Outdoor Life on Developing Loads

[DaleA]

Well, it's over MY head but I thought I'd put it out here for others.

Headline:
Node Nonsense: How You’ve Been Wasting Your Time and Money on Load Development

https://www.msn.com/en-us/news/crime...1&sc=shoreline

P.S. I've never subscribed to Outdoor Life but I do remember it used to have a bit of credibility...I may not be current.

P.P.S. I think I'll give this article another read after I've had a chance to digest the first read.

[Unclenick]

Sample size is a big issue. The velocity of "Satterly type 1" tables with one shot per charge weight are worthless, as you discover if you try to replicate them. If you shoot ten or more at each charge weight (a lot of shooting) then take the average of each set of ten and plot those ten-shot averages, you get a straight line and no significant flat spot. The YouTube channel, Winning in the Wind has a fired example, and I've generated them with random numbers in the standard bivariate distribution model, and the same thing happens. You always get what looks like a flat spot and often appear to get nearby ones in some of the subsequent "firings," but when you include all ten and take the average it turns out those apparent flat spots are just randomly generated. Assisted by the human brain's propensity to see faces in clouds and other kinds of patterns where there really isn't one, you think you are seeing them.

The OCW method's three-shot groups and, indeed, any three-shot group is plagued by how high the standard error is for a sample of three. In other words, the standard deviation of the location of the mean is high enough to misidentify a gun's mean POI. But what you can do is use running average plots to narrow that by actually taking nine holes at a time into consideration.

There is, in the end, such a thing as synchronizing bullet barrel exit to muzzle deflections. It is how barrel tuners work, for example. But thus far, shortcut methods of finding them fall short of doing so reliably.