Harbour Dog
Registered User
Seriously? Something is WAY off if they're saying that Kreider's shot there only goes in 1 out of 8 attempts on that play.
Do the numbers literally just look at the type of shot and the area of the ice it's taken from?
Post-Game Talk: The Official Matt Rempe Hart Trophy Campaign Thread
- Thread starter Machinehead
- Start date
Seriously? Something is WAY off if they're saying that Kreider's shot there only goes in 1 out of 8 attempts on that play.
Do the numbers literally just look at the type of shot and the area of the ice it's taken from?
That's what I was thinking about. Maybe it's taking into consideration plays where the scorer is in that part of the ice on the PP for example where they are getting defended and not taking into account a 2 on 0. The last 2 on 0 I remember us actually scoring on was like Hlavac against the Sabres in 2001. I'm sure we've scored on them since then and 2 on 0 is rare but they're not automatic.
As far as I know, Moneypuck and most (all?) public expected goal models don't take pre-shot passing into account, because that's not tracked in the NHL's public data. It's why previous iterations of the Rangers that favored quality over quantity were generally viewed more favorably in private models that took those things into consideration.
- Mar 1, 2018
- 365
- 684
A lot of negativity on here despite playing the whole game without taking a bench minor.
Machinehead
Fox Mafia Consigliere
Pretty much.
The fact that it's a 2 on 0 makes the puck movement extremely unpredictable, but a lot of models are not going to differentiate Kreider's shot from an in-zone chance from the same location.
This is definitely an issue and the type of modeling that Moneypuck uses is already a bit antiquated.
I don't agree with all of Vally's takes nor do I think his model is without its own warts, but his model does way more to differentiate quality.
