All fair points but since NET isn't a public formula, I do not know what they are looking at or if there are caps on margin. But even if there is, that isn't the same as a cap on efficiency in terms of total possessions vs points, right?.
Even if there was a cap, lets say 20 or 25, there is still a strong correlation to scoring margin and the efficiency sites and the overall rankings. So the cap is almost irrelevant unless you are destroying good to great teams.
In other words, the more that a team beats the spread on the efficiency sites spreads (not odds makers), the more the teams overall ranking is going to be effected on the efficiency sites and on the NET.
That's why when Pitt beat North Carolina and Virginia, Pitt hardly moved on torvik, pomeroy, and on NET when it it was 2 great wins. Why? Because the efficiency metrics of a site like tovik had the spreads very small already, much smaller than the oddsmakers. And Pitt barely won. Thus, little movement in the actual ranking. Efficiency sites are not odds makers.
When Pitt significantly beats the spread on the efficiency sites actual lines on the games, then, you are going to get significant movement.
Efficiency sites don't care about wins and losses like they do about efficiency. So you could play a bunch of teams that are rated high in efficiency, lose every game by a close score, and still be rated high on the efficiency sites. The same cannot happen with the RPI, you actually have to win too to get a high rpi ranking.
Thats why if Pitt is up 20 on a great team, inserts all subs and the lead goes to 10 at games end when its already over, its a real bad idea if you want to gain ground on the efficiency sites and on the net and actually improve your ranking. The win margin against good to great teams is very important.