This recent discussion has fascinated me.
First, I "picked" Luneau in the first round
here.
Second (you can quit reading because I admit to not watching juniors), I do think there are ways to improve what professional scouts do.
1)There was obviously artistic license, but I believe "Moneyball" depicted some truths about scouts. I believe many (not all, maybe not most) professional scouts are blinded by their own expertise. This belief stemmed from my fascination with behavioral economics and its findings--like those depicted in "Predictably Irrational" and "Thinking Fast and Slow." For instance, pre-draft interviews are still used and some think they are really important. Most academics who have studied the subject have found that interviews don't help, and often hurt, candidate selection in many fields. One of the main reasons I starting paying attention to prospects is because I was trying to create a model that identifies some defensive prospects that scouts likely under-value.
2) I have been refining the model for 5+ years. The results have been strong (at least from my perspective). For instance, last season the model identified Jake Furlong, Adam Engstrom, Tomas Hamara, and Frederic Brunet as D who should have been
drafted anytime after the middle of the second round.
3) There is currently no model for forwards (and goalies are, well, goalies). But I am tinkering with several factors that may be valuable. The most intriguing is GWP% (games with points). I started wondering if players who scored more consistently were undervalued. I tried using this in my picks in last year''s Basement Stakes. I would say the one resounding success was picking Gendron in the 7th. My list is heavy on D, since that is my focus. The objective to pick as Buffalo. With Dahlin, Samuelsson, and Power all being left-shots there is little reason to believe the Sabres will have room for another in the next decade. That explains why Luneau was high on my Basement Stakes list (though behind Gaucher, a player mentioned several times in this discussion) and also Salin's position. My model didn't single Salin out because his "consensus" ranking was end of the 2nd round--so him going in the 5th means he was in fact under-valued by the NHL scouting community.
4) No way would I argue that I am better than most NHL scouts at drafting. I would argue that my model has come far enough that it does somewhat better than the majority of teams at drafting D-men after the 2nd round. Though there is a team or two each draft that seems to do what my model suggests: Brock Faber was drafted higher than his pre-draft rankings as my model indicated he should be; Montreal did a great job taking Adam Engstrom in the 3rd even though he wasn't on any top 100 list.
5) It is at least arguable that the three first round picks I made in the Basement Stakes (Savoie, Kulich, Luneau) were as strong as Buffalo's picks (Savoie, Ostlund, Kulich) given the facts that most now have Kulich rated at least equal to Ostlund and Luneau looks like a perfect future partner for Power. I also think Salin/Furlong are preferable to Lindgren/Komarov at 106 and 134.
I have two arguments for why it might be possible to do as well as professional scouts. 1) They likely have biases for things like size (which I use as one of many factors) and intangibles that lead to group think. 2) A model-heavy approach can "see" every game while most scouting staffs only get at most a dozen viewings of a player taken after the first two rounds.
As I said, I am fascinated in both the Luneau and the scouting parts of this discussion.
