I don't find measuring case volume with water to be difficult, but it is time-consuming and I don't do it for all of the hundreds of cases I use in competition. On numerous occasions I have measured case volume with water for cases of known weight. There is always a general inverse trend for case weight versus case volume. However, there are always a few outliers where case weight does not match up well with case volume (i.e. they're well off the trend line).
Regardless, I still sort cases by weight and I figure even if case weight doesn't always correspond perfectly to case volume, the weight sorted case will still be more uniform in terms of volume than if I do nothing at all. I can weight sort a few hundred cases in well under an hour. Sorting the same number of cases using water volume would take considerably longer.
In response to Seymour's original question, I think the best answer lies in the concept of limiting source of error. Which condition would have the greater effect on velocity/pressure: 1) varying charge weight by a tenth of a grain in cases of the
exact same internal volume, OR 2) having the
exact same charge weight in cases where the internal volume may differ by a tenth up to a few tenths of a grain?
In this example, I think Quickload might give some insight to which factor (case volume versus charge weight) might be the larger source of potential error, at least in as much as you have faith in QL predictions/outputs. I used one of my .223/90 VLD loads as a starting point and 1) adjusted charge weight up by 0.1 gr, or 2) adjusted case volume up by 0.1 gr. Here are the QL predictions/outputs:
As you can see from the Table, increasing charge weight by 0.1 gr had a 3-fold greater effect on predicted velocity/pressure than did changing case volume by 0.1 gr. In this example, I know from testing that the predicted increase in velocity going from 24.2 gr to 24.3 gr is spot on the increase I actually measure with the chronograph. I have not ever loaded cases that differed in volume by ~0.1 gr and tested those to determine how good QL's prediction there actually. However, these predictions at least suggest that as measured in grains, charge weight will have the greater effect on velocity/pressure by a factor of slightly more than 3X.
During the reloading process, there are many things we do as reloaders that probably go far beyond what we can realistically assess with confidence on the target. Identifying and understanding the the largest potential sources of error can be a useful exercise in deciding what steps you consider "essential" and those you may not always choose to implement, except perhaps for really critical matches. It is important to note that the one element of this process that cannot be defined other than in individual terms is the peace of mind and/or confidence you get by going the extra mile. I'm positive that on a theoretical basis, some of the reloading steps I carry out for important matches have an effect so small as to be statistically insignificant. However, I place a high value on having confidence when I get behind the rifle for a match that there is little more I could have done to improve my ammunition. For that reason alone, it is personally worth it to put forth the extra effort. Everyone must decide for themselves what steps they consider "essential" and what steps may not be worth the extra time and effort.