Basically it’s a trick question, that requires a very simple complicated answer. Without using the big words and scientific jargon, you have to accept that there are basically two definitions or results of bullet “Stability” or not.
Maximum efficiency of flight.
Bullets sideways in the target.
There are also different ways to predict mathematically how any given bullet will perform.
Stability calculator
Stability predictor
The most common is the Stability predictor. It uses four inputs, bullet length, weight, velocity and twist. I call this a predictor because it makes a lot of assumptions. Simply put, it doesn’t factor in bullet shape, it assigns one. The formula seems to work great until you take the exact same bullet, load it backwards and forwards expecting the same results on target. This type of “Stability calculator” does not factor in actual bullet shape, and why a bullet assigned a factor of 1.2, May out perform a bullet assigned 1.4.
A Stability calculator that includes a drag function will require specific bullet measurements. With this type of calculator, the above mentioned anomaly of a bullets assigned 1.2 out performing a bullets assigned 1.4, disappears. It will generate different numbers if you reverse the inputs and use the base for the tip. It knows if your shooting the bullet backwards or forwards.
If you’re willing to accept that very basic premise, then the idea that there is no such thing as a marginally stable bullet, helps the next idea. Really more than idea, experiential results, without scientific formulas.
There’s no such thing as marginally stable, it either is, or is not.
Stability/drag is important for long range precision. It’s more about maintaining the most efficient bullet flight from muzzle to target. Since velocity drops off faster than spin, the longer it travels, the better it flies. In theory.
Stability/wobble throws a curve to the normal idea of a bullet being more stable as velocity declines, when bullet spin to velocity ratio increases.
The idea of a bullet starting out with a slight wobble and then “going to sleep” down range, starts becoming a reality you can see with sub-sonic shooting. Most people never worried about this until they bought a suppressor and blew the baffles out with a bullet that showed perfectly stabilized by a Stability predictor. When people started checking for keyholes no farther from the muzzle than where the muzzle blast would not tear the paper, interesting things started showing up.
Setting up targets every 25 yards out to 2-300 yards shooting sub-sonic allowed you to track stability. Round holes at 25, a slight smear at 75, full on key hole at 150, completely sideways at 200. The fact is that the bullet was unstable at the muzzle, it never went to sleep, the wobble just kept increasing instead of decreasing until it tumbled. Sometimes the opposite. A slight smear at 10 yards, a round hole at 100.
It’s a long way of saying, both answers, an increase or decrease in “stability” over distance can be correct. The type of formula used to predict that is very important. Actual bullet shape is a must, as is actual field testing.
There are plenty of examples of bullets that have exceptional performance at 100, 200, 600 yards that people won’t consider using at 1000. It’s not because of spin rate falling off. Something else in the bullets shape and balance points, cause a less than ideal efficiency of flight. That’s why I say a “stability predictor” with limited inputs is more a general idea. Where a “stability calculator” that includes bullet shape will make a more accurate prediction and help make the numbers make sense. Knowing the actual drag curve is helpful.
Short version is that some bullets that are labeled “marginally stable” that have problems at distance, were actually probably never truly stable to begin with. It just takes longer for the instability to show up.