Aaron K
@ladyotreesyrup.bsky.social
Stars fan, data fan, handle to the tune of "Radio Free Europe"
He/him
https://puckovertheglass.substack.com/
He/him
https://puckovertheglass.substack.com/
Pinned
Aaron K
@ladyotreesyrup.bsky.social
· Dec 19
Something Strange Is Going On with the NHL's Play-By-Play
The Box Scores, They Are A-Changin'
open.substack.com
New post: Something Strange Is Going On with the NHL’s Play-By-Play
Looking at changes in shot locations and scorekeeper bias and a guess at what’s behind the change
open.substack.com/pub/puckover...
Looking at changes in shot locations and scorekeeper bias and a guess at what’s behind the change
open.substack.com/pub/puckover...
Reposted by Aaron K
Here's a thread of some scattered thoughts about Jordan Spence, first star of last night's Ottawa win over Utah. He is a very good and very unusual player, which is to say: my favourite kind of player.
November 10, 2025 at 6:27 PM
Here's a thread of some scattered thoughts about Jordan Spence, first star of last night's Ottawa win over Utah. He is a very good and very unusual player, which is to say: my favourite kind of player.
Reposted by Aaron K
New tool up on LB-Hockey, the Steals & Chokes Report! 🚨🏒
Our first goalie-exclusive viz looks to capture their game-breaking ability by showcasing how their stolen & choked games over the years, and providing info on those individual performances.
🔗 lb-hockey.com/steals-and-c...
Our first goalie-exclusive viz looks to capture their game-breaking ability by showcasing how their stolen & choked games over the years, and providing info on those individual performances.
🔗 lb-hockey.com/steals-and-c...
November 6, 2025 at 4:47 PM
New tool up on LB-Hockey, the Steals & Chokes Report! 🚨🏒
Our first goalie-exclusive viz looks to capture their game-breaking ability by showcasing how their stolen & choked games over the years, and providing info on those individual performances.
🔗 lb-hockey.com/steals-and-c...
Our first goalie-exclusive viz looks to capture their game-breaking ability by showcasing how their stolen & choked games over the years, and providing info on those individual performances.
🔗 lb-hockey.com/steals-and-c...
Reposted by Aaron K
Let me write a short thread as it is of interest to many. I'll start off by saying that the landscape has changed from when I got into hockey analytics and when I got hired by the Calgary Flames.
Hey @hockeyanalysis.com, just saw a couple of your posts on X, and I was just wondering if you happened to have any advice for someone aspiring to work in the hockey analytics space or something along those lines
November 4, 2025 at 7:58 PM
Let me write a short thread as it is of interest to many. I'll start off by saying that the landscape has changed from when I got into hockey analytics and when I got hired by the Calgary Flames.
Reposted by Aaron K
Bat speed aging curve, but this time I account for the fact that we are less likely to observe very young/old players. The corrected curve declines less rapidly early and more rapidly later compared to the regular curve, which seems like the correct inference:
November 1, 2025 at 8:05 PM
Bat speed aging curve, but this time I account for the fact that we are less likely to observe very young/old players. The corrected curve declines less rapidly early and more rapidly later compared to the regular curve, which seems like the correct inference:
Reposted by Aaron K
New season, new NHL EDGE landing page! The NHL has updated both the UI and the backend for their EDGE stats, converting the data transfer from a web socket message/response system to API endpoints
October 9, 2025 at 3:30 AM
New season, new NHL EDGE landing page! The NHL has updated both the UI and the backend for their EDGE stats, converting the data transfer from a web socket message/response system to API endpoints
New season, new NHL EDGE landing page! The NHL has updated both the UI and the backend for their EDGE stats, converting the data transfer from a web socket message/response system to API endpoints
October 9, 2025 at 3:30 AM
New season, new NHL EDGE landing page! The NHL has updated both the UI and the backend for their EDGE stats, converting the data transfer from a web socket message/response system to API endpoints
Reposted by Aaron K
Top 50 Projected Players for 2025-26 by SPAR: 31 to 40
A slew of RWs leads this section, including some from the next wave in Jarvis & Boldy. Meanwhile, Josi looks to bounce back from injury and prove he’s still one of the most dynamic dmen out there.
📈 lb-hockey.com
A slew of RWs leads this section, including some from the next wave in Jarvis & Boldy. Meanwhile, Josi looks to bounce back from injury and prove he’s still one of the most dynamic dmen out there.
📈 lb-hockey.com
October 4, 2025 at 3:55 PM
Top 50 Projected Players for 2025-26 by SPAR: 31 to 40
A slew of RWs leads this section, including some from the next wave in Jarvis & Boldy. Meanwhile, Josi looks to bounce back from injury and prove he’s still one of the most dynamic dmen out there.
📈 lb-hockey.com
A slew of RWs leads this section, including some from the next wave in Jarvis & Boldy. Meanwhile, Josi looks to bounce back from injury and prove he’s still one of the most dynamic dmen out there.
📈 lb-hockey.com
Reposted by Aaron K
New post: Introducing RAPM for transition stats. Which players are pushing the puck up ice driving creating more time in the offensive zone? It turns out play on entries has a far greater impact that play on exits.
open.substack.com/pub/puckover...
open.substack.com/pub/puckover...
Introducing RAPM for Transition Statistics
Which players drive zone entries and exits with possession?
open.substack.com
September 29, 2025 at 1:02 PM
New post: Introducing RAPM for transition stats. Which players are pushing the puck up ice driving creating more time in the offensive zone? It turns out play on entries has a far greater impact that play on exits.
open.substack.com/pub/puckover...
open.substack.com/pub/puckover...
Reposted by Aaron K
Top 10 Projected Players by Position for 2025-26: Defencemen
Quinn Hughes begins the season as the Norris favourite for 2025-26. Dallas is the only team with multiple showings here, with Harley at 10th and Heiskanen in the top 5.
📈 lb-hockey.com
Quinn Hughes begins the season as the Norris favourite for 2025-26. Dallas is the only team with multiple showings here, with Harley at 10th and Heiskanen in the top 5.
📈 lb-hockey.com
September 29, 2025 at 5:13 PM
Top 10 Projected Players by Position for 2025-26: Defencemen
Quinn Hughes begins the season as the Norris favourite for 2025-26. Dallas is the only team with multiple showings here, with Harley at 10th and Heiskanen in the top 5.
📈 lb-hockey.com
Quinn Hughes begins the season as the Norris favourite for 2025-26. Dallas is the only team with multiple showings here, with Harley at 10th and Heiskanen in the top 5.
📈 lb-hockey.com
New post: Introducing RAPM for transition stats. Which players are pushing the puck up ice driving creating more time in the offensive zone? It turns out play on entries has a far greater impact that play on exits.
open.substack.com/pub/puckover...
open.substack.com/pub/puckover...
Introducing RAPM for Transition Statistics
Which players drive zone entries and exits with possession?
open.substack.com
September 29, 2025 at 1:02 PM
New post: Introducing RAPM for transition stats. Which players are pushing the puck up ice driving creating more time in the offensive zone? It turns out play on entries has a far greater impact that play on exits.
open.substack.com/pub/puckover...
open.substack.com/pub/puckover...
Reposted by Aaron K
It appears the NHL has spent some time standardizing its shift reports over the past few seasons. Short OTF shifts (10 seconds or less starting and ending OTF without a stoppage in between) have steadily dropped over each of the last three seasons to just over 50% of the previous rate
September 19, 2025 at 7:01 PM
It appears the NHL has spent some time standardizing its shift reports over the past few seasons. Short OTF shifts (10 seconds or less starting and ending OTF without a stoppage in between) have steadily dropped over each of the last three seasons to just over 50% of the previous rate
It appears the NHL has spent some time standardizing its shift reports over the past few seasons. Short OTF shifts (10 seconds or less starting and ending OTF without a stoppage in between) have steadily dropped over each of the last three seasons to just over 50% of the previous rate
September 19, 2025 at 7:01 PM
It appears the NHL has spent some time standardizing its shift reports over the past few seasons. Short OTF shifts (10 seconds or less starting and ending OTF without a stoppage in between) have steadily dropped over each of the last three seasons to just over 50% of the previous rate
Reposted by Aaron K
New post: Jason Robertson is really, really good at hockey, even if it's not always obvious puckovertheglass.substack.com/p/just-how-g...
Just How Good Is Jason Robertson?
From a statistical perspective
puckovertheglass.substack.com
September 16, 2025 at 8:41 PM
New post: Jason Robertson is really, really good at hockey, even if it's not always obvious puckovertheglass.substack.com/p/just-how-g...
New post: Jason Robertson is really, really good at hockey, even if it's not always obvious puckovertheglass.substack.com/p/just-how-g...
Just How Good Is Jason Robertson?
From a statistical perspective
puckovertheglass.substack.com
September 16, 2025 at 8:41 PM
New post: Jason Robertson is really, really good at hockey, even if it's not always obvious puckovertheglass.substack.com/p/just-how-g...
Reposted by Aaron K
Why points are overrated as a stat (but still useful), and also a quick overview on how content at The Five Hohl will look like for this season.
link: open.substack.com/pub/thefiveh...
link: open.substack.com/pub/thefiveh...
September 15, 2025 at 6:28 PM
Why points are overrated as a stat (but still useful), and also a quick overview on how content at The Five Hohl will look like for this season.
link: open.substack.com/pub/thefiveh...
link: open.substack.com/pub/thefiveh...
Reposted by Aaron K
New post: A statistical deep-dive into how Jake Oettinger compares to his peers over the past three seasons and where he should rank going forward puckovertheglass.substack.com/p/just-how-g...
Just How Good Is Jake Oettinger?
From a statistical perspective
puckovertheglass.substack.com
September 15, 2025 at 2:08 PM
New post: A statistical deep-dive into how Jake Oettinger compares to his peers over the past three seasons and where he should rank going forward puckovertheglass.substack.com/p/just-how-g...
New post: A statistical deep-dive into how Jake Oettinger compares to his peers over the past three seasons and where he should rank going forward puckovertheglass.substack.com/p/just-how-g...
Just How Good Is Jake Oettinger?
From a statistical perspective
puckovertheglass.substack.com
September 15, 2025 at 2:08 PM
New post: A statistical deep-dive into how Jake Oettinger compares to his peers over the past three seasons and where he should rank going forward puckovertheglass.substack.com/p/just-how-g...
Reposted by Aaron K
⬡⬢⬡ New Model at HockeyViz ⬡⬢⬡
This year's new version of my model for estimating player impacts on 5v5 chance rates.
Read all the details here: hockeyviz.com/txt/magnus9EV
This year's new version of my model for estimating player impacts on 5v5 chance rates.
Read all the details here: hockeyviz.com/txt/magnus9EV
August 27, 2025 at 5:01 PM
⬡⬢⬡ New Model at HockeyViz ⬡⬢⬡
This year's new version of my model for estimating player impacts on 5v5 chance rates.
Read all the details here: hockeyviz.com/txt/magnus9EV
This year's new version of my model for estimating player impacts on 5v5 chance rates.
Read all the details here: hockeyviz.com/txt/magnus9EV
The Modano is 14.99? Does Mike Babcock run this deli?
There’s a deli in south Minneapolis that’s North Stars themed (complete with showing old games) 35 years after the team relocated to Dallas. Pretty badass.
August 25, 2025 at 7:01 PM
The Modano is 14.99? Does Mike Babcock run this deli?
Reposted by Aaron K
I'm gonna put a chapter in my (upcoming! soonly!) shot rate model write-up about this but as a preview: the player talent distribution in an apex league is not normal and in fact it's more skewed than would match the way we "normally" talk about talent. (This is forwards, generating offence.)
August 22, 2025 at 4:51 PM
I'm gonna put a chapter in my (upcoming! soonly!) shot rate model write-up about this but as a preview: the player talent distribution in an apex league is not normal and in fact it's more skewed than would match the way we "normally" talk about talent. (This is forwards, generating offence.)
Reposted by Aaron K
New Post: Estimating shot speed by shot type using game-by-game NHL EDGE data that I scraped myself. Plus, controlling for man power and blocked shots. Some surprising results and a possible source of improvement for public xG models
puckovertheglass.substack.com/p/nhl-edge-e...
puckovertheglass.substack.com/p/nhl-edge-e...
NHL EDGE: Estimating Shot Speed by Shot Type
How much power does a slapshot add? A wrist shot?
puckovertheglass.substack.com
August 21, 2025 at 5:01 PM
New Post: Estimating shot speed by shot type using game-by-game NHL EDGE data that I scraped myself. Plus, controlling for man power and blocked shots. Some surprising results and a possible source of improvement for public xG models
puckovertheglass.substack.com/p/nhl-edge-e...
puckovertheglass.substack.com/p/nhl-edge-e...
New Post: Estimating shot speed by shot type using game-by-game NHL EDGE data that I scraped myself. Plus, controlling for man power and blocked shots. Some surprising results and a possible source of improvement for public xG models
puckovertheglass.substack.com/p/nhl-edge-e...
puckovertheglass.substack.com/p/nhl-edge-e...
NHL EDGE: Estimating Shot Speed by Shot Type
How much power does a slapshot add? A wrist shot?
puckovertheglass.substack.com
August 21, 2025 at 5:01 PM
New Post: Estimating shot speed by shot type using game-by-game NHL EDGE data that I scraped myself. Plus, controlling for man power and blocked shots. Some surprising results and a possible source of improvement for public xG models
puckovertheglass.substack.com/p/nhl-edge-e...
puckovertheglass.substack.com/p/nhl-edge-e...
Reposted by Aaron K
New article up on LB-Hockey! 🚨🏒
We showcase players’ willingness to engage in high-contact areas and ability to succeed when pressured, along with how it correlates to team success & player development. It also introduces a new free visualization: the Battle Grid!
🔗: lb-hockey.com/2025/08/13/d...
We showcase players’ willingness to engage in high-contact areas and ability to succeed when pressured, along with how it correlates to team success & player development. It also introduces a new free visualization: the Battle Grid!
🔗: lb-hockey.com/2025/08/13/d...
August 13, 2025 at 4:00 PM
New article up on LB-Hockey! 🚨🏒
We showcase players’ willingness to engage in high-contact areas and ability to succeed when pressured, along with how it correlates to team success & player development. It also introduces a new free visualization: the Battle Grid!
🔗: lb-hockey.com/2025/08/13/d...
We showcase players’ willingness to engage in high-contact areas and ability to succeed when pressured, along with how it correlates to team success & player development. It also introduces a new free visualization: the Battle Grid!
🔗: lb-hockey.com/2025/08/13/d...
Reposted by Aaron K
I also think it's worth emphasising that, as with all of these lists, the numbers are just random guesses. There is no empirical method, it's just "How likely do you think it is that computers will be able to do this task in the future?" and then the poll respondents pick a number at random.
Just a reminder this list is drawn from a piece written by four researchers for MICROSOFT (data scientists, CS people, economists) who have never stepped foot in an archive/never taken an oral history/studied material culture, by their CVs.
Unlike them, I do my research, and don't talk out my ass.
Unlike them, I do my research, and don't talk out my ass.
August 4, 2025 at 11:43 PM
I also think it's worth emphasising that, as with all of these lists, the numbers are just random guesses. There is no empirical method, it's just "How likely do you think it is that computers will be able to do this task in the future?" and then the poll respondents pick a number at random.
Reposted by Aaron K
New post: What happens when Analytics Darlings are promoted out of depth roles?
puckovertheglass.substack.com/p/do-analyti...
puckovertheglass.substack.com/p/do-analyti...
Do Analytics Darlings Scale Up the Lineup?
How do players who put up good numbers in limited minutes perform when given a bigger role?
puckovertheglass.substack.com
July 8, 2025 at 1:09 PM
New post: What happens when Analytics Darlings are promoted out of depth roles?
puckovertheglass.substack.com/p/do-analyti...
puckovertheglass.substack.com/p/do-analyti...