-09	WSH vs NYR	Home -1.50	+199	2	+1.99
2026-10-09	CBJ vs PIT	Home -1.50	+228	1	-1.00
2026-10-09	WPG vs ANA	Home -1.50	+223	-4	-1.00
2026-10-10	STL vs CBJ	Home -1.50	+223	2	+2.23
2026-10-10	CAL vs ANA	Away +1.50	+217	-2	+2.17
2026-10-10	BOS vs PHI	Home -1.50	+194	4	+1.94

The model's highest unit gains arrived via the St. Louis Blues (-1.5 at +223, margin of 2, +2.23 units) and the Anaheim Ducks (+1.5 at +217, margin of -2, +2.17 units). Additional covers were recorded by the Boston Bruins (-1.5 at +194, margin of 4, +1.94 units), the Edmonton Oilers (+1.5 at +167, margin of -2, +1.67 units), and the Carolina Hurricanes (+1.5 at +113, margin of -3, +1.13 units).

Conversely, favored lines failed to deliver. The Buffalo Sabres (+1.5 at -217) carried the highest confidence score in the set (0.70) but missed with an actual margin of -3 (-1.00 unit). Similarly, the New York Islanders (+1.5 at -159, confidence score 0.62) lost with a -2 margin (-1.00 unit). Florida (-1.5 at +208, margin -3) and Dallas (+1.5 at +156, margin 1) also dropped one unit each.

Sample Size Reality

Evaluating a model on 12 total plays is strictly a descriptive accounting exercise. A +5.13 unit outcome off a 6-6 record highlights how price selection impacts short-term returns, but drawing firm conclusions about nhl_selective_v1 requires a far broader dataset across a full season.