Without further comment at the moment I wanted to draw forum reloader's attention to this article:
https://bisonballistics.com/articles...nt-for-f-class
Printable View
Without further comment at the moment I wanted to draw forum reloader's attention to this article:
https://bisonballistics.com/articles...nt-for-f-class
Interesting article tx.
His graphing is really good, and I like the fact that he adds incremental data into the mix and gets a very visual representation of how that alters the picture.
I do load first and then play with seating depth - 2-step process. Can see how combining the 2 could eliminate some errors. But 185 rounds is a fair chunk of barrel life for f-class !
I think this approach acknowledges that even a single 5 shot group has quite a variable size and limited predictive value for the long term average size for a given load. Rather than firing 5 x 5 shot groups to home in on the real potential of a load the author uses trends among nearby loads, which is probably pretty good provided both charge weight and seating depth are equivalent in their contribution to group size.
I would guess the most common error in load workup is discarding a good load because of one poor shot due to shooter error. Who can fire 20 or 50 rounds without a couple of less than perfect let offs? So this method does cover that problem.
I think running a 2 dimensional survey like this would give me a lot of confidence, provided I had already settled on bullet powder and primer etc. Stepping in one variable at a time as most people do requires faith that each successive parameter is dominant and independent so that its superiority will be maintained as you tweak the next variable. For example the best powder weight at one seating depth will also be the best weight at other CBTO (I didnt know a word for this before), then that changing the primer won't change the best powder and depth after that ...
My suspicion is that most people follow a one dimensional random walk across the load development landscape and could miss those green spots, if they truly exist.
A key decision would be deciding what increments of powder and CBTO to use in the first trial to get enough data points to be sure you haven't missed a sweet spot. Probably a surface plot in excel would show the trends well enough. The fact that the author "just happened to know how to use R" shows the line of work he's in.
Yes 100 rounds is a lot and not justified for hunters but if you need 0.5 to be competitive then you will use that and more testing loads at random.