Shorty: I applaud your effort to measure bullet weights. That is alot of work and we all appreciate it. But the entire force of acceleration is due to the chemical energy content and burning performance of the powder and primer. Your post acknowledged that the resolution of the weighing device was not sufficient to discern any differences. But 0.1 grain resolution we know a-priori, is not sufficient resolution for very small volume/mass of powder. As the volume and mass of powder decreases, the tiny differences in powder mass from case to case create proportionally larger acceleration force differences. 0.1 grain of powder difference in a .338 magnum caliber cartridge is close to negligible. But that same 0.1 grain powder difference in a .22 rimfire case could be highly significant, and therefore requires 2,3, maybe 4 decimal places to explain the kinetic energy differences caused from the powder burn dynamics).
Many a science experiment fails from the get-go due to non-suitable apparatus for the required accuracy, and methodology flaws, leading to invalid results and conclusions.
What comes first is the force applied to accelerate the bullet, and that has everything to do with the energy conversion from chemical to kinetic, which all comes from the powder and primer (assuming an even firing pin strike force).
None of this comment negates your evidence of significant differences in bullet weight. I would guess those differences in your data do indeed affect trajectory and POI to very significant numbers. However without the powder weights controlled for, they are hopelessly confounded.
(Aside: Your graph of bullet weight proportions per brick is good work, but stats-wise it should be shown as a bar graph. A line graph connecting data points implies a continuous relationship between mass and proportion, which is not the case for these variables. They are distinct. I know I know - picky statistics nerd stuff!)
(Aside 2: The "line" graph, which should be a bar chart, does show what looks like approaching a normal distribution bell curve. That is a good sign indicating that your sample size was approaching sufficient to describe the distribution. There is still that jagged double peak in the middle which is concerning because it will confound the mean and SD. Obviously the bell curve would become more refined and symmetrically shaped as the sample size increased, and likely bell-up that double peak in the middle to reflect a more accurate mean, and thus a better SD. Perhaps a few dozen more bullets in the sample would likely smooth that out and fill in the double peak. Adequate sample size is not easy to attain in this rimfire science, so I appreciate the effort).
Last edited:
) alone for a couple weeks and this is what I come back to? Good grief. Evidently, any further commentary I might have on the matter would only fall on deaf ears and thus would be a waste of energy. Some folks just have to come to their own realizations, however, if they're stuck in a bias they're unlikely to do the work necessary to bring about such a realization, unfortunately.
