i really hate stats. unfortunately, knowing the basics of stats is HUGELY important in scientific testing.
the reason we use confidence intervals and other stats methods to look at our data (deals with gun consistency) -
so you see, we use confidence intervals to look at how different our sampled mean (the data points we take in a test) to the true mean (if we had data on every single event under those conditions).
the equation for a two sided confidence interval is (mean) +/- (tvalue x standard deviation)/ Sqroot(number of events)
the t value is a textbook value based on a normal distribution, and we can decide how inclusive our results are by judging if we want to 90%, 95%, 99%, or even 99.9%. what that means that if we take a sample size of say muzzle velocities, then calculate the CI around it, we choose how inclusive our range is. if we calculate the CI based off a 90% t value, then we can be 90% sure that our next group of samples of velocity will fall in that range. with a CI we can look at exactly how good our sample is compared to the "true" values of ALL shot velocities.
I would also add that the higher the confidence level (99% instead of 90%, etc), the wider the confidence interval becomes. And the larger the sample size, the smaller the interval comes.