The Minnesota Wild bet heavily on players the analytics community would bench, and for the second consecutive game, it worked. Yakov Trenin, who entered Tuesday's 4-2 victory over Dallas with a -7.8 goals versus expected mark - worst among Minnesota forwards - drew the primary assist on the insurance goal that effectively ended the Dallas Stars' season-saving comeback attempt. The Wild now lead the first-round series 3-2 with Game 6 set for Thursday in Minnesota.
Trenin's 46% on-ice expected goals share ranks dead last among Wild regulars who logged meaningful minutes this season, per MoneyPuck. Yet head coach John Hynes gave him 12:04 of ice time in a game Minnesota couldn't afford to lose, and the 29-year-old Russian winger responded with seven hits - most among Wild forwards - and the setup on Michael McCarron's third-period goal that made it 3-1 at 7:47.
The sequence didn't come from a structured zone entry or a crafted play. Trenin won a forecheck battle along the left boards, forcing Stars defenseman Thomas Harley into a rushed clearing attempt that Vladimir Tarasenko intercepted at the blue line. Tarasenko's quick feed found McCarron alone in the slot for his first goal of the series. Four seconds from initial contact to puck crossing the goal line.
That cushion proved essential. Jason Robertson's power-play goal at 16:39 pulled Dallas within one, but Kirill Kaprizov's empty-netter at 18:00 sealed the win. Without McCarron's marker, the Stars would have spent the final five minutes chasing a one-goal deficit with their season on the line. Instead, they burned clock trying to overcome two goals they never seriously threatened.
Minnesota had analytically sounder options available. Ryan Hartman posted a 58% on-ice expected goals share this season with a goals-versus-expected mark of +0.3. Mats Zuccarello carried a 63% on-ice xGF%. Both played fewer than 17 minutes Tuesday. The Wild could have leaned harder on possession-driving forwards. They deliberately chose not to.
The decision looks sharper when considering what happened to Dallas' most analytically impressive players. Mikko Rantanen entered Tuesday with a 62% on-ice expected goals share, best among Stars regulars. Matt Duchene sat at 58%. Robertson carried a 65% mark. Those three combined for zero goals at 5-on-5 despite playing 68 combined minutes. Rantanen managed two shots on goal. Duchene recorded zero.
Minnesota's physical pressure - 28 hits to Dallas' 20, with Trenin leading the way - forced the Stars into exactly the kind of hockey their skill players struggle to win. Dallas generated just 22 shots on goal, their lowest total of the series. Their 50.8% expected goals share at 5-on-5 during the regular season suggests they should control roughly half the quality chances in neutral games. They didn't approach that Tuesday.
The contrast between McCarron and Trenin's analytical profiles and their actual impact raises legitimate questions about what expected goals models capture in playoff situations. McCarron's -5.6 goals versus expected and 41% on-ice xGF% mark him as nearly as poor a possession player as Trenin. Yet both logged meaningful minutes in a must-win game, and both contributed directly to the winning margin.
The argument isn't that expected goals lack value. Kaprizov's +6.6 goals versus expected and 63% on-ice xGF% help explain his three-point night. Quinn Hughes' 60% on-ice expected goals share correlates with his ability to quarterback Minnesota's power play, which converted its only opportunity. The model identifies high-end talent effectively.
Where it struggles is measuring disruption value. Trenin doesn't create high-quality chances for his linemates - his 46% on-ice xGF% confirms that. But his seven hits in 12 minutes forced Dallas into mistakes that created transition opportunities for other Wild players. The McCarron goal didn't originate from a controlled breakout. It came from Harley making a poor decision under physical pressure, exactly the type of play that pre-shot metrics don't capture.
Dallas' roster construction amplifies this vulnerability. The Stars added Rantanen at the trade deadline to bolster their skill advantage. They already had Robertson, who posted a +7.3 goals-versus-expected mark this season, third-best in the dataset. Miro Heiskanen quarterbacked their power play to a 42.1% conversion rate in the playoffs. On paper, they should overwhelm opponents with shot quality.
In practice Tuesday, they managed two goals, both on the power play. Their 5-on-5 offense generated nothing against a Minnesota team willing to sacrifice possession time for physical engagement. The Wild finished with a 51.9% faceoff win rate and 26 blocked shots, both modest advantages that understate how effectively they clogged shooting lanes and forced Dallas to the perimeter.
Minnesota has now outscored Dallas 11-3 at even strength through five games, according to NHL.com game notes. That margin exists despite Minnesota icing two of the league's worst finishers by expected goals - Trenin and McCarron - in meaningful minutes. The Wild aren't accidentally getting contributions from analytically questionable players. They're deliberately deploying physical depth pieces who force mistakes, then capitalizing on the chaos those mistakes create.
The coaching staff's decision to play Trenin despite his analytical profile suggests they've identified something the models miss in compressed playoff hockey. Teams defend harder, goaltenders elevate performance, and the margin between winning and losing often comes down to who makes fewer mistakes under pressure. Trenin's physicality creates that pressure. Whether that value exceeds the offensive production he costs his team is impossible to quantify with current metrics, but Tuesday's 4-2 victory offers evidence supporting Minnesota's approach.
The risk, naturally, is that one game proves nothing. Trenin's -7.8 goals versus expected over 82 regular-season games represents a much larger sample than his 12:04 Tuesday night. If Minnesota continues deploying him heavily and he costs them goals through poor finishing or defensive breakdowns, the analytical case against him strengthens. The Wild are betting that playoff hockey rewards different skills than regular-season hockey, and that Trenin's disruption matters more in a seven-game series than his shooting efficiency.
For Dallas, the path forward requires solving Minnesota's physical pressure that limited them to 22 shots, or relying on a power play already responsible for both their goals Tuesday. The Stars finished the regular season 50-20-12 (112 points), eight points clear of Minnesota's 46-24-12 (104 points). That regular-season edge came partly from superior shot quality metrics that haven't translated through five playoff games.
Game 6 returns to Minnesota on Thursday with the Wild one win from advancing. Trenin will almost certainly draw back into the lineup. What he does with those minutes will tell us whether Tuesday's performance represents sustainable playoff value or statistical noise the models will eventually correct.