STRIKER 22-250 100 YARDS

John Alexander wrote: Ok Joe there are a few groups that look pretty big.  But very few that are much bigger than twice the average. But nearly as man that are less than half the average. Group sizes aren't likely to be normally distributed.

I think that group size IS normally distributed. Many of my calculations assume this.

I need help to understand what you are doing with the numbers.  Maybe I'm the only one but maybe not.  

With 4 barrels, 2 bullets and one powder; and 80 sets of 5, five shot groups-in addition to some sets of 4 and 3 groups, the data in a spreadsheet or table loses all meaning-at least for me. The data, put into the charts, shows accuracy vs. powder charge. The charts show group sizes that are unusually large, compared to others with the same charge. These are sometimes caused by fliers. The arithmetic required to identify a flier was done, in the past.

 What criteria do you use to decide a shot is an outlier?

EXPECTED AND UNEXPECTED FLYERS               When shooting groups, there is  always one shot, a flyer, that determines group size. With five shot groups that flyer turns a smaller four shot group into a much larger five shot group. That's generally just arithmetic antagonizing the shooter.               When we shoot a set of groups during an experiment, we're trying to estimate the average group size for a given shooter/equipment /conditions system. Sometimes that flyer is so far out that we wonder if it should be included in the calculations.               Some flyers are expected, are just part of the normal variation in bullet hole location. Other flyers are unexpected, a bullet hole very far away from the others.   Here's how to tell an expected flyer from an unexpected flyer.               Measure the group including the flyer.             Measure the group without the flyer.             Divide the group sizes, with flyer/without flyer; and get the “ratio".   For five shot groups: If the ratio is greater than 1.6, we're 95% sure that the flyer is unexpected. If the ratio is greater than 1.8, we're 99% sure that the flyer is unexpected.   For ten shot groups: If the ratio is greater than 1.4, we're 95% sure that the flyer is unexpected. If the ratio is greater than 1.5, we're 99% sure that the flyer is unexpected.

Why do you give averages for as many as four different loads?  it seems to me that that would naturally lead to groups that might be further above the average and in a homogenous set.

To show/see group size vs. powder charge.

JohnI shoot a lot of groups, and I see some fliers. I don't know what causes them, or how to eliminate them. They cause odd big groups. 1 flier in a 5 shot group biases the other 4 shots. 1 flier in a 10 shot group biases the other 9 shots. An argument for 2 shot groups?

The charts sorta suggest that group size variation decreases as group size average falls, but the numbers are very fuzzy.

The charts seem to show that group size vs powder charge is “U shaped", as expected. 

The charts also show that my test plan was/is crappy, much wasted shooting. Maybe there's a hint in there about a better test plan.

joe b.