I remember seeing Ken Oehler's article, but didn't grasp the significance of what he discovered until I read this thread.
It seems most long range shooters missed the significance of Ken's article. Simply stated, one drag function is as good as another for predicting trajectory at supersonic velocities. A lot has been made of using G7 in the last few years, but unless you intend to shoot to ranges where your bullet goes subsonic, the numbers in the original post show it's not significantly better at predicting trajectory then G1.
I remember one test where the measured BCs for 10 shots were all within half a percent of each other, but much higher than expected. Turned out we entered the wrong distance to the target. I bring this up because your statement that "...he and others confuse repeatability with accuracy" rings true in my own experience. I image you've learned that lesson from your own experience.
Confusing repeatability with accuracy is an easy mistake to make particularly for shooters for whom repeatability seems almost synonymous with accuracy. As you discovered, however, an incorrect measurement or assumption can produce highly repeatable numbers that are grossly inaccurate. That's one reason scientific papers are peer reviewed before they are published. Even then, the results are sometimes garbage.
