I enjoyed reading Joe Brennan's article in TFS, on group shooting and mathematics. Lots of good info and I am going to keep a hard copy at my desk for reference.
That said, I do find mean radius a useful metric (and apparently the military agrees with me!). Yes, it takes a few minutes to calculate mean radius (about 5 minutes per target once you set up a spreadsheet to crunch the numbers, or you can even get software where you merely hold the target up to your computer screen and click on the bullet holes).
As the example below illustrates, mean radius is indeed “better” because it gives equal weight to every shot where as group size is often distorted by “flyers". Re: “when you know GS, you know MR, and vice versa.” Well, only if you shoot a zillion shots. The advantage of MR over GS is that you get more data for a given number of shots.
I use GS when groups are lousy -- no need for fancy analysis to prove that a big group s**ks. :D I switch to MR and/or 
