Keep in mind this is a zombie thread. It hadn't had a post since 2012 until Schleeb posted, so many of the participants have moved on. Bart stopped posting here awhile back, but posts on other forums. I don't know why he dropped us, but I did contact him to make sure he was OK and he said he is.
Bart's post, is basically correct. Bullet balance in this context is mass symmetry around the longitudinal axis of the bullet. Because a bullet spins around its center of gravity when it is in flight, one whose CG isn't centered will have its surface spinning eccentrically. This results in what is commonly called bullet trajectory line wobble, as the trajectory then has a corkscrew shape to the extent the nose position flies eccentrically.
Bart posted in the past about Mid Tompkins's experiment with D46 FMJ's some decades ago. He got a special collet made for a Dremel tool that held the bullets. When he spun the tool up, the eccentric wobble of the unbalanced ones created enough side loading on the Dremel bearings to measurable increase the current draw of the motor. Mid used this to sort bullets, with the lowest current draw bullets segregated out as being best. From one of Bart's posts on this:
Bart B said:
The perfect ones put 10-shot groups from .7 to 1.5 inches at 600 yards, the others {were} inside 10 inches with {a} few in the middle.
Harold Vaughn did a version of this experiment for his book Rifle Accuracy Facts, in which he put dynamic headphone elements on a small tube machined and drilled to suspend a bullet on an air bearing and spin it using compressed air. Eccentricity caused the bullet to spin closer to the side of the tube, creating a sound output at the spin frequency that got louder with greater asymmetry. He, too, found bullet seated this way group better than unsorted ones.
There was also a tool designed by Vern Juenke that used ultrasonic sensors to identify jacket thickness irregularities. It was concluded by him that this was the primary source of mass asymmetry in cup and core bullets, and his apparatus turns bullets slowly while ultrasonic transducers measure jacket thickness by wave reflection.
Someone bought him out, but their site doesn't leave you confident they have any real idea what they're doing. They sell only parts kits for close to 2 large, then tell you that you'll need Rainman skills to make it work and won't guarantee you can make it work and won't fix it if it doesn't. I get the impression they don't know themselves how to put one together and make it work reliably. I haven't seen these kits widely available. It's yet another bullet sorting scheme, as the other two methods are.
So, why does wobble reduce accuracy? It does increase drag (decrease effective bullet ballistic coefficient) a little because the CG and center of pressure of a bullet are co-located in a balanced bullet, but not in a wobbling one. That off-center CG results in the bullet constantly having some amount of coning motion, where its nose and base describe small circles whose centers pass through the CG. In a perfectly balanced bullet this settles out into the yaw of repose, but never settling out completely means always exposing a tiny amount of side profile to the air stream, which increases drag in proportion to the square of the sine of the angle off true. So that causes vertical stringing.
Of more significance, though, is that the bore of a rifle holds a bullet centered on its geometric axis even if the bullet mass is off-center. This means the center of gravity orbits the centerline of the bore as the rifling turns it. When the bullet exists the muzzle, that eccentric CG lobs the bullet perpendicular to the axis, tangent to the location of the CG in its path around the bore axis, at a speed equal the the angular velocity of the CG times its distance from the bore axis. This sideways lob is drift. It is small and too slow to create enough drag for it to slow significantly during the bullet's time of flight, so the bullet just drifts in that direction at that speed.
Suppose, for example, the CG were 0.001 inches off axis in a .308 caliber bullet fired at 2700 fps from a 10" twist barrel. The common bullet RPM formula,
Where:
MV is feet per second
P is inches per turn of the rifling
720 × MV / P = 194,400 rpm
Divide by 60 to get revolutions per second (rps):
194,400 / 60 = 3,240 rps
Angular velocity is 2 × pi × 3240 = 20,356 radians per second
Times the radius of rotation is:
20,356 rad/s × 0.001 inches/rad = 20.4 inches per second.
So, if the bullet has a transit time to 600 yards of 0.87 seconds, it will drift 20.4 inches/s × 0.87 s, or 17.75 inches away from the mean point of impact by the time it gets there. In what direction it will drift, you can't predict unless you know what side of the bore the CG was closest to when it cleared the muzzle.
The point of bothering with the calculation is to show how little error there needs to be in the CG location off-axis to have it show up on target. Bullet symmetry is critical to accuracy.
Also note that most modern bullet shapes don't have their CG and geometric center in the same place. This means that if the bullet enters the bore even slightly less than straight, the CG will have some amount of eccentric spin and some amount of resulting drift.
