We read 8 NHL preseason games: every home team won — and 3 did it with fewer shots

We read the 8 games ESPN listed as Final when we pulled the scoreboard on the morning of September 22, 2026 — all of them played Monday, September 21, early in the NHL preseason. Eight games, eight home-team wins. Across those 8 games: 398 total shots on net, 37 goals, a combined shooting percentage of 9.3%. In 5 of the 8, the home team also outshot its opponent and won. In the other three, the home team had fewer shots and still won — each time because it scored on a bigger share of them: Colorado Avalanche at Winnipeg Jets — Colorado outshot Winnipeg 32 to 19. Winnipeg won 3–1. Colorado scored on 1 of its 32 shots (3.1%); Winnipeg scored on 3 of 19 (15.8%). Minnesota Wild at Chicago Blackhawks — Minnesota outshot Chicago 22 to 20. Chicago won 4–3. Minnesota scored on 3 of 22 shots (13.6%); Chicago scored on 4 of 20 (20.0%). St. Louis Blues at Dallas Stars — St. Louis outshot Dallas 31 to 27. Dallas won 2–1. St. Louis scored on 1 of 31 shots (3.2%); Dallas scored on 2 of 27 (7.4%). In each of the three wins on fewer shots, the home team scored on a bigger share of its shots — by at least four percentage points. Power play — the stretch after a penalty when one team skates a man up: the 16 teams drew 47 such chances and scored five times, a 10.6% rate. Five of the 8 games ended with neither team scoring on the power play. The three games that produced any power-play goals: Washington vs. Philadelphia (2 goals on 6 chances), Pittsburgh vs. Buffalo (2 goals on 5 chances), Montreal vs. Ottawa (1 goal on 10 chances).

The following is analysis, not fact. The home-team sweep in 8 games is an extreme outcome — in the NHL regular season, home teams win roughly 53% to 55% of games ^mmodeled, so an 8-for-8 run in any sample is unusual. Three explanations fit, and these 8 games alone cannot separate them ^sspeculative. Away teams tend to travel in early preseason with more developmental and tryout players than they send in October ^sspeculative. The visiting club may have less at stake. Or the sweep is simple variance in a small sample. The conversion gap in the three wins on fewer shots is more interpretable. Colorado put up 32 shots to Winnipeg's 19 and scored once. Winnipeg scored three times from 19 attempts. The shots-total column would have pointed to Colorado; the goals column did not. A gap that large tends to narrow over an 82-game season as shooting percentages even out ^mmodeled, but a large gap is itself common early in preseason, when two goaltenders often split each game and scoring lines haven't built their regular-season chemistry ^sspeculative. The power-play rate of 10.6% is well below the NHL's recent regular-season average of roughly 20% ^mmodeled. In 5 of these 8 games, the power play scored nothing across every combined opportunity. That is consistent with teams deploying second- and third-unit personnel on the power play rather than their established first units — rehearsal that may trade execution sharpness for developmental ice time ^sspeculative.

Preseason results carry little demonstrated predictive value for the regular season ^mmodeled. Coaches typically use early preseason to evaluate depth and fringe players, keep veterans healthy, and experiment with combinations that will not appear in October ^sspeculative. Colorado's 32-shot effort against Winnipeg likely included substantial ice time for players who will not open the regular season on the roster ^sspeculative. The home-team sweep and the power-play collapse are consistent with a slate designed for development, not for outcomes that travel to the standings ^sspeculative.