Ben Csiernik

While Team USA didn't create a ton of chances via volume, they were able to capitalize on a couple high danger chances. Sometimes that's all it takes in a hockey game.

Team Canada controlled possession for a majority of the game, but Team USA kept them from getting to the dangerous spots over the course of the game extremely well.

A game report showing Team USA scoring 2 goals on 1.66 expected goals on 33 shot attempts, 33 shots, and 6 scoring chances. Canada had 0 goals on 1.13 expected goals on 49 shot attempts, 35 shots on net, and 1 scoring chance.

A game probability worm showing the most likely score in the game would have been 1 to 1.

This past Sunday, team USA won the 2026 IIHF U18 Women’s World Championship 2-0 over Team Canada. A really competitive game from two teams filled with players with bright hockey futures. Here is the game report and game probabilities based on my modelling work in women's hockey.

I don't have many thoughts on Gascon aside from I think she's excellent, so that one may just be an experience angle.

As for Primerano, it was primarily deployment. Limited minutes generally, but limited OZ starts and PP use. Hard for her to show her best with less friendly opportunities.

Signs were there from the rivalry series, but I’ve been wanting a little more opportunity for Nicole Gosling and Kendall Cooper too, though that seems extremely far away now.

The most obvious one is how fundamentally different advantaged rushes are compared to non-advantaged rushes. They're essentially non-comparable, and hard (impossible?) to tease out without detailed data. Here are the two most relevant density maps showcasing this from my article.

An article looking at rush offence using women's Canadian University is up on my website. Had a lot of fun looking at rush offence in response to the discussion in the NHL.

While not directly applicable to the NHL tactically, I think there's important overlaps.

sites.google.com/view/bencsie...

I regularly record shots from inside the crease in my modelling and tracking. A 3 foot increase from low slot to inside crease can make my models xG value jump by ~0.2 at times.

It makes a big difference as you’ve clearly established

This is fundamentally a blocked shot and I have a hard time seeing it in any other light. If a puck hits a defenders stick and doesn’t go at the net then it was impeded.

In my tracking I wouldn’t include that as a shot attempt (nor do I think many people would).

Curious the thought process from the NHL data team here

We can only dream. I’d imagine the NHL has no intention of doing that for a number of reasons, which is sad but that’s show business.

I plan on writing up a longer piece on the relationship between rush shot decision making and overall offensive shot choices, especially at it pertains to women's hockey, where shot location decision making seems to have slightly larger impacts than in men's hockey.

Removing breakaways gives us a much clearer idea of team philosophy on advantaged rushes. Shot choices are always interesting.

A dot line shot on a 2 on 1 grades out to be worth about 0.045 xG in my model, while a shot from the slot at the same X axis positioning is worth a hair under 0.14.

Back to overall trends. Just under 40% of all advantaged rush attempts are breakaways in my data set, which means looking at a density plot of all advantaged rushes leads to a very narrow net front distribution, as you would expect. Not super helpful really.

As @hockeyviz.com has often said, rush offense is often indicative of a team's overall offensive philosophy. This is seen through comparing a high volume approach (Team A) with a lower volume middle focused approach (Team B).

Some approaches mean if the shot isn't there, you establish possession

Women's college hockey scores almost 50% of its goals from the inner and middle slot, which I regularly am overwhelmed by.

With that being said, when it comes to rushes, if we remove advantaged rushes (e.g. 2 on 1) our density map notably changes. Shots from the top of the circles are not ideal

I personally define a rush chance as a team generating a shot attempt within the first five seconds of entering the offensive zone. In these situations, teams don't have typically have time to establish their defensive structure.

The density map is unsurprisingly very net front focused

Lots of good discussion on rush chances in hockey the past few weeks from @ladyotreesyrup.bsky.social and @hockeyviz.com . Thought I'd throw my hat in the ring with some more detailed data, albeit, not NHL. So, a rush breakdown from the detailed data that I do have in women's college hockey.

New year means the second half of the OUA women's hockey season is about to start and I am caught up on all 101 games from the first half with 10 thousand + shot attempts watched.

Looks like scoring has stabilized around expected, and team shooting profiles look very different.

We have an unpublished undergraduate thesis kicking around from last year with those estimates, so a big shoutout to the student who did a great job watching 500+ rush attempt videos, and now plays pro hockey.

We spent more time discussing rush tactics philosophy than research but it was fun.

Did someone say 35%?

The Olympic hockey roster argument that makes my eyes glaze over more than any other is "you need players to play a bottom-six and fourth-line role."

No you don't.

You literally don't.

Not enough grit. Can’t measure heart with numbers.

In the NHL, I’d argue this is probably negligible at the riskiest.

At other leagues and levels, definitely. My model in womens hockey has a number of shot types with xG values at or around .002 which is…not good!

Both of their incoming recruiting classes for next year are overwhelmingly good too if that makes you feel better

Fair and reasonable.

If you could sum up your gripes with the WCHA in 300 characters, what are they

I’d likely imagine that the rush percent shot attempt number should be much closer to 35%. But I have NO idea from public play by play data how we could get there if that is the true number

All this to say. Public play by play data makes this complicated I think

9% feels so low. I supervised an undergraduate student who looked at NHL rush chances.

We used various cut offs to identify rushes but ultimately settled on 10 seconds from an event outside. 524 unblocked shot attempts were reviewed by video with 485 featuring a shot off the rush entry..