The Athletic - NHL awards watch: Leon Draisaitl leads a fascinating MVP field

[Craig Button]

Marner of Kaprizov for Hart

Kaprizov Mitch Marner Shorthanded Time On Ice | StatMuse

Kirill Kaprizov has been on the ice for 0:12 while shorthanded this season.

www.statmuse.com

Kaprizov Mitch Marner Average Short Handed Time On Ice 2024-2025 | StatMuse

Kirill Kaprizov has spent 0:00 on ice per game while shorthanded this season.

www.statmuse.com

Kaprizov Mitch Marner Takeaway 2024-2025 | StatMuse

Kirill Kaprizov has gotten 9 takeaways this season.

www.statmuse.com

Pretty sure this model values center's more than wingers for defensive metrics because they have more responsibility.

That's only true if they don't know anything about hockey

 

[nowhereman]

Hughes is the answer for both the Hart and Norris

 

[Regal]

Why? It’s reasonable to assume that Makar wins the Norris. He may be having a slightly worse defensive year but he’s having a better offensive year which offsets that. Makar has proven to play at a consistent level whereas I think Hughes utilization is going to show signs of fatigue in the 2nd half. He’s getting overworked due to the current defensive structure of the team and it’s already showing.


Side note: 99% Dom’s models are typically trash though.

Makar’s defensive numbers are slightly better this year. I don’t think I’d say Makar’s been better offensively either given the seasons their top forwards have had and the fact Quinn has a better PPG. I’d say it’s Hughes’ 5v5 play driving that’s the biggest plus in his favour. But it’s close enough that it’s fair to give it to Makar for staying healthy, though the Avs have also just played more games so far.

 

[syz]

How the hell is Shesterkin so high on the Vezina list?

There aren't many good goalies in the NHL anymore so somebody has to fill the list.

 

[JuJu Mobb]

Dom's model...

 

[jani]

Hellebuyck at 9 (Hart) is a joke. Should be 1.

 

[ijuka]

Dom's model...

The main flaw with it, I believe, is how much more it weighs offense over defense. Almost all the good players have massively higher offensive contributions than defensive ones.

Shouldn't the system be normalized so that offensive and defensive contributions both are centered on the average, and deviations are represented as standard deviations from the expectation? That would make far more sense.

Also, the idea of just adding the game scores together the way he's doing is also not exactly statistically valid. It ensures that only the players who have played the exact same number of games are comparable.

My preferred method would be representing every player's scores with both mean and variance parameters, assuming the stat itself is even worthwhile to track.

 

[Craig Ludwig]

Adam Fox on the dumpster fire that is the Rangers? Replace with Hutson on a much improved Habs team, where he is their #1 Dman and logging 23 minutes a game. And yes I am a Habs fan but Hutson has been a huge addition to the team.

 

[NinjaKick]

those damn Germans... what are "they" up to now






aaaaaaaaaaaand I'll show myself out

 

[Regal]

The main flaw with it, I believe, is how much more it weighs offense over defense. Almost all the good players have massively higher offensive contributions than defensive ones.

Shouldn't the system be normalized so that offensive and defensive contributions both are centered on the average, and deviations are represented as standard deviations from the expectation? That would make far more sense.

Also, the idea of just adding the game scores together the way he's doing is also not exactly statistically valid. It ensures that only the players who have played the exact same number of games are comparable.

My preferred method would be representing every player's scores with both mean and variance parameters, assuming the stat itself is even worthwhile to track.

I think it suggests that the best offensive players can have more of an impact over average than the best defensive players, which makes sense. Defense is still more team dependent and less individualistic than offense. GAR charts on evolving hockey use standard deviations above and below average and no one has defensive impacts as high as the highest offensive impacts.