Originally Posted by patteeu
Yes, I know the article doesn't specifically discuss Geno Smith, but it treats draft pick slots as generic players and talks about value of that pick in terms of previous statistics instead of the individual attributes of that particular prospect. I'm not criticizing the article, I actually think it's very good. I'm pointing out that it has the same predictive flaw that those who use the draft position of prior super bowl winning QBs to make a case for "trying" by taking a QB in the first round (or with the first pick) in a particular draft.
It's probably pretty good for predicting results on average over a long period, but it's not so good for predicting results of a specific pick. For example, if the 1.1 pick is used on a QB this year, it's pretty unlikely that they'll end up with a QB whose Career Approximate Value is extremely close to that of Brad Johnson (like +/- 1 or 2 points). Like someone else said, it's probably more like a bell curve distribution around BJ.
I'm not being critical of the article. I'm being critical of anyone who would try to misuse the article.
Therein lies the rub. First of all, the sample size is really small, just 62 number one draft picks; secondly, in terms of QBs the sample size is shrunk even more since not all #1 picks were Qbs; thirdly, there is a wide variance of standard deviation because of extreme outliers like Peyton Manning and JaMarcus Russell.
I suppose that is why I asked what are people's definitions of a "franchise QB."
Is a franchise QB defined by how many pro-bowls they are selected?
Is a franchise QB defined by how many Super Bowls they win?
Is a franchise QB defined by how many times "they will their team to win"?
Is a franchise QB defined by how much extra revenue they generate for their team?
How long do you give a QB to prove he is a franchise QB?
I don't know.