Everyone's making good points, and no, there is no one right way to do this. I'm just going to explore around, statistically, and see what it yields.
.
Ideally, we'd have yards per game...but that's not readily available. I may have the list move around and evolve some, then include yards per game for the top 20, then finalize it. We'll see.
.
Imagine an all-time top 100 offensive linemen, with NO stats to go by. I'm sure we'd all come to a consensus on that one, right?
.
.
.
And a note on the Barry Sanders thing. We all know his 88 season was epic, but an underrated aspect of this is volume. If you don't account for volume, Felix Jones was better than Darren McFadden. If you don't account for volume, most any RB with fewer carries can be considered better than the guy with more carries. It's a statistical thing (which many here cannot seem to grasp). No, Sanders would not start sucking, had he started 3 years instead of 1, but he also wouldn't have continued averaging what he did in 1988.