entoptics
Well-Known Member
- Joined
- Jan 16, 2018
- Messages
- 846
Just a little pot stirring. If I had a dime for all the times I've heard people comment on the accuracy of their particular gun/setup, with the statement "capable of" or "when I do my part", I'd be able to buy a nicer rifle...
As an analytical chemist, providing data to researchers across the globe, I can't imagine releasing data with the caveat "our instruments are capable of XXX precision, when I do my part...". Instead, we monitor long term precision and accuracy, using rigorous blinded testing procedures, and that is what is reported. Period.
Each sample we analyze is a unique event, just like each round down range, and the only way to reliably predict the precision and accuracy of the outcome of such an event, is to aggregate the results of many of them and apply statistics.
As an example of how misleading small data sets are, see the following groups. These were fired back to back today, in good conditions, bipod, rear bag, prone in a wheat field (200 yds).
So, it's obvious that my rifle is "capable of" half MOA with this load "when I do my part".
Or is it...
Behind the trigger, I thought I pulled a shot in each group. Not much, but a whisker off center when the recoil started. More importantly, I've shot and recorded a half dozen 4 shot, and a couple of 10 shot groups with this load recently (in similarly good conditions), and they are averaging about 0.8 MOA (which I'm very happy with, btw). Had I been doing load development with only the above sample to reference, I would have sworn that I'd found the "node" or some such.
I have spreadsheets full of data (both for rock analyses and shooting results) that show this type of wishful thinking.
In summary, even a blind squirrel finds a nut sometimes, and worse for "common sense statistics", sometimes they'll find one twice in a row. The important part though, is to know when you're the squirrel that might need glasses...
As an analytical chemist, providing data to researchers across the globe, I can't imagine releasing data with the caveat "our instruments are capable of XXX precision, when I do my part...". Instead, we monitor long term precision and accuracy, using rigorous blinded testing procedures, and that is what is reported. Period.
Each sample we analyze is a unique event, just like each round down range, and the only way to reliably predict the precision and accuracy of the outcome of such an event, is to aggregate the results of many of them and apply statistics.
As an example of how misleading small data sets are, see the following groups. These were fired back to back today, in good conditions, bipod, rear bag, prone in a wheat field (200 yds).
So, it's obvious that my rifle is "capable of" half MOA with this load "when I do my part".
Or is it...
Behind the trigger, I thought I pulled a shot in each group. Not much, but a whisker off center when the recoil started. More importantly, I've shot and recorded a half dozen 4 shot, and a couple of 10 shot groups with this load recently (in similarly good conditions), and they are averaging about 0.8 MOA (which I'm very happy with, btw). Had I been doing load development with only the above sample to reference, I would have sworn that I'd found the "node" or some such.
I have spreadsheets full of data (both for rock analyses and shooting results) that show this type of wishful thinking.
In summary, even a blind squirrel finds a nut sometimes, and worse for "common sense statistics", sometimes they'll find one twice in a row. The important part though, is to know when you're the squirrel that might need glasses...