Quote:
Originally Posted by Itse
This.
It's also worth remembering that its extremely unlikely that stats could even exist which would somehow reliably predict Tanev being better in Calgary than in Vancouver. Statistics can only really tell you about things that have already happened, the rest is always interpretation. Stats can also only tell you what the numbers are, they can't really tell you why they're like that. There's always going to be things ypu can't measure, or don't think of as a possible explanation.
This is really the major reason why I dislike so much of the current fascination with statistical analysis; so many of the conclusions people draw about what's a good trade or bad for example are based on the fundamentally faulty premise that it's possible to reliably predict how a person will perform in a completely different situation using stats alone. It's just not what stats really do.
Decisions like "is Tanev a good signing for Calgary Flames for the 2021 season" will always in the end be an educated guess. Better stats can make that a more educated guess instead of a less educated one, but that's as far as they can take you.
Right now it seems that the Flames made a good educated guess when it came to Tanev.
It's what we pay the pro scouts for, so that's probably where the credit belongs.
|
Not sure I agree.
We keep hearing about proprietary models that can take goalie x and his save percentages by type, and plop him into team y and their propensity to give up a certain type of shot.
The math isn't that hard to run after you have those two sets.
So why not have that for defensemen?
"Chris Tanev is poor on the rush, and poor when stuck in his own zone for more than 65 seconds. His strengths are blocking out lanes, and taking away cross seam passes. Needs a partner to compliment that by also being strong positionally, and somewhat predictive"
Quinn Hugues and Noah Hanifin are very different. Calgary spends less time getting lit up than Vancouver did last year ... Vancouver was in the top five in many "against" categories, including shots and shot attempts. They spent a lot of time in their own zone.
So yeah I think you can have a model that would allow you to plop a certain player into a different situation.
If they don't ... it's coming.