Yes. This is a good example of why we work loads up. Even though my calculation called for a bigger charge, you got there with less. This error happened because the 4-shot and 5-shot samples for the lighter loads of N32C had standard errors that made the 95% confidence of the accuracy of the mean number only about as good as their standard deviations. So, a larger sample could well have produced mean values closer together, decreasing the velocity per grain of powder. If you fired, say, 10 shots with the first two loads, a more accurate calculated charge value would result. The 95% limit as a percent of spread in a sample looks like this:
You can see how quickly increasing sample size narrows down the possible errors. Shooting 9 is about twice as precise as 5. Shooting 30 really puts a nail in the coffin of uncertainty.
You can see how quickly increasing sample size narrows down the possible errors. Shooting 9 is about twice as precise as 5. Shooting 30 really puts a nail in the coffin of uncertainty.
