You have a pretty serious multiple endpoint problem here. Your endpoint cuts out the two most successful RBs of the past 20 years, and you count Zeke as having only one top season. Well, OK, Zeke has had only one top season so far, but it's reasonable to expect that he'll have a career at least as good as Lynch's, and possibly as good as Peterson's or Tomlinson's. So even excluding Faulk and Tomlinson, out of 15 picks you are looking at three massive studs vs. only one clear bust, with two other very promising young RBs in Gurly and Gordon. You could get five studs out of 15.
I'd say top RB picks look like they're hitting at about the same rate as top picks at other positions.
In other threads I have pointed out that the hit rate on RB's has averaged typically about 2-3 per year based on having a Career AV of 50+. (More on that in a second). But if as you say, I am cutting out the endpoints of a 20 year stretch (in this case really 15 years), you are only reinforcing my point, not debunking it. If I am excluding guys at the ends of nearly a 20 year range, that means there were not a lot of successful players in the middle.
That being said, my 15 year cross section of RB1's included effectively all of Tomlinson's career and the primes of several other backs (who just happened to have been drafted before 2002). I didn't include the names in the yearly RB1 chart because it was the spot of the draft selection that was more important than the names, but there were plenty of early draft picks on that list.
Going back to the 2000 draft, here are all the drafted RB's that posted a Career AV of 50 or more. Granted, that is a different metric and has no real practical application to fantasy football, but it at least
2000 Lewis, Alexander, TJones,
2001 LT, McAllister
2002 Portis, Westbrook
2003 McGahee, LJohnson
2004 SJackson, Turner
2005 Gore, Sproles
2006 MJD, Bush, DWilliams, Addai
2007 ADP, Lynch
2008 Forte, CJohnson, Rice, Charles
2009 McCoy (Moreno 33, RJennings 31, DBrown 25)
2010 None (RMathews 43, Spiller 34, Starks 27)
2011 Murray (Ingram 34, Ridley 24, Vereen 23)
2012 None (DMartin 33, LMiller 30, AMorris 30)
2013 None (Bell 43, Lacy 34, Bernard 29)
2014 None (Freeman 29, Hill 20, JWhite 13)
2015 None (DJohnson 24, Gordon 15, Gurley 12)
2016 None (Elliott 16, Howard 9, Booker 5)
(Only 2 guys that were UDFA that came into the league since 2000 met the 50 AV criteria: Fred Jackson 57 (2007) and Arian Foster 63 (2009).)
That's 25 players with a lot of the more recent guys too soon to evaluate. Of those 25 players (27 if we count the UDFAs), only 6 were Top 15 overall draft picks (Tomlinson, Peterson, JLewis, Lynch, TJones, and Bush). Early RB picks have had a better hit rate than other positions, but I would have to research more where higher ranking players were actually drafted to get a better sense of where the best places are to draft certain positions.
By comparison, here's the overall positional breakdown of Career AV across all positions in that same time span (players drafted from 2000 - 2016 with CareerAV of at least 50):
LB: 40 (13 of those were Top 15 picks out of 28 total selected as Top 15 picks = 46.4%)
OT: 34 (11 Top 15 picks out of 36 total selected = 30.6%)
DE: 33 (9 Top 15 picks out of 32 total selected = 28.1%)
DB: 33 (10 Top 15 picks out of 35 total selected = 28.6%)
WR: 29 (6 Top 15 picks out of 33 total selected = 18.2%)
DT: 27 (9 Top 15 picks out of 32 total selected = 28.1)
RB: 25 (6 Top 15 picks out of 19 total selected = 31.6%)
QB: 23 (12 Top 15 picks out of 29 total selected = 41.4%)
G: 23 (0 Top 15 picks out of 4 total selected = 0%)
C: 9 (0 Top 15 picks out of 0 total selected)
TE: 5 (0 Top 15 picks out of 5 total selected = 0%)
PK: 0 (0 Top 15 picks) (0 Top 15 picks out of 0 total selected)
P: 0 (0 Top 15 picks) (0 Top 15 picks out of 0 total selected)
Note that the players are listed by the position they were listed in when they were drafted, so if someone switched positions they would show up under their original position.