Offseason Time!

Now that we are more into the dead season, it’s time for me to look back and forwards at both this publication and myself. I am mostly out of writing ideas in the pipeline though the recent roster moves and film studies have renewed a certain itch.

For now, my intention is to take the rest of the offseason to learn some new skills and prepare for a big 2024-25 where I intend to implement changes and improvements to my overall process.

The Bottleneck

As I’ve got into increasing volume of specific player breakdowns and in trying to answer questions I may have, I’ve discovered a few bottlenecks. Primarily, I am not proficient in writing the sort of code that helps me answer those questions and which could quickly generate charts and tables for each player quickly. I could significantly increase my agency if I developed that specific skill.

In school, mostly via quantitative neuro and computational neuro classes, I was exposed to some MatLab and basic statistical analysis software but it’s been a while and I was never really satisfied with the depth of my knowledge there. I’m not sure I can fundamentally pick up those skills in the limited time between now and pre-season, but I’ll do my best to at least get a foot in the door.

Really, I’ve found a lot of enjoyment in examining cohort analysis, associating it with video, and understanding how everything comes together from a development perspective. Doing day-to-day film reviews only serves to continue to better understand the context required to drive more specific conclusions and understandings.

If I had the skills, I would love to put a lot of time into analyzing post-draft and into NHL growth and development.

To do this, I have a hunch that there may be “higher-hanging fruit” in the specifics of Quality of Competition. I think we should be able to use AllThreeZones microstats to create more specific, though not necessarily better, Quality of Competition and examine performance through the stylistic differences there.

Furthermore, I think understanding the skills that drive success at sub-NHL levels and which skills might be easier to translate would be a fruitful course to pursue. In order to do that we might need to understand the differences in fundamental structure between leagues, Mitch Brown and Lassi Alanen’s tracking data could move toward that, but we also might need to dig into systems analysis and role asks.

For example, if a player is successful because they generate 20 controlled entries in the game but the maximum an NHL player, Connor McDavid or Jack Hughes, only gets 15, how do we expect that to change their projection? Are they a rare talent that can contribute without puck touches and leave space for elite puck carriers? How do these factors translate to NHL performance.

There’s some preliminary work being done by Mitch Brown and Will Scouch but I think there’s plenty more fruit on the vine. I have a hunch that time-to-contact, cohesive pressure and plenty of less-measurable factors (unforced defensive errors, poor positional discipline, on-puck experimentation) probably signal a big difference. The difficulty in measuring prospects across different leagues then would be in creating a data disparity but, for the moment, I digress.

In order to understand how a prospect might fit, and how their skills contribute or detract from certain responsibilities, we would have to be able to compare them to like players. In order to compare them to like players we’d have to have a way to define those like players. Ideally, that means using microstatistics to generate simlar role players. We can also just use expert knowledge to assign roles but that would take manpower to be used across all prospects/NHL players and be prone to specific scout-subject bias (or take even more manpower).

Plenty of this curiosity comes primarily from mistakes the Blue Jackets have made. Why did Cole Sillinger and Kent Johnson slump? Why didn’t Patrik Laine and Johnny Gaudreau work together? Is there anything we could have done to anticipate the success or failure prior to them joining?

Why do some transactions fail? Why did Vladislav Gavrikov “look terrible” via analytics and then look like one of the best defensive defenseman in the league when joining LA?

You might notice that plenty of these research questions come from Blue Jackets specific problems. I believe having intimate and specific knowledge of a single team is a fruitful path forward to understanding greater league-wide conclusions as well. My breakdown of Florida’s defensive system came from a desire to understand what they were getting right about playing defense because the Blue Jackets seemed to struggle so much.

Moving between large gestalt work and small specific work is a cycle that I find helps my creative process. Finding the rhythm between big picture and detail is critical.

Research Questions

• League Differences in Events

  • Transition Attempts

  • Transition Pace

  • Puck Recovery Pace

  • Offense Volume

• Difference between Early NHL Season, Late-Season and Playoffs

• Microstats Role Assignment

  • Development Cohort Growth Curves by Key Performance Stats

    • Puck Carrier growth in Transition Game Score

    • Forecheck Gamescore Growth

  • Role Changes from Junior to NHL

  • Role Rarity (Establishing Market Supply)

    • In-Demand Roles

    • Drafting or Trade Value Impact on Role Rarity

  • Role Stat Benchmark

  • Line and Roster Construction w/r to Roles

• Quality of Competition

  • Stat vs Skill(A3Z)

    • xGF(or whatever) vs Buckets of Denial% Defenders

    • xGA(or whatever) vs Buckets of Carries/60

  • Player Specific Performances

    • Share Stat (GF%, xGF%, CF%, etc) vs Bucketed GAR by Position

• Linking A3Z tracking (shot assists, game state) with xG Model

  • could also create On-Ice Metrics

I wont go too far into this but suffice to say that I’m highly curious and want to learn a lot about any of these various research questions but don’t yet have the capabilities to explore them. I don’t know how quickly I can get there but I’m eager to get moving.

Some of these are already being done but I think learning to do them myself would help to expand my knowledge base and empower my ability to continue to do them going forward. I find visualization a very intellectually stimulating exercise and would really enjoy being able to engage in that medium.

Some of this research could also be helped with more expansive and specific data, which brings me to the next section.

Tracking Project

I started this past season believing I would track every game. I even tracked the first 5 games, some multiple times as I dialed in the breadth of the tracking project, but moved on as they were taking significant time (I wanted to track every pass) and I didn’t have the knowledge, or data flow, to easily extract conclusions from there. I didn’t think I could develop that knowledge while trying to keep up with the workflow and ESPN started deleting the archive of games.

From there, I pivoted to the Next Day Reviews partially so I could have a selection of video to work from.

From here, in addition to the above skills, my primary goal of the upcoming break is to prepare the tracking project for day-to-day use. Prior to that, I would love to do a sort of “White Page” for the tracking with information behind the inclusion of the statistics done from whatever studies I can find.

For the most part, these tracked stats should be used only to understand more behind what’s happening on the ice. Unless proven, they don’t necessarily predict anything but they can help us understand the “whys” while finding room for adjustments to create better results.

Ideally, I’ll create a very good dataset with which to draw conclusions in the summer of 2025 and one that can help connect some interesting dots with regards to systems, roles, etc if rigorously defined and paired with other forms of analysis. This, also, brings me to the next part.

Next Day Reviews

I started next day reviews primarily to get into a working rhythm but also to bridge the gap between the feeling of the outcome of a game to the slower, more methodical understanding outside the immediate emotional context (sleep is a fantastic way to reset the nervous system).

We all know the ups and downs from comeback victories and devastating losses, those won’t be going away. We can then extend the “content” by taking a different perspective the next day. Really, I wanted to tell the story of the game within the game or the game of tactics within each NHL game and that’s only accessible away from the real-time.

Knowing when I was going to work helped me find better time to relax without worrying about whether I was creating at a sufficient volume. Furthermore, I believe I do my best work when I’m immersed and through that immersion, along with free association brain time, find lines of inquiry (immersion→inspiration). Not only was I consistently expressing thoughts and opinions but I was also cultivating an understanding that made the next ideas come freely as well.

Also, I just like watching film. It was the primary vehicle for self improvement in my ultimate frisbee career and I love taking time to watch and rewatch multiple events from multiple perspectives. I believe that there are beautiful details to be discovered and it’s a level of detail and nuance that I am consistently frustrated is not made accessible to a regular viewing audience.

Sports fandom is difficult and, in my opinion, made more difficult, if not radical, simply because people don’t have deeper outlets with which to express their fandom (82 games in a season is also difficult to build a balanced life around). TV broadcasts do not allow for deeper understanding and, at least based on camera angles and color commentary tendencies, don’t want it either. That’s largely fine but I certainly believe there’s more joy to be found going deeper.

From there, I decided to move forward with Next Day Reviews. I think it went well and it certainly helped me improve my understanding of the game. However, it was often rooted in only things I was looking for and that made the overarching film analysis, especially at the end of the season, slow and difficult. The clip library was just poorly organized.

From here, I believe the tracking project and Next Day Reviews can both work to enhance each other. Instead of starting with exclusively matchups and shot data, though it is probably the best starting point, we can also utilize game-to-game differences in the tracked data. If the Blue Jackets decided to clear the zone repeatedly instead of exiting with control, why?

Similarly, if we can utilize Next Day Review as a sort of “journal style” subjective analysis and add tags for research questions (like “Matchup Line”, or “Elite Forecheck” or 1-3-1 Neutral Zone Trap or “Elite Puck Carrier”), perhaps there’s also more fruit so to speak.

If we take a survey of fan sentiment post-game and search for associations in the real on-ice data perhaps we can discover the bridge that drives sentiment? Do certain things take more weight in our subconscious or emotional reactions? Can we use that to help drive better immediate takeaways?

Designing a sort of season long survey with appropriate tags or characteristics might be quite difficult but it sounds like an interesting challenge and one that should be fun to tweak over time. Perhaps this can be a step towards systems matchup analysis or maybe a way to coach coaches, eventually.

At the end of the day, I would love to create 10 game breakdowns or “analytics/tracking/systems” check-ins. From what I understand, this is the interval for analysis that Jared Bednar and the Avalanche use for their players so why not imitate and see what we can do?

Post-Game Shows

An empty space, at least for the Blue Jackets, and one that could provide an immediate counterpart to the slower brain “Next Day” deeper content is a post-game show. Television networks love post-game breakdowns with interviews, and rightly so, but they are leaving the possibility of live-stream interactivity on the table.

There are a variety of places on YouTube that I believe are already doing this, the first in my awareness being Emerald City Hockey who covered the Kraken.

In these shows we can spend time talking about quick hits and interacting. If done well we can acknowledge the ups and downs of fandom and then move forward into understanding. This could be a way to “crowdsource” questions to answer in the slower review the next day or simply a way to have some good discussion with fans from anywhere.

If the tracking project is lightweight enough (likely just transition data based on my previous games tracked), perhaps post-game reports can be included as part of the show. If not, HockeyViz has excellent resources to springboard a discussion.

I am sure I am not the only person who hasn’t been able to find great hockey discussion in the most frequently available in-person or online spaces. If I could provide that or build a sort of community, I think it would be a fruitful endeavor.

Bringing it All Together and Thank You

I started writing in this space because I was spending a lot of time looking at data (shoutout HockeyViz, EvolvingHockey and AllThreeZones for nearly endless tables and visualizations) and watching Blue Jackets games and I wanted to share what I felt like were good insights. I’ve listened to and read very smart people talk about what they were doing (shoutout Dimitri Filipovic and the PDOcast, Alison Lukan especially she still worked for the Blue Jackets, any and all talented writers at the Athletic but especially Jack Han, Ryan Stimson and Darryl Belfry and too many others I can’t even name). Furthermore, I think the actual doing, the writing in most cases, helps me clarify and develop my own ideas.

Previously, I have had the tendency to want to make sure I’m perfectly prepared to take the next step into whatever I was doing. I didn’t want to be a leader for a team unless I was absolutely sure I was the best, I didn’t want to start a project unless I knew I could do it. In many ways, I usually revel in doing the things that need done rather than doing things with higher-order preparation or ambition. In starting this blog, I eschewed many of those perfectionist procrastinating habits and just started doing it.

From that perspective, I’m absolutely thrilled with how the first season and a half went. The feedback that I get is nothing short of incredible and I think I’ve pushed myself out of the comfort zone and am now reaping the accomplishment rewards. Now, I think, is the time to do all of the dirty work to bring my little publication forward.

When I started, I at least wanted a sort of portfolio of the sort of things I could do and wanted to continue doing. Something I could show someone instead of just saying “trust me I can do this”. I wasn’t, and to an extent, am still not sure that this thing is the final place for doing. As I’ve watched The Athletic disintegrate and “journalism” continue it’s late-capitalist doom spiral, I’m still left without a clear idea of where it all goes.

Create first, worry about what it all means later.

I should maybe come up with a better name, probably get some actual branding here and on the potential YouTube and maybe even complete the Substack landing page and whatever emails would help explain what exactly I do here.

Tracking, data analysis, post-game shows, film breakdowns… it’s all quite a lot. I’m often overwhelmed by my sheer desire to do things. Picking a college major was difficult for me (I landed on Astronomy and Physics before Ohio State introduced Neuroscience) because I had too many interests that I wanted to pursue and not because I had no idea.

So now, here I am once again in a familiar position. Too many things, not enough time. While I’d love to do each and every one of these things, it’s most likely that I’ll have to pick and choose. I have the vision for all of it but I don’t know enough to even have an estimate at what’s possible. I must, once again, wrestle with the long and the short, the big and the small, the gestalt and the detail.

Perhaps the tracking project will have to be stripped down, the post-game shows simply not feasible or Next Day Reviews having a changed format. Perhaps I need to admit that if I was actually going to learn Python to the depth I’d like it would have happened some time ago. I’m not sure yet. What I do know is that I’m really enjoying the work and, especially if you’re still reading, I appreciate anyone paying attention.