Better off shooting a 20, 30, 40 round group on a target at 200m and then count how many rounds land inisde 1MOA, 2MOA, 3MOA of the point of aim. What percentage fall inside of each circle in 1MOA increments will be more telling of the accuracy of the rifle and ammo combination.
If the rifle, ammo and shooter are too good this will not be very telling if you are basicly stacking rounds so close that you get a massive circle of nothing left.
In the case of a BR or competition rifle and ammo you would want to build a jig to stack targets in perfect allignment and shoot 3, 5, 10 shot groups. You would repeat this say 10 times and compare the groups from datam line. You would aggregate the total and individual group.
As an someone that has worked in engineering in automotive manufacturing and aviation that is how I would set up testing accuracy of a rifle if I had to create data for a client. That said there are many ways to skin this cat depending on what data you are actual looking for.
I would not try to do a meta analysis of data that was not created by myself unless I knew all the details of how that data was generated and measured and what model was used to process the data. Sometimes you do not have a choice but often for empirical data you are often better off generating fresh data so long as you can have enough sample size to be significant for purpose. Often we are looking more for trending data than anything absolute unless we have some unknown we are trying to discover and solve for so it can be accountedf for and solved for!
Estimating group size is a fools folly in my opinion. In my youth I lifted weights and played sports. Their was a chart for coverting weight lifted by reps to failure into predictive load capcity at other rep counts. So if you could do 200lbs. for 10 Reps to failure it was suppossed to predict what you could lift for say a 3RM 3 repetition max or 1RM and the other way as well. I never found a single person that this wall chart was ever accurate for. I asked the coach if he had ever met an athlete where that chart predicited their conversion accuracy and after heming and hawing for what seemed like forever he said "No not really." so outside of it breaking up the bad paint job int he gym like wallpaper or a painting on the wall it was not worth while and the very premise of it is absurd.
I have been reloading since 1995 and to be honest only with a few cartridges does reality shake out with predicitive expectations. If you doubt me watch video's about people that reload and then test loads on youtube. Even using loading software looking for nodes and then doing a shooting ladder if you watch enough video's of this it quickly becomes evident that it is not as empirical, logical and intuitive as some would have you believe it is.
There is a very real reason tunners are more popular than ever amoung compittors than ever. Likewise depending on which shooting sport you are talking about some high level shooters are actualy using factory loaded match ammo which no one would even consider 20 years ago in any shooting sport that did not require it like Palma did for some time.