DCOrange Posted April 5, 2018 Posted April 5, 2018 (edited) Really interesting article, basically arguing the opposite of my Scouts Inc thread...this article basically says that this year's QB class is weak, and notably gives guys the following comparisons based on projected success at the next level (listed in order from high to low success rate). As you'll see below, the comparisons are downright frightening, but I would urge you to read the article anyways because it's a really interesting look at things using some advanced metrics. Baker Mayfield: Russell Wilson Sam Darnold: Colt McCoy Mason Rudolph: Jared Goff Logan Woodside: Nick Foles Riley Ferguson: Zach Mettenberger Nick Stevens: Mike Glennon JT Barrett: Jake Locker Luke Falk: Brett Hundley Lamar Jackson: Brandon Weeden Mike White: Jacoby Brissett Josh Rosen: Brandon Weeden Chase Litton: Tim Tebow Josh Allen: Deshone Kizer https://www.sbnation.com/nfl/2018/4/5/17046116/2018-nfl-draft-quarterbacks-josh-allen-sam-darnold-projections-hype Edit: I should mention that this study focuses solely on passing, so for example, it would be reasonable to expect Lamar Jackson and probably Josh Allen and JT Barrett to outplay the expectations from this study. Edited April 5, 2018 by DCOrange
4merper4mer Posted April 5, 2018 Posted April 5, 2018 Based on this input only, if given the choice, I would opt for Mayfield. Lalueleluelaluletta doesn't even make the board? 1
DCOrange Posted April 5, 2018 Author Posted April 5, 2018 Just now, 4merper4mer said: Based on this input only, if given the choice, I would opt for Mayfield. Lalueleluelaluletta doesn't even make the board? I'm not certain, but I don't think the stats used for this study exist at Lauletta's level of competition.
Thurman#1 Posted April 5, 2018 Posted April 5, 2018 (edited) That's a bizarre use of stats. Football Outsiders uses these stats to grade the whole offense, not the QB. The writer pretty much agrees with this, saying you can't separate players out. Yet he's doing so. He also seems to use pro stats, but only from the first four years. Which would really help fast starters and grade down guys who started slower. And because he uses some (reasonable) ways to cull the guys he might have looked at and is only looking at QBs from 2010 or later, he has only looked at 38 QBs college-to-pro results so far. He points out that a guy like Rosen who sat for a year in college has a big advantage over guys who played early, and says the same thing happens for QBs who sit for a while as a pro at the beginning of their career, these numbers are better when they start later. I read most of the article, but it's night over here. I'm beat. Looks like an interesting avenue of approach but I don't he's putting this out there as much more than an interesting idea that he has to work more on, that shows some promise. He likes it better in showing ceilings. This is early in looking at this method. It's interesting, though. Edited April 5, 2018 by Thurman#1 1
DCOrange Posted April 5, 2018 Author Posted April 5, 2018 2 minutes ago, Thurman#1 said: That's a bizarre use of stats. It is, but the section about a QB's ceiling was particularly interesting to me considering that there were zero outliers in the dataset.
Cash Posted April 5, 2018 Posted April 5, 2018 41 minutes ago, Thurman#1 said: That's a bizarre use of stats. Football Outsiders uses these stats to grade the whole offense, not the QB. The writer pretty much agrees with this, saying you can't separate players out. Yet he's doing so. He also seems to use pro stats, but only from the first four years. Which would really help fast starters and grade down guys who started slower. And because he uses some (reasonable) ways to cull the guys he might have looked at and is only looking at QBs from 2010 or later, he has only looked at 38 QBs college-to-pro results so far. He points out that a guy like Rosen who sat for a year in college has a big advantage over guys who played early, and says the same thing happens for QBs who sit for a while as a pro at the beginning of their career, these numbers are better when they start later. I read most of the article, but it's night over here. I'm beat. Looks like an interesting avenue of approach but I don't he's putting this out there as much more than an interesting idea that he has to work more on, that shows some promise. He likes it better in showing ceilings. This is early in looking at this method. It's interesting, though. Yeah, credit to him for showing his work in a way that Football Outsiders often doesn't, but I think he reaches some dubious conclusions. First off, a 0.27 correlation is barely anything. A century ago when I was in school, we were taught that 0.3 was the minimum correlation that could be considered significant. Second, despite correctly noting that marginal explosiveness has essentially no correlation to QB success, he still throws into into his analysis anyway.
Recommended Posts