Not a fan after watching some of them last year, haven’t seen this years. Seems like they just watch a game or two of film and use single plays to make huge sweeping statements about the players, I disagreed with several of them, though everyone has different opinions.
I really think the Lucius kid is going to be a 1st line caliber player.
Lucius projects highly in this class. There are 10 players in this class that project a higher "star potential" than Owen Power, and Lucius is one of them.
That list has Matty Beniers at a 43% chance of being an NHLer?Lucius projects highly in this class. There are 10 players in this class that project a higher "star potential" than Owen Power, and Lucius is one of them.
Looks like it's an NHLe-score based list to me, which IIRC is totally production based (i.e. looks at NHL-"equivalent" production based on the prospects point totals in a given league at a given age). Fun for easy quick charts to look at like Byron Bader throws out all the time, but fairly useless for any actual player eval in my opinion.
makes sense i suppose. just brings you back to trusting your eyes i guess. he wont be a zegras but he sure looks like he can be a quality C in the league.
I think you would be quite surprised how much scoring rate in feeder leagues actually translates to NHL success. But like with anything, it's only 1 piece to use for evaluation.
Wouldn't be that surprised really, everyone remembers that "potato vs. Canucks GM" article from years ago. In the aggregate it will generally get you more productive pros. When looking at individual players, it doesn't provide a ton of info though. As you say, one piece (though obviously a big one).
The problem with the chart above is how do they account for Power, Beniers, and Johnson being in the NCAA’S a year ahead of schedule?
You don’t see players with his size and skating outside the NHL, which makes me think he’ll certainly play. But, I wonder if he’s more than a bottom pairing guy.
How are you determining that they're in college a year ahead of schedule? This model is standardized allowing for an apples-to-apples comparison between the scoring rate of two prospects in different leagues. Age, weight, height are factored into it as well.
Are there any good retrospective analyses of previous NHLe projections? I've seen the model thrown around for a few years now so I assume someone's gone through and looked at whether it hits more than it misses, but I haven't been able to find one.
Patrick Bacon's (the one I linked) is self-analyzed. He dives into that in the 4-part series. He touches on previous NHLe models, advancements, pros and cons of older models & alternative models, how he addresses biases, etc, etc.
Very few players spend their draft year in the NCAA. The fact that guys get more credit for tearing up the USHL than they do for being good players in the NCAA’S tells you that the formula is flawed
I see where he explains some of the testing he did to calibrate the model against existing data, which is most of Part 3. But the test of a good predictive model is usually making predictions and seeing how often those predictions come true. Running an area-under-curve analysis of draft classes where we already know that 70-some % aren't ever going to make the NHL seems like kind of a weird way to test it (which, to the author's credit, he seems to cop to in Part 3).
The USHL has an NHLe of .143 whereas the NCAA has an NHLe of .194
Because scoring at a PPG pace in the NCAA is quite a bit more impressive at 18 than it is at 23.
