Originally Posted by
arcfide
Schwarz, I know that you have said that you have 890+ sample data points for this test, and I greatly appreciate the ANOVA results. I have a few more questions about that though. I assume, hopefully, that these are all the data points, and that no cherry-picking or removal of outliers has been done? The other question, which I don't recall you answering, is how these data points are distributed across different manufacturers, weights, bullet designs, temperatures, &c. It's less convincing if most of these data points came from a relatively small sample size of different types of projectiles in a small range of velocities. If, on the other hand, you did something like 10 rounds for each individual cartridge loading, and thus tested 89+ different rounds of substantially different design and loading, then that would be more compelling I think, especially if they covered a wide enough range of calibers, including .22 and .44 magnum and everything in between.
If I understand you right, these models will only work for bullets designed for straight line expansion without tumbling and little to no fragmentation, yes? Would they, for instance, accurately predict penetration for Liberty 50gr 9mm +P for instance (impact velocity: 2000fps, 0.37" diameter, 27gr retained weight)? Sorry, I haven't read through your formulae carefully enough to plug in the numbers myself yet. What about for tipped rounds that include polymer tips, such as Critical Defense, Critical Duty, or Speer G2 rounds? Does this model scale to the velocities and bullet weight combinations found in non-fragmenting, non-tumbling rifle rounds, such as the 50gr TSX Black Hills and the 64gr BSB from Nosler or the 52gr Federal FBI round? Where does the model break down? Is there an upper or lower velocity, weight, or expansion threshold at which the predictive power of the models begins to break down?