Quote:
Originally Posted by SebC
Please, show that NHLe has diminished predictive value for "projects".
|
The purpose of NHLe is to project the numbers of points a player could expect
next year.
It uses averages from all the players that have walked that same path in the past.
A project, by definition, is someone who is going to take longer to develop.
So averages of other players their age are kind of pointless as an argument against them. Yeah, we know they are going to have lower point totals right now because, you know, they're a project.
Quote:
|
That's the point. Projects are part of the NHLe dataset. Basically what the strength of NHLe's correlation tells us is that a low-scoring "project" is code for "bad pick" - unless of course there actually is evidence otherwise, in which case it should be possible to present such evidence.
|
Um no, it's not code for a bad pick. A late bloomer isn't going to be comparable to others at a young age, because... late bloomer.
So throwing averages from young players at them as an argument that they are no good, is rather pointless IMO. Well, to be more clear, it's silly if you are doing so as the argument for a player being a bad pick (like the article did).
The most important thing with prospects is progression. Jankowski continues to progress nicely. If someone wants to use NHLe as one means of showing where he is currently vs his peers, I'm all for it. And yes, he still has a ways to go.
But that is not what the author did.