Mastodon
@New York Rangers

Random predictions are not models



Random predictions are not models

by IgorSaysOK

5 Comments

  1. new-jersey7

    I could be wrong but I think their garbage model likes this team a lot better than last

  2. jahauser

    Hockey is such a dynamically flowing sport, where the positioning of one player a foot to the right or left has a ripple effect changing how every other player positions themselves and where the puck goes. And all the interdependent free flowing action happens by minuscule decision making, decision making which is so fast that we need a term ‘hockey IQ’ to describe a player that just does things with the right intuition.

    An innocent and independent seeming play like the particular way a defensemen exists the zone may not result in a goal, or any real hockey event, but the entire rest of the game has completely changed…every subsequent play is impacted by it.

    I both find the predictive and advanced analytics interesting but also don’t fully buy in. I don’t see hockey as being a sport that can be analyzed in the same way as say baseball. Call me old fashioned but that’s why I do think the eye test is important. Games are won on the ice and can happen in the most random ways. The butterfly effect aspect of such a dynamically flowing sport makes it tough for me to put too much faith in the numbers.

Write A Comment