joeb33050
posted this
21 January 2014
ACCURACY AND ERRORS
In a perfect world rifles would shoot one-hole groups. It isn't and they don't. There are a number of imperfections, called here “errors", that affect accuracy. Errors might include barrel quality or bullet dimensions or powder charge weight variation. I don't know how many errors there are, or the exact effect of these errors, singly or in combination, on accuracy. (Here “accuracy” is defined and measured by average group size.)
The shooting folklore includes, (among others), two notions.
First and most commonly heard is that there are many errors affecting accuracy, so many, a number so large, that we can never deal with them all. Or identify them all.
Second is the “synergy” contention, typified by this quote from a shooter using “features” in place of errors: “BUT sometimes three or four features that are individually insignificant will collectively make a large difference; as in tuning a rifle."
(While discussing these notions, both the wringing of hands and the throwing up of arms is frequently observed.)
There are some things that we do know about these errors.
Errors vary in effect.
Powder choice affects accuracy more than does shooting-bullets-in-the-order-cast. Bullet choice affects accuracy more than does consistent/inconsistent flash hole diameter.
Errors come in (at least) three flavors.
The “choice” flavor:
IMR4227 or SR4759 ot A#9 are choices, WLP or WLR or Rem 2 1/2 are choices, weight segregating brass or not is a choice.
The “amount” flavor:
12/IMR4227 or 12.3/IMR 2227 or 12.6/IMR4227 are differences in amount.
The “tolerance” flavor:
(Units are grains) Weight segregating bullets into categories +/- .5, or +/-.3, or +/-.1 are differences in tolerances of the categories. Weighing powder charges to +/- 0 or +/- .1 or +/1 .3 are tolerances of the charges.
Controlling some errors is essentially free.
(Thanks to gunlaker on the ASSRA forum for helping get my thoughts organized on this.)
Cleaning primer pockets, uniforming flash holes, uniforming primer pockets, segregating cases by weight, trimming cases to length, turning case necks, visually inspecting/rejecting bullets, weighing/dribbling hard-to-meter powders”¦
There are a lot of errors that may or may not affect accuracy, that require little effort or expense. If we take the actions to eliminate these errors, we don't have to worry or think about them.
Error effect on accuracy is constant.
For example, imagine a rifle shooting 2” groups with load A and 1/2” groups with load B. Loads A and B use the same Remington brass. If using Lapua brass rather than Remington brass reduces load B groups by 1/8", to 3/8"; then using that brass will reduce load A groups by the same 1/8", to 1 7/8". Now imagine another rifle, using the original load A and B brass, and shooting 1” groups. Changing to the Lapua brass will reduce groups by 1/8", to 7/8".
Changing brass as described changes accuracy as described, for any load or rifle.
Multiple error effects don't simply add.
Changing multiple errors to affect group size yields less total affect than the summed effects of the individual changes.
Example: Load A produces 2” groups. Changing from Remington to Lapua brass reduces group size by .125", from 2” to 1.875". Changing the powder charge amount reduces group size by .5", from 2” to 1.5". Changing both, one would suspect, would reduce group size from 2” by .125” + .5” = .625", or to 1.375". They don't. They change group size by .515", to 1.485". (The total change is the square root of the sum of the individual changes squared.)
The above is true only when the errors are independent.
Dependent error pairs include powder charge amount and cartridge case volume; because different cartridge case volumes AND different powder charge amounts both affect pressure and velocity. Both errors affect pressure and velocity; they have a common effect.
Independent error pairs include primer pocket depth and bullet weight variation, since they do not have a common effect.
When multiple dependent errors are changed, the accuracy effect of the changes is (almost always) LESS than the square root of the summed squared effects.
This means that it is likely that small group size changes from small error changes get reduced or even lost in the arithmetic. This may be why data is lacking on the effects of such errors as shooting-bullets-in-the-order-cast, weighing powder charges or bullets to .1 grain, orienting bullets/cases/both, or the myriad of other vigorously defended marginal error changes.
.jpg?width=20)
