This is the sG model (synthetic goals)…

It looks at player performance levels by combining multiple different things:

  • Finishing
  • Setting
  • Offensive transition
  • 5v5 offensive chance generation (expected Goals)
  • Power play offensive chance generation (expected Goals)
  • Penalty drawing
  • Penalty taking (or lack there of)
  • 5v5 defensive chance prevention (expected Goals)
  • Penalty kill defensive chance prevention

It adjusts all of these based on usage factors (who you play with, against, zone starts, game trailing/leading, etc.) but also prorates everyone to the exact same baseline of ice time deployment.

Enjoy.

4 comments

  1. I was expecting the Ehlers vs. Connor chart, and nearly had a conniption fit when I saw who the orange line was actually modelling

  3. I appreciate that you haven’t expanded speculatively on this presented data. It could be something for everyone:

    — Petan-truthers were right
    — Connor is no good
    — analytics are garbage
    — analytics are nuanced, not monolithic
    — environment and context matter
    — nothing matters, god is dead
    — cats and dogs are different and it doesn’t matter
    — cats could be dogs with different deployment
    — cats shouldn’t be forced to be dogs, let people enjoy things
    — etc

    A true firecracker, and a true vacuum of context.

Leave a Reply