JD

You and Brent may have opened a can of worms here.  I have an Excel spread sheet going back about nine years to compare group sizes with different ammo brands and lot numbers fired in specific rifles.  I shoot 10 shot groups and measure the outside spread of each group.  

Many times I'll have a 5/8 inch group with nine shots and one will blow the group out to one and an eighth inch.  That group is recorded as being one and an eight inch.  It seems to me that if you do anything otherwise, it will throw all of the comparative information out the window.

I have a general working knowledge of statistical formulas as applied to economic analysis and have them programmed into an Excel spread sheet to calculate various things.  Some of you who are more adept at such things may know how to do comparative accuracy testing without including all shots fired in the group, but I don't. It would be a fatal flaw in the data collection and comparison to do anything less, but then, I'm not a statistician and I don't know everything about this subject.  What little I do know leads me to believe statistics gives a decision making edge where experience does not serve or is missing. The numbers are unemotional and without prejudice.

So will someone please explain to me how you can do comparative accuracy testing and not include all shots, including flyers, in your data collection and statistical analysis?  If I dropped the flyers out of my data, I could look like a better shooter than I am.  The only trouble with that approach is the scoring judges count those flyers at matches around here.

What is wrong with those guys?

To my mind, if you can honestly "call" a flier out because you KNOW you shot it wrong, it's one thing---basically an AD because you misread the wind or something. 
  However when seriously testing ammo its the frequency of "non-shooter induced fliers" caused by mfg variations that we are specifically trying to identify in order to determine quality of brands and batches.
  Some ammo just won't group worth squat in a given rifle. Then there is the heartbreaking inconsistant good stuff that will shoot 80 or 90% into nice tight groups then throw the odd round or two outside somewhere.   Seems like the lower end target grade stuff is good but inconsistant the big buck stuff is both good AND consistantly so.
I've been shooting Wolf a lot.  In general, (for me and my rifles that is) the M/T and the M/E seem to shoot about the same size  groups,  the difference between them is that the M/E produces fewer "fliers"

JD

It looks like you are as anal about this as I am.  I keep all my targets arranged by month, by year, and compare groups fired in the same month and year just for the heck of looking at the data that way to see if there is any correlation to group sizes I fired this year in May with those I fired last year in May with the same ammo and rifle.  

I also note the velocity of each round fired and where it strikes the target.  I correlate the velocity SD with the group size SD.  I throw out the high and the low velocity and then take a measurement of the group size without those being included.  I correlate those group sizes with those that include the high and low velocity.  I look at every statistical aspect Excel is capable of including in its formula programming just because it's easy to do once you write the formulas and put them into a cell, column, or row.  Then all I have to do is put in the data and there are all of these crunched numbers to play with and graphs to look at.

But I never fool myself by thinking I am any better shooter than the simple average (mean) says I am because on any given day, that is what I am going to shoot.  Try as I might, I have not figured out a way to use all of those numbers to make me shoot a dam bit better and when I blow a match and tell the judges what my average is, they just smile and wish me better luck next time.  Those rascals only count the bullet holes in the target I shot that day during the match.  They are completely unreasonable, or have you noticed that? 

Quote:


If I honk one out and call it out before looking through the scope, I don't include it in the group. My (arbitrary) rule.
With some revolvers, one chamber may throw a shot out of the group. I record "6 shot group size" and "best five shot group size".
With, particularly a Ruger MK II, the first shot in 10 shot 50 and 100 yard groups is often out of the group. I'm trying to find out why and fix it, frequently measure "10 shot" and "best 9 shot" groups.
With 22 RF ammunition I find big weight variation in loaded ctgs, little weight variation in fired cases.   
I wonder if it is possible to weigh/measure the 22 ctgs. and see which are "fliers".
joe brennan