Explaining the problem in a different way might be useful.
At subsonic speeds, one can calculate a "drag coefficient." With this one number, you can predict trajectory knowing only the mass and frontal area of the projectile.
As one approaches the speed of sound, the drag coefficient rises and peaks around the speed of sound. As the speed increases, the drag coefficient drops. A projectile's shape, especially near the nose, determines the shape of a plot of drag coefficient versus speed, and it varies quite a bit. A single number will not accurately describe drag over the supersonic range of speeds, therefore we refer to standard projectile shapes for which the drag curve is known and assume our bullet's drag curve is proportional to that. This is only an estimation, but it works. The accuracy of the estimation is improved if you pick a reference projectile whose drag curve is actually similar to that of your bullet. However, to compare bullets of markedly different shapes, we adjust the same drag curve (G1) to all of them. This is okay for relative comparisons, but if you want to calculate drop at 400 yd down to some tiny fraction of an inch, you should use a better matched drag curve, such as G7.
Estimating the fudge factor for the G1 drag curve using chronograph data is not reliable. The atmosphere is quite variable and throws off the subtle calculations. As MD points out, today radar measurements can accurately read velocity changes over a small distance and remove much of this variation. You get a more accurate BC number, but any BC number references a particular drag curve.
Today's computing power allows one to estimate the true drag curve of a bullet starting with its shape. Hornady has a trajectory program, but I think it uses stored drag curve measurements for each Hornady bullet instead of doing calculations to form the curve. Regardless, this will be more accurate than using the drag curve of a reference bullet. However, out to 100 or 200 yd, it will not be a big difference; your hold will be a greater variable.
A little work with google
brings this page which shows the great difference in drag curves. They all have loosely the same shape, which is why a single BC number can give respectable predictions, but the details are enough different to affect calculations at long distances. Note that at low subsonic, the drag coefficient is constant.
It's kind of funny how they tout the accuracy of such modeling. It's not as if we take a weather measuring station into the hunting fields to let us compute true air density.