When viewed from the perspective of an individual player outward, I can understand how it would seem that "making it" to the next level is less predictable than it really is.
When viewed as a group from a birds eye view, a group of players has a quantifiable attrition rate. Don't quote these numbers, but it is something like 30% of five star players out of high school, 15% of four star players and 5% of three star players go on to the pro's. That doesn't mean that the players will go on to be successful, just that there is a quantifiable rate of pass-through from one level to the next. That also shows that the majority of players don't actually pan out, not if playing in the professional ranks is the measure of success.
I read something last night that blew my mind. In Stumbling on Wins by D.J. Berri and M.B. Schmidt, they look at the economy of the NFL draft. The findings are basically that the NFL draft, especially when looking at quarterbacks as an example, is really bad at determining actual value.
When looking at the performance of the first QB's selected from 1990 to 2004, only 5 had stats after five years that would consider them "above average." Those five were: Carson Palmer, Michael Vick, Chad Pennington, Peyton Manning and Steve McNair. When viewed another way, that is the top QBs that entered the NFL for each of those given years, only 6 of 15 were the first QBs chosen. Those were Carson Palmer, David Carr, Peyton Manning, Tony Banks, Steve McNair and Drew Bledsoe. Only Carson Palmer, Peyton Manning and Steve McNair were both the first pick at their position and had careers that could objectively be considered "above average."
Two of the top QB's produced during that period were undrafted (Jake Delhomme and Jeff Garcia), and 8 of 15 of the best QB's produced during that time were drafted in this order: 3,3,4,5,5,8,NA, NA. In other words, being first drafted at your position, especially if you are QB, has a weak relationship to your success.
For UT fans there is an interesting note. Both the BEST and WORST QBs drafted first during that period, using the methodology described in the book, came from UT. That is Peyton Manning (1998, first round, first pick) and Heath Shuler (1994, first round, third pick). Both were Heisman runner's up, right? In other words "draft position gets the quarterback on the field, but draft position doesn't appear to tell us anything about how well the quarterback plays when he's on the field." About Shuler, the authors noted that "f the Redskins were looking for leadership-and if people in Congress are examples of leaders-the choice of Shuler might make sense. With respect to production on the field, though, it is pretty clear that Washington's choice...didn't quite work out." After a comprehensive study, the authors' bottom line, is that "teams in the first round, especially at the top of the first round, should be making every effort to trade down."
Another interesting note: the authors addressed the factors that are statistically significant when drafting a QB, and factors that are statistically insignificant (in other words, here are the actual things that the numbers indicate matter to those choosing the QB, even though the data shows that where the QBs are chosen in the draft is unrelated to his performance).
Significant
1. Player's performance at the combine
2. Height
3. Body Mass Index (between 27.8 and 29.1 is most valuable).
4. 40 yard dash time
5. Wonderlic score
6. Playing for an FCS school
Insignificant
1. Performance of the Quarterback's college team
2. Race of Quarterback