Nashville Predators
Minnesota WildThe Dallas Stars finished third overall last season with 112 points, led the NHL in power-play goals, allowed the second-fewest goals in the league, and lost in the first round to the Minnesota Wild in six games. Somewhere in the middle of that series they went 254 consecutive minutes without scoring an even-strength goal.
That last number is not an aberration. It is the season, concentrated. At 5-on-5 the Stars generated 2.36 expected goals per 60 minutes, 22nd in the NHL, which is to say they created chances at even strength like a team fighting for a wild card. They were sixth at preventing them, at 2.29, and that is the whole act: a 51 percent team that suffocated opponents, scored on the power play, got 16 goals of goaltending above expected and 11 goals of finishing above expected, and converted a genuinely ordinary even-strength profile into the third-best record in hockey.
None of that is cheating. Special teams and goaltending are real skills and Dallas is excellent at both. But the playoffs shorten the power play, and when Minnesota stopped taking penalties the Stars had nowhere to go.
The summer made it worse, because it had to
Dallas went into the offseason with seven players at $8 million or more and no room. Jim Nill's line was that you get the calculator out and the numbers do not add up. What that meant in practice was selling.
On July 1 the Stars sent Mavrik Bourque and Ilya Lyubushkin to Nashville for a 2027 second-round pick and a 2028 third. Bourque is 24. He scored 20 goals and 41 points, and at 5-on-5 he was on the ice for 2.70 expected goals for and 2.23 against, a 55 percent share that was second among Dallas forwards behind only Jason Robertson. In all situations his share was 58 percent. Nashville immediately signed him for six years at $5.5 million.
So a team that finished 22nd in even-strength chance creation sold, for picks, one of the three forwards on its roster who was good at even-strength chance creation. Alexander Petrovic went to Florida, Michael Bunting left, and the only NHL addition was Joel Kiviranta at $1 million.
The other half of the cap story is Jason Robertson, who signed for one year at $12 million out of arbitration in late July. That deal is not an overpay so much as a delaying tactic: it keeps him for 2026-27, pushes the long-term decision into a summer when the cap rises and Tyler Seguin's $9.85 million expires, and leaves Dallas with roughly $390,000 in space right now. Robertson earned it. He scored 45 goals and 96 points on 294 shots and a 65 percent expected-goal share, the best on the team by a distance, and his 15.3 percent shooting is in line with a five-year record that has never dipped below 12.6.
The goaltending that carried them was not the goaltender they pay
Dallas got 16.1 goals of goaltending above expected at 5-on-5, fifth in the league. The distribution is the interesting part.
Oettinger's .899 was the worst full season of his career, and the trend under it is not comfortable.
The case for calm is that Oettinger still finished plus-5.7, which is above water, behind a defence that limits volume but concedes quality on the rush. The case for worry is that Dallas is paying starter money for a goaltender whose save percentage has fallen three years running, and that the 16 goals of team goaltending they banked last season were substantially the work of a 35-year-old backup on a $1 million expiring contract.
What is actually good here
A great deal. Miro Heiskanen and Esa Lindell were both 55 percent players with sub-2.10 expected goals against per 60, which is elite two-way defending. Thomas Harley at 25 is signed to 2033-34. Roope Hintz was a 56 percent player and plus-17 in only 53 games. Robertson is a top-ten winger. Wyatt Johnston at 23 is on the second line of a team that has been to three conference finals in four years.
The exception on the blue line is Tyler Myers, acquired at last season's deadline from Vancouver with half his salary retained. He played 1,195 minutes at 5-on-5, was on the ice for 2.79 expected goals against per 60, the worst of any Dallas regular, and finished minus-28 in goal differential. At $1.5 million retained he is cheap. In top-four minutes he is a problem.
The case for the summer, stated fairly
Dallas was cap-locked by its own success and had to choose between keeping a 24-year-old third-line centre and keeping a 45-goal winger. It kept the winger, which is correct. Bourque at six years and $5.5 million was never affordable in Texas. The Robertson bridge buys a year at a time when the cap is about to jump, and Dallas projects close to $39 million of space next summer. Seguin returning from a torn ACL is, in effect, a free addition. Glen Gulutzan has said the specific fix is neutral-zone structure to manufacture more even-strength offence, which is at least the right diagnosis.
Where this lands
Robertson kept at the cost of one year, an elite defensive spine intact, Harley locked to 2034, Seguin back, and a $39 million war chest next July. Nill sold a third-line centre, not a core piece, and got two picks for a player he could not have re-signed anyway.
A team that ranked 22nd in even-strength chance creation responded by selling its second-best even-strength forward and adding a $1 million winger. The 112 points rested on the league's best power play and 27 goals of finishing and goaltending luck, and the goaltending half of that came mostly from a 35-year-old backup on an expiring deal.
What it means
The Stars will win a lot of games again. The defence, the power play and Robertson guarantee it, and the Central is not deep. Watch expected goals for per 60 at 5-on-5, which was 2.36. If Gulutzan's neutral-zone changes push it toward 2.55, Dallas is a genuine contender rather than a regular-season one, and the Bourque trade will look like a price worth paying. If it sits at 2.36 again, then this is a team that needs a power play and a hot goaltender to win in April, and it has just committed $12 million to one winger and $8.25 million to a goaltender whose save percentage has gone the wrong way for three years. The record will not tell you which. The chance data will, by American Thanksgiving.