In their most recent game, Friday December 20 against the New York Rangers, the Dallas Stars played 17 minutes on the power play and were outscored 1-0; only the third time since 2009-10 that a team had that much time on the power play, and the only instance in which a team was outscored with that much power play time. That toothless showing was the low point of what’s been a frustrating power play all season. As of Saturday, the Stars sit 23rd in the league in goals for per 60 minutes on the power play; a stunning drop after finishing 5th and 6th in the two previous seasons, per Evolving-Hockey.
Under the hood, however, the numbers don’t look so bad. By both expected goals and corsi for per 60 minutes, the Stars are third in the league according to Evolving-Hockey. HockeyViz has the Stars slightly lower at sixth in expected goals rate, but still roughly in line with their previous performance under Pete DeBoer and Steve Spott.
There’s no doubt that the ineptness of the power play is hurting the Stars. In all, their expected goal differential on the power play is 13 goals higher than their actual differential, equivalent to roughly two wins. Digging into specific games (not generally a good practice, but somewhat enlightening in this case), there have been five games this season in which the Stars’ expected goal difference on the power play was greater than the score difference in the game, including three of the last four losses. Given the recent depths the power play has sunk to, it’s not surprising that frustration is at a peak right now.
So let’s take a look at what’s vexing (or hexing) the power play and what is causing the discrepancy between their underlying numbers and their top line numbers. Is it just luck, or are the underlying numbers misleading?
All statistics are as of December 21.
Zone Time and In-Zone Play
One of the biggest issues for Dallas has been their lack of offensive zone time on the power play. Currently, the Stars sit sixth worst in offensive zone time percentage on the man advantage with just 55.6% of their power play time spent in the attacking zone according the NHL EDGE stats. Historically, that has not correlated well with power play success.
Two things stand out here, besides Dallas’s poor offensive zone time this season. First, Dallas generates expected goals at a much higher rate than would be expected given their zone time. That means the Stars is making the most of their limited attacking time. However, even though they generate chances, the Stars are scoring even fewer goals than would be expected only based on zone time.
A related issue for the Stars is that they skate a lot on the power play. This is likely due to their lack of zone time as regroups and re-entries mean returning to the neutral or defensive zone and building up speed to attack the defending blue line. Once again, the Stars are at the wrong end of the spectrum when in comes to the historical trend. Teams that skate more on the power play tend to have worse outcomes and the Stars skate more than any other team this season.
This actually appears to be a feature (or weakness) of the DeBoer/Spott power play, which looks to create chances off the rush. Of Dallas’ 17 power play goals, eight have been scored off the rush compared to eight scored off in-zone formation (and one off a faceoff win). Recently though, the Stars haven’t been scoring off the rush. Dallas was the top skating team by average speed in both 2022-23 and 2023-24. Before that, Vegas was seventh in the league in 2021-22. In terms of offensive zone time, the 2022-23 and 2023-24 Dallas power plays were 19th and 30th while Vegas was 27th. Dallas was able to buck the trend the last two years, but it seems like some negative regression is at play here bringing them back down to what might be expected.
Looking at when the Stars generate their chances shows the effect of the lack of zone time on their power play production and their reliance on the rush attack.
Dallas has fixed (regressed to the mean) their poor power play faceoff percentage from earlier in the season, and that is evident in the surge of early chances they generate in the first 10 seconds of the power play. In fact, Dallas generate more chances off the opening faceoff of the power play than any team except the (similarly struggling) Bruins. From there, the chances slow down as Dallas is unable to generate in-zone pressure before spiking again when the second unit takes the ice around the one minute mark (usually on a regroup/re-entry) and generate a flurry of chances off the zone entry.
Shooting and Goaltending
Another issue for Dallas has been that the shooters taking shots haven’t been the most dangerous shooters. We can estimate shooter talent by taking the shooting component of Evolving-Hockey’s skater xGAR metric and dividing it by the number of individual unblocked shot attempts (fenwicks) each player took. Then we can take an average of the shooting talent for each player who took shot attempts on each team’s power play, weighted by the number of unblocked shot attempts they took. Since it’s only about 30 games into the current season, I combined the 2023-24 and 2024-25 seasons to estimate shooter talent.
Dallas ranks eleventh in average shooter talent for the 2024-25 season. That’s a notable drop from 2023-24 when they ranked second. The reason for the drop is mainly that the players taking shots for Dallas this season have seen their estimated shooting talent drop from last year to this year. The top shooters for Dallas by volume last season were Jason Robertson, Miro Heiskanen, and Joe Pavelski. This season, Wyatt Johnston has replaced Joe Pavelski as the third most frequent shooter. The problem is that Robertson is still struggling to find his form after late offseason surgery, Heiskanen has never been a particularly strong finisher, and Johnston has been in a slump all season. Together those three players account for 42% of Dallas’ fenwick events, and all three have struggled to finish their chances.
Another issue for Dallas is that they’ve also faced a tough set of opposing goaltenders. Using a similar method to estimate shooter talent, we can take Evolving-Hockey’s shorthanded goalie GAR component and estimate impact per fenwick faced. Then we can average that impact per goalie faced by each team’s power play weighted by the number of fenwick attempts against that goalie. Again, I’m combining 2023-24 and 2024-25 GAR numbers because of the limited sample from this season.
With this process we see that Dallas has faced the second toughest goaltending in the league so far, compared to last season when they faced the second easiest.
This is not to make excuses for the power play getting tough matchups. A good power play should be able to score against good goalies. In the playoffs, of course, the average level of competition is going to be higher and if the Stars have aspirations for the Spring, they need to be able to overcome good goaltending with their power play.
Luck
Lastly, the role of luck in a game as chaotic as hockey needs to be considered. Looking at expected goal and actual goal rates through the first 170 power play minutes of each season relative to the rest of the league (to account for changing scoring environments) for every team since 2009-10 puts this year’s Stars in rarefied air in a bad way.
Currently, the Stars generate chances at an above league average level (about 115% of the rest of the league’s rate), but finish them at an abysmal rate relative to the rest of the league this season (roughly 78%). The difference between their relative xGF per 60 rate and their GF per 60 rate is nearly -37% (only Boston and Anaheim are worse). The good news is that teams with a difference that large at the start of the season universally see that difference improve over the rest of the regular season, so some positive regression should be in order for the Stars. The bad news is that teams with such a large negative difference still tend to have a negative differential over the rest of the season. So although the Stars should expect their actual goals to converge to their expected goals, it shouldn’t be a surprise if they continue to underperform because although luck plays a role, being this poor is indicative of quality.
Lastly, still on the topic of luck, while Dallas is generating high total expected goals, they’re doing so through volume rather than by creating high quality individual chances. This season, The Stars rank 16th in expected goals per corsi on the power play, compared to 4th and 2nd in 2022-23 and 2023-24. This has the effect of increasing the variance of the outcomes, meaning luck plays a larger role than it would if they were generating lower volume, but higher average quality chances. If Dallas wants their skill to play a larger role, they should be more selective in their shot selection.
Conclusion
The Stars power play is hurting them this season, to the tune of somewhere between two and five additional regulation losses that are the difference between contending for the division lead and fighting for a wild card spot. Behind the struggles are issues with getting into the offensive zone, sustaining offensive in formation, and a poor stretch of finishing off the rush. In addition, the Stars are not getting the puck to their most dangerous finishers and their most frequent shooters are fighting some bad finishing luck. That’s being compounded by one of the toughest slates of goaltending in the league. Finally, there is likely a decent amount of poor luck involved which should correct itself to some degree over the rest of the season. Dallas’ power play is likely too skilled and its underlying numbers are too strong to stay in the bottom ten, but there are real process problems that need to be resolved.