I store weight sorted primers in their original containers with the weights marked on them.
Borrowing 64R's CCI diagram, I start by weighing one and zeroing the scale with it. I'll weigh about 10 more trying to find the center point of the weight distribution. After selecting one that looks like it'd be in the 5.22 - 5.24 range, I do a final zero on the scale and set that primer aside to use as a drift checking standard. The horizontal axis of 64R's diagram will now be labeled -.04, -.02, 0, +.02, +.04. Only 1 digit to react to makes the process go faster and is less prone to error. There isn't any point in doing this on a strain gage scale and a triple beam would make it impractically tedious.
It takes me 3-4 hours to get through 1000. After sorting, the primers are loaded back into trays starting at one end of the distribution moving to the other. The trays are marked with the weight range in them. As pointed out above, 100 primers covers a larger range at the ends of the distribution so I mark the first and last trays. I call them the torn ends. 4 hours is a bit tedious, but it's under 30 minutes per 100.
I'll only put this effort into ELR ammo. Even then, any benefits would be really tough to see in ammo that had SDs much above maybe 6 fps. I'm not a bench rest shooter, but my take on it for 100 yards is if a 3/8" group doesn't completely ruin your day, I doubt you'll see it. For those of us that don't shoot bench rest at that level, the chrono is a better way to look for primer sorting effects.
Let's say unsorted primers contribute 25 fps to the velocity ES of 60 fps. That'd be middle of the road commercial match ammo. Next, let's assume sorting completely eliminates the 25 fps spread. The new spread will not be 60 - 25 = 35. Because the rest of the contributors to the 60 fps ES are random and doing their own thing, it'll likely be closer to:
(60^2 - 25^2)^1/2 = (3600 - 625)^1/2 = 54.5 fps ES
That 5.5 fps drop probably won't be seen unless you're really looking for it and using very large groups to do it. It's probably not yet time to be sorting primers. I load a lot of 223 ammo with 10 shot ESs around 25 fps so I don't buy 26 fps from primer sorting. My guess is it's 1/4 to 1/3 that.
My 223 long range practice load typically has a 10 shot SD of 8-9 and ES's between 25 and 30. LC brass bought new, no neck turning, no primer sorting. This setup is for loading relatively quickly. At 1000 yards, 25 fps gives 6" of vertical with the 75 ELDm. It won't but let's say this load will hold 10" of vertical and that we eliminate the velocity spread. Same error budget / RMS math as before.
(10^2 - 6^2)^1/2 = (100 - 36)^1/2 = 8"
The 2" reduction isn't much unless you're competing at a very high level. Moving from the chrono to a target brings the real world into it. This calc is just for the equipment. You'll never actually eliminate the velocity spread and primer weight is only a small part of it at the 25 fps ES level.
So why sort primers for ELR? Because the contribution of velocity spread to vertical becomes more significant. Even here, it's probably questionable unless your 10 shot SDs are consistently mid single digit for 10 shot strings and you have ready access to a range with a lot of really nice days.
