• Moderators, please send me a PM if you are unable to access mod permissions. Thanks, Habsy.

Around the League 2019-2024 Edition

I guess one of the core issues is whether the model is meant to be predictive or reactive...and of/to what.

A 3rd pairing defender who absolutely molests cans for 14-15 5v5 minutes per game can have a bigger influence on the game than a bad 1st pairing defender getting caved in. It's entirely appropriate for a ranking system to rate that player higher than the bad 1st pairing guy.
 
yeah but the models were clearly overvaluing sheltered dmen, no question.

someone needed to do some un-obvious legwork on this instead just calling all coaches and execs dumbasses, and Dom did it and seems to have been rewarded with a better working model for now.
 
zeke said:
yeah but the models were clearly overvaluing sheltered dmen, no question.

someone needed to do some un-obvious legwork on this instead just calling all coaches and execs dumbasses, and Dom did it and seems to have been rewarded with a better working model for now.
Click to expand...

But again, that comes back to asking what the model is trying to do. If it's just reporting the on ice impacts of each player as they're being played....maybe that player type is actually underrated (can crushers) and guys who do heavy usage poorly hurt their team more than we think.

If the model is trying to be a rating system that accurately predicts how gud everyone actually is, like WAR but for hockey....well yeah, that needs sorting out.
 
the LOF perspective, I suspect, would be that because Patches is an American, the Toronto-Montreal rivalry is not relevant to him. I suspect.
 
You must log in or register to reply here.
Back
Top