Had to do some training with constructing an AI model + dataset in my spare time to prep/POC for an actual work initiative. Fairly basic for those devs floating around here, but I ended up throwing it some team/individual data that I was able to scrape (but not verify...because time/pain), as well as a spreadsheet I created with my own prospect rankings.
View attachment 919431
I haven't gone very far here with my prospect list — I've accomplished what I needed for the original purpose, but might keep tweaking for fun. Thomas and Kyrou are essentially the controls here. And I just took the 8 "prospects" that I've 1.) gotten the best views of and 2.) foresee being a part of the logjam.
Aside from the team/individual performance data, I also had it look at Armstrong's trade history, and found a separate dataset for all NHL trades the last 3 years. I messed up here, as I didn't run a control on the results with/without that data set, but: Stated objective was to create a top 6 with some spare (ideally, unique) parts that can float between bottom and middle 6. And to add a tradeability and redundancy score next to each player.
Prospect | Position | Tradeability Score | Justification |
---|
Dvorsky | C | 1 | Core future player |
Snuggerud | W | 3 | Potential asset for trade unless trending toward ceiling |
Stenberg | W | 2 | Valuable. Tradeable in a package deal for elite talent. |
Neighbors | W | 2 | Important. Only tradeable once trend stabilizes and only if talent (>) |
Bolduc | W | 3 | Neutral; depends on team needs |
Dean | C | 4 | Expendable; trade candidate |
Peterson | C/W | 3 | Versatile but tradable. Potentially ascendent due to physical characteristics. IQ is the cap. |
Kyrou | W | 2 | Tradeable but requires (>) and known production coming back |
Thomas | C | 1 | Cornerstone team member |
High confidence in the Center positions. Wings lack elite skill, capping shot quality and reducing production. Recommend trading non NHL assets — ideally 3 and below — for Elite Wing talent.
Known Targets: Brock Boeser and Travis Konecny
Emerging Players: Kirill Marchenko and Jack Quinn could be strategic acquisitions for both immediate impact and future growth.
Not meant to mean much, but wanted to share for fun. Obviously, defense is a driver here that is not represented, and, my rankings certainly swing the model more significantly than the generalized production data would've on its own. Last note: I added Buch to the model, and it didn't change a whole lot other than giving him a score of 2 (due to his age).