The Seattle Kraken are riding a nine-game point streak, going 8-0-1 in that stretch. On Dec. 19, just before the streak began, Seattle was tied for last in the NHL with 30 points after matching a franchise-worst stretch at 1-9-1. Since then, the Kraken have collected 17 of 18 possible points and now sit third in the Pacific Division.
As fans suffered through Losing Streak Camille and Losing Streak Darren, there were still reasons for optimism. All but two games, both against Edmonton, were one-goal losses once empty-net goals were removed. The Kraken were competitive but consistently finished on the wrong side of the result. Interestingly, by the eye test, Seattle has not played as clean during the current points streak as it did during the losing stretch.
To explore that discrepancy, I compared game-level data from three segments: the current point streak, the losing streaks, and the games prior to Camille and Darren. With the point streak at nine games, the losing stretch at 11, and the pre-Camille segment at 21 games, all comparisons use per-game averages. Data is based on all situations and sourced from MoneyPuck.
- Pre-Camille: 21 games from Oct. 9 to Nov. 22
- Losing streaks: 11 games from Nov. 23 to Dec. 18
- Points streak: nine games from Dec. 20 to Jan. 6
Expected and actual goals
On Nov. 22, just before Losing Streak Camille, the Kraken ranked third-to-last in the NHL in goals scored. With only 2.58 expected goals for per game, Seattle relied heavily on defensive structure and low-event games.
Several trends stand out in the data. During the pre-Camille stretch, expected and actual goals for and against tracked closely. As Seattle pushed for more offense during the losing streak, expected goals for increased by 16.9 percent. The results did not follow. Actual scoring fell to two goals per game, while expected goals against jumped 27.4 percent to 2.91.
That trend reversed during the turnaround. Expected goals for climbed again to 3.16, but the finishing finally arrived. Over the past nine games, the Kraken have averaged 3.89 goals per game while allowing just 1.89.
There are also red flags. During the point streak, expected goals against sit at 3.69 per game. Across the season to date, a 3.69 xGA would rank last in the NHL. Vancouver currently holds that honor at 3.52. As the eye test suggested, Seattle often posted higher expected goals for than expected goals against during the losing streaks, but the scoring support was not there.
Goaltending
For this section, the focus is on team-level goaltending rather than individual performances. Team save percentage shows a strong relationship with goal support, particularly when Seattle scores first.
- Pre-Camille: 13 of 21 games with the first goal, 61.9 percent
- Losing streaks: three of 11 games, 27.3 percent
- Point streak: eight of nine games, 88.9 percent
Goaltending has been a major strength during the point streak, after dipping below .900 during the losing stretch.
Playing with the lead also continues to set this team up for success. Seattle has scored first in 24 of 41 games this season.
Despite expected goals against reaching their highest point of the season, the goaltenders have delivered. Over the last nine games, the Kraken have posted 16.2 goals saved above expected. Across the rest of the season, including pre-Camille and the losing streaks, they sit nearly even at minus 0.4.
Shots
Before the losing streak, Seattle generated only 24.4 shots on goal per game. If that pace held through Jan. 6, the Kraken would rank last in the NHL. During that stretch, there was a clear shift toward urgency and putting more pucks on net.
Shots on goal increased from 24.4 to 28.3 per game, a 16 percent jump that closely mirrors the 16.9 percent increase in expected goals for. Shot attempts rose even more sharply, up 19.5 percent.
That increase came with tradeoffs. Missed shots rose 23.1 percent and blocked shots climbed 21.9 percent. Both exceeded the growth in overall shot attempts, meaning a larger share of shots failed to reach the net.
The point streak tells a different story. Shot attempts are at their lowest point of the season, driven in part by Seattle frequently playing with the lead.
During the point streak, the Kraken have trailed for just 3.1 percent of total game time, or 17:03.
Even with fewer shot attempts, shot quality and efficiency have improved. Missed shots are down 39.5 percent and blocked shots are down 33.1 percent. Shots on goal have dipped by only one per game, to 27.3, and remain 11.9 percent higher than during the pre-Camille stretch.
Kraken can’t rest on their laurels
The Kraken continue to benefit from the current point streak, but there is little margin for complacency. Offensive progress is real, and goaltending has been the anchor. The balance between offense and defensive structure remains a constant tug-of-war. Limiting high-danger chances and bringing expected goals against back down should be a priority as Seattle looks to sustain its scoring gains.
The Kraken will look to extend their point streak to 10 games Thursday against the Minnesota Wild at Climate Pledge Arena. The Wild are coming off two losses to the Los Angeles Kings (one regulation and one shootout loss).
Feel free to leave any comments or questions below, and Go Kraken.
Blaiz Grubic
Blaiz Grubic is a contributor at Sound Of Hockey. A passionate hockey fan and player for over 30 years, Blaiz grew up in the Pacific Northwest and is an alumni of Washington State University (Go Cougs!). When he’s not playing, watching, or writing about hockey, he enjoys quality time with his wife and daughter or getting out on a golf course for a quick round. Follow @blaizg on BlueSky or X.
This is fantastic! Thanks Blaiz. This is the sort of granularity that is typically missing in most analytics breakdowns.
Like most things, you can take from it what you want. Just today Luszczyszyn was blasting the Kraken for their horrendous 5v5 xGF% without stopping to notice their actual GF% is tenth in the league. It seems pretty clear to me what Chicago coach Jeff Blashill brought up and analytics guru Jack Han confirmed is true… the numbers folks simply don’t know how to quantify defense.
I could certainly be mistaken on this, but it seems to me for the third season in a row we’ve seen that “shooting the puck more”, for this team, is a net loser. It seems to me, if this team is going to continue winning games, they’re going to have to focus on defense and rely on situational hockey.
Go Kraken!!!
…and thanks again Blaiz, good stuff.
Dom’s model is historically shit and he can’t be bothered to watch the games to see if it’s actually working right (hint: it often isn’t).
To be fair to Dom, and I hate saying that, the publicly available data is limited and that combined with defense being hard to quantify makes it brutal. The API doesn’t provide realtime player tracking so you have no idea if the goalie even saw the shot or was out of position, for instance.
I couldn’t agree more.
This one made me chuckle… Kraken are a 12% long shot to make the playoffs, while the Kings are a 85% safe shot.
https://x.com/domluszczyszyn/status/2009282008728449097?s=46&t=8Fcw-MbaRBz3DOcpSJuXjw
…and in this week’s Power Rankings Dom has Seattle at 15 and LA at 19. While Dom’s model now has Seattle all the way up to 14% to make the playoffs, MoneyPuck now has them at 66%. I can’t imagine even Luszczyszyn believes his model knows how to account for the Kraken, but he’s certainly not going to bring that up.
I’m always surprised how folks can look at data and draw conclusions that make no sense. There’s nothing in this data that suggests shooting more was bad. Most likely it doesn’t matter.
Xg for hockey under values defense a little but it under values excellent goal tending a lot. That’s the key to the win streak.
Is the current goal tending sustainable?
“As Seattle pushed for more offense during the losing streak, expected goals for increased by 16.9 percent. The results did not follow. Actual scoring fell to two goals per game, while expected goals against jumped 27.4 percent to 2.91.”
So you think 4 shots a game caused the defense to collapse?
But 3 shots more a game was the sweet spot
Counting shots is, to me, exactly the problem. It’s not about the volume of shots so much as the quality and focus. Four more shots, while significant (32nd v 10th), isn’t the issue. Four more shots is eight more shot attempts. If they’re simply throwing the puck at the net to prop up volume that’s a lot of posession they’re just giving away. If you compound that with a “shoot first” mentality displacing a “defense first” mentality… it seems to me that’s a problem.
I’m certainly no analyst, but I don’t think this team has the “top end” talent to “outscore” teams, but they do have a strong blueline and defensively responsible forwards. I think worrying less about shot volume and more about keeping the puck out of the net plays to their strengths and is in their best interest.
I clearly said shot volume didn’t matter. I’m not saying they should just chase more shots. I am saying increased shot volume didn’t create the losing streak. Cold goaltending did.
Again:
“As Seattle pushed for more offense during the losing streak, expected goals for increased by 16.9 percent. The results did not follow. Actual scoring fell to two goals per game, while expected goals against jumped 27.4 percent to 2.91.”
None of the numbers here, the expected numbers and the goals scored, are affected by Kraken goaltending. They may have had the losing streak for a variety of reasons, but the point I was following up on from the article is the one from above.
I’m curious if there is anything interesting in the data regarding special teams (which have been better) – is this just 5v5 data or EV or all situations?
This is all situations. I did not look at special teams specifically, but do have some data on the penalties for and against.
Penalties called on the Kraken (Kraken on the PK)
Pre-Camille – 3.1 penalties per game for 6.6 minutes
Losing streaks – 3.6 penalties per game for 8.2 minutes
Points streak – 2.8 penalties per game for 5.7 minutes
Penalties called on the opponent (Kraken on the PP)
Pre-Camille – 3.1 penalties per game for 6.6 minutes
Losing streaks – 4.3 penalties per game for 9.1 minutes
Points streak – 2.3 penalties per game for 5.0 minutes
So while on the current points streak, they are on the PP and PK alot less that when on the losing streak.
There is. 5×5 xgf% for Seattle is terrible. It would suggest that at even strength we’re not very good. Total xgf% isn’t bad. Our power play had been effective. That and empty netters is the difference.
At even strength their actual goals for is 13th in the league. Certainly the goaltending has been a big part of the disparity between expected and actual, but then you have to ask why has the goaltending improved so much? Obviously there’s some unsustainability to it, but I think Grubauer has clearly changed his game, the defensive structure is allowing a volume of low-danger shots that add up to more but less perilous expected goals, and the structure is being deployed situationally. Most of all, I don’t think the xG analytics do a very good job of accounting for defense.
Maybe they’re just on a lucky streak. Obviously they’re not an 18 of 20 point team, but they’ve played tight, low scoring games against all quality of opponents this season and to me it seems like a repeatable approach.
Where did you get that stat? When I look there actual gf is 31st.
There is a huge gap between their actual ga and xga, 36. This is why their xgf% is so bad. The gap is due to both their defensive structure and their goaltending. I believe it’s mostly goaltending. I think Blaise’ data makes that obvious. Particularly with the big increase in xga during the win streak.
Is this sustainable? Only if Seattle has figured out to be more offensive without sacrificing too much defense. The goaltending will revert to a save % around 910 eventually.
Maybe with our injured players coming back, the offense will be more productive. Kakko is clearly helping Beniers. McCann looks to be healthy. Otherwise, we will regress to the lower third of the league. That is what the numbers and the eye test suggest.
NST 5v5 GF% 50.72 13th in the league.
On the season Seattle is second only to Minnesota in 5v5 SV% at .931. Certainly this isn’t sustainable, but they’re also 26th in SH%. Can they balance the regressions? I hope so. I also think, without digging in, shooting more actually hurt their shooting percentage.
Obviously 18 of 20 isn’t a realistic expectation, but given the opportunities you mentioned – injuries, pairings, etc – can this team hit Blaiz’s “Strive for 95”? and not revert to the bottom third? I think so.
You need new eyes if you think that’s what the eye test is telling you.
It is a coincidence. I was just listening to the most recent Emerald City Hockey podcast in which they talked about the bizarre analytics that underpin the Kraken’s games since mid-way through the losing streaks, and in it RJ threw his hands up and said, “I give up. None of this makes sense.” Apparantly, in the past thirteen games the Kraken have won all but one game where the stats (MoneyPuck and HockeyBiz) said they should have lost (eight-zero-one) and, even more inexplicably, lost every single game (four) they should have won. If anything, the stats are completely upside-down. The team gets better results on the scoreboard when their advanced stats are worse and vice versa. The elite goaltending from Daccord and Grubauer may explain how some of the expected losses became real-world wins, but it also flies in the face of the real-world losses that were expected wins. The 2025-26 Kraken seem to be a statistical anomaly. That or, as Daryl W said, analytics are just really awful at quantifying defensive play.
Expected goals and the way they are calculated are a big part of it. This is why using your eyeballs is still required. Most of the recent games haven’t felt lopsided in favor of either team visually but the analytics would tell you otherwise. That tells you the advanced stats are wrong which isn’t a surprise. They’re trying to extract meaning from very limited publicly available data.
A lot of things don’t make sense to ECH. They’ve been full of god awful takes and never question the data.
Good stuff! Would love to see the shooting percentage and PDO numbers broken down like this too.
its such a shame that xGF% isnt weighted for leading/trailing, when a team is down they are more willing to take riskier shots on net compared to when leading. this will favor the trailing team in model because more “chances” are created.
Many expected goal models do adjust for score effects, and there are sources (HockeyViz, Natural Stat Trick) that break things down by score state. How would you change those adjustments?
The odds of you being alive are 1: 400,000,000,000,000. The simple fact that you’re reading these words right now is some type of statistical miracle…said someone on the Internet.
If statisticians ran the planet, no one would get out of bed in the morning.
What’s your point?
If the goal is to win the hockey game in front of you, which I think it is, some of the stats (most?) seem meaningless. I haven’t run the numbers (because I have other things to do…and I actually just like watching hockey games) but one would THINK there is some sort of correlation to puck possession (CF%?) and time spent in the offensive zone where the opposing team’s net actually is? But it’s probably a 50-50 indicator of who actually won the hockey game.
Or maybe it’s kind of like teams with the most coveted high scoring players don’t always win the Stanley Cup, or even their division.
The things that actually lead to one team prevailing over another are REALLY difficult to measure with stats of the day, but so many “experts” use them with such gravitas.
They are hard to measure, that’s why it’s called data science. Hard doesn’t mean it isn’t useful or worthwhile.
You’d probably be surprised how much of the world is run by statistician. You might want to go back to bed.
LOL. I understand stats and their value. How are they helping you win a hockey game? In other words, if I am a coach, what statistical trends demonstrably ensure my team wins more often than not? Do we know what those are? Is there any direct link to any of these stats and winning more frequently, other than goals for and goals against?
Seems you need to have the puck to score a goal. What percentage of puck possession do I need to achieve to ensure I win? Is there a correlation to winning and puck possession? If there is, why does Jamie Oleksiak love to give it to the opposing team whenever possible? In fact, if we need the puck, why do teams often dump it (give it away) when they need a line change rather than just circle back or even just fire it back to their own goalie so they can regroup in their own end…with the puck?
Don’t take it too seriously.
I’m more of a baseball than a hockey fan I like to translate the statistics. The baseball equivalent to getting “goalied” is getting “BABIPed”. The baseball equivalent to a high danger scoring chance is a hard hit ball with a high xBA.
Shot volume has to be the most useless hockey stat of all. Nobody would care about shot volume if it wasn’t right there on the score bug. It’s like pitcher W-L record in baseball. If a pitcher wins more games than he loses, there’s probably some correlation to something good happening, but it could be a million different things. Putting up 40 low danger shots is the same as 40 ground balls to shortstop. The shortstop will probably boot one of them and you’ll get a hit, but you would have been better off swinging harder and striking out 20 times instead of getting 40 useless ground outs.
The challenge with advanced metrics in hockey is that even a “high danger scoring chance” is difficult to evaluate. When the Kraken were slumping it seemed like they couldn’t hit the side of a barn with their shots. Even the high danger chances were being flubbed, hitting the glass, or going right into the goalie’s chest. Same at the beginning of the season when Eeli Tolvanen was getting a bunch of good looks but there was a near 0% chance that he was actually going to convert them because he seemed to have forgotten how to shoot.
Imagine you play the Avalanche and you have an even shot volume, an even number of high danger changes, and Moneypuck says that expected goals are even. Are you going to win that game? Absolutely not.
Just looking at the way Jared McCann is shooting the puck right now tells you all you need to know about why we’re winning now vs. a month ago. The whole team is playing with the confidence of Jared McCann’s wicked shot.
Exactly! I love this take.
Well they might win that game. That’s what you’re trying to see past right?
Hot goalies is probably the biggest reason they are winning now. I agree McCann is helping.
The weird thing is… over this ten game stretch, the Kraken are 8th in the league at 3.63 G/60. I know it’s a small sample size and the goaltending is a big part of their +17 goal differential over those 10 games, but it seems like write this off as “hot goaltending” might be missing a few things.
The weird thing is… this team that can’t score is 8th in the league at 3.63 G/60 over the last 10 games. That’s not the goalies, is it?
Strange double-tap.
I don’t believe it’s just the goalies either. It’s never just one thing. Scoring is up over that short sample. Injured players are back. We’ve played some easier teams.
My interpretation of the data is the hot goaltending is the biggest factor.
I’m trying to figure out how good this team is and it’s really weird. Most of the defensive data is pretty good. Most of the offensive data is really bad. Watching the games i see a very up and down team with an offense that had rarely put it together for 2 periods in a row.
Some of that may be defensive structure but our numbers are so bad I don’t think that’s it.